![[PDF]](/pb-assets/Icons/adobe-pdf-icon-1498649896243.png){kind=icon}
Abstract
New emphasis is being placed on the importance of parent involvement in children’s education. In a synthesis of research on the effects of parent involvement in homework, a meta-analysis of 14 studies that manipulated parent training for homework involvement reveals that training parents to be involved in their child’s homework results in (a) higher rates of homework completion, (b) fewer homework problems, and (c) possibly, improved academic performance among elementary school children. A meta-analysis of 22 samples from 20 studies correlating parent involvement and achievement-related outcomes reveals (a) positive associations for elementary school and high school students but a negative association for middle school students, (b) a stronger association for parent rule-setting compared with other involvement strategies, and (c) a negative association for mathematics achievement but a positive association for verbal achievement outcomes. The results suggest that different types of parent involvement in homework have different relationships to achievement and that the type of parent involvement changes as children move through the school grades.
In the past decade, the importance of getting parents involved in their children’s education has received considerable attention from policy makers, educators, parents, and the mass media. Central to this heightened awareness is the No Child Left Behind Act of 2001, in which parent involvement was identified as one of six areas requiring reform. Other national initiatives that have advocated partnerships between parents and schools include Project Appleseed, a nonprofit group that asks parents to sign a promise to be involved in their children’s schooling, and the National Coalition for Parent Involvement in Education, funded by the Ford Foundation and Union Carbide. A Time magazine article titled “How to Make a Better Student: The Eight Secrets of Success” (Wallis, 1998) emphasized parent involvement as being a key component of efforts to enhance achievement.
Parent involvement in education can take many forms. For example, parents may be involved by communicating with the school, volunteering at school and participating in school decision making, or supporting learning at home (Epstein, 1995). To the latter end, encouraging parent involvement with homework is a strategy commonly practiced by schools and teachers in an effort to facilitate academic achievement (Cooper, 1989; Roderique, Polloway, Cumblad, Epstein, & Bursuck, 1994). Parents agree that they have an important role in homework (Hoover-Dempsey & Sandler, 1995, 1997). Of all the types of school involvement, parents report that helping with homework is particularly effective for enhancing achievement (Epstein, 1986; Sanders, Epstein & Connors-Tadros, 1999). Parents’ behavior corroborates this belief; 90% of parents report setting aside a place for homework, and 85% report checking to see that homework has been completed (U.S. Department of Education, 2005). Finally, students generally feel that when parents help, they do better in school. Balli (1998) found that 95% of students reported that they did better in school at least some of the time when they received help with homework from their parents.
Despite widespread beliefs that parent involvement in homework is linked positively with achievement, research suggests that the relationship may not be a simple one. This is because the association between parent involvement in homework and achievement may be influenced by numerous factors, such as the involvement strategy parents use, the child’s age and ability level, resources in the home, and the parents’ own mentoring skills. In the following section, we review theory and research addressing the effect of parent involvement in homework on achievement and achievement-related outcomes.
Review of Theory and Research Addressing the Effects of Parent Involvement in Homework
Homework can be defined as tasks assigned by schoolteachers intended for students to carry out during nonschool hours (Cooper, 1989). Research has clearly shown that students, especially adolescents, who do homework, benefit in school by exhibiting improved achievement (Cooper, 1989; Cooper, Robinson, & Patall, 2006). However, researchers have suggested both positive and negative effects of parent involvement in homework (see Cooper, 1989; Corno, 1996). Table 1 presents a list of potential outcomes of parent involvement in homework, suggested by empirical research or by theoretical formulations. These sources have also suggested that parent involvement in homework can affect both achievement outcomes and outcomes related to achievement. As such, we discuss theory and research on the effects of parent involvement separately for these distinct sets of outcomes.
Effect on Achievement
The most commonly cited benefit of parent involvement is that it can be used to accelerate learning by increasing the amount of time students spend studying and making homework study more efficient, effective, and focused. A cursory look at research on the effect of parent involvement in homework on achievement suggests mixed findings. Some studies indicate that parent involvement in homework has a positive effect. For example, quasi-experimental and longitudinal studies have been conducted on the effectiveness of the Teachers Involve Parents in Schoolwork (TIPS) program, in which interactive homework assignments that require the involvement of parents are assigned. These studies found that parent participation in TIPS significantly enhanced student writing scores (Esptein, Simon, & Salinas, 1997), as well as homework grades and class grades in science (Van Voorhis, 2003). Some correlational evidence also supports the effectiveness of parent involvement in homework, suggesting that setting a place and time for homework and providing direct aid with homework are effective ways parents can enhance achievement (Deslandes, Royer, Potvin, & Leclerc, 1999; Yap, 1987).
The results of other studies suggest that parent involvement in homework has a negligible or negative relationship with achievement (Epstein, 1988; Ginsburg & Bronstein, 1993). Given that a great deal of the research on the relationship between parent involvement in homework and achievement has been correlational, one potential explanation for contradictory findings is that the relationship may be bidirectional. That is, although parent involvement in homework may improve student achievement, it may be also be that low-achieving students are more likely to receive greater parent involvement. There is some evidence to support this hypothesis. Pomerantz and Eaton (2001) found that children’s poor academic performance predicted heightened involvement in homework 6 months later, which predicted improved achievement over time after initial achievement was taken into account. However, this explanation may not entirely explain contradictory results. Even in some research using more powerful experimental or longitudinal designs, null or negative effects of parental homework involvement have been found (Balli, Wedman, & Demo, 1997; Levin et al., 1997).
Finally, other studies suggest that it is the type of parent involvement that is critical. Parents engage in many different types of involvement strategies, such as providing space and materials for homework; interacting with the teacher about homework; providing general oversight or monitoring of completion; making rules about when, where, or how homework is done; responding to questions about homework and giving feedback; or actually providing direct homework instruction (Hoover-Dempsey et al., 2001). Different forms of parent involvement are likely to have distinct effects on student achievement, and these effects may even vary depending on characteristics of the student, which could explain contradictory findings in the literature. According to motivation theorists (Grolnick & Ryan, 1989; Pomerantz, Grolnick, & Price, 2005), forms of parent involvement that support the child’s autonomy, as well as provide structure in the form of clear and consistent guidelines about homework, will be the most effective. However, forms of involvement that are experienced by the student as controlling will have little impact or a negative impact on motivation and achievement. Research supports this assertion. For example, a study conducted by Cooper, Lindsay, and Nye (2000) found that parent involvement in homework in the form of support for autonomy was associated with higher standardized test scores, class grades, and homework completion, although direct aid was associated with lower test scores and class grades.
Effects on Achievement-Related Outcomes
In addition to direct effects on achievement, parent involvement may have positive effects on a number of desirable achievement-related outcomes, as well as indirect effects on achievement through such variables. As would be expected, parent involvement in homework has been linked to the most proximal achievement-related measures, including improved homework persistence, understanding, performance (Callahan, Rademacher, & Hildreth, 1998; Forgatch & Ramsey, 1994; Hutsinger, Jose, & Larson, 1998; Natriello & McDill, 1986; Voelkl, 1993), and in particular, homework completion (Cooper et al., 2000). Given research suggesting that students who do homework outperform students who do not, the effect of involvement on completion rate may be one route by which parent involvement influences achievement outcomes.
Students may experience homework as more pleasant when parents are involved. When parents are involved, students report being more attentive and having a more positive mood (Leone & Richards, 1989), greater homework enjoyment (Shumow, 1998), and perceive their homework activities as less difficult and more manageable (Frome & Eccles, 1998). When homework is more enjoyable, students may get more benefit from engaging in it.
Parent involvement may lead to enhanced achievement by facilitating the communication of expectations to children and providing opportunities for reinforcement of desired homework behaviors (Hoover-Dempsey et al., 2001; Maertens & Johnston, 1972), as well as facilitating communication between parents and teachers (Epstein & Van Voorhis, 2001). Parents who are involved in homework have an opportunity to demonstrate their belief that schoolwork, homework, and learning are important (Epstein & Van Voorhis, 2001) and to show support for what their children are learning (Balli, Demo, & Wedman, 1998; Levin et al., 1997). Research has shown that parent involvement in the form of indicating positive attitudes about homework is related to the development of positive attitudes about homework and schoolwork for the student (Cooper, Lindsay, Nye, & Greathouse, 1998). As such, parent involvement in homework has also been linked with more positive behavior during school (Sanders, 1998), suggesting that involvement may affect achievement by affecting learning during school hours as well as at home.
Parent involvement in homework may also influence long-term achievement and related outcomes by promoting the student’s ability to engage in adaptive self-regulation. That is, parent involvement may promote the development of cognitive, affective, and behavioral strategies—including goal-setting; planning; and the management of time, materials, attentiveness, and emotions—necessary to achieve academic goals (Zimmerman, 2000). Research supports this assertion, suggesting that students demonstrate more effective study habits when their parents are knowledgeable about the homework task (McDermott, Goldman, & Varenne, 1984; Xu & Corno, 1998). Consequently, parent involvement may be particularly important for younger students who lack self-regulatory skills and are in the process of developing self-management and study habits.
Despite these positive benefits of parent involvement, some researchers have highlighted that involvement may lead to negative experiences for parents and students. Levin et al. (1997) found that greater maternal help with homework was related to increased fatigue, frustration, and disappointment for the mother and caused tension between mother and child, particularly when the child was a low achiever. Furthermore, emotional costs and tension between the parent and child may be compounded when parents, although believing homework is valuable, experience frustration about having inadequate skills for helping children with homework, especially older students for whom the curriculum is most challenging (Dauber & Epstein, 1993; Reetz, 1990), and have constraints in terms of their own time and energy (Hoover-Dempsey, Bassler, & Burrow, 1995; Kay, Fitzgerald, Paradee, & Mellencamp, 1994; Reetz, 1990).
Some forms of parent involvement in homework may be adaptive in that homework completion and learning are facilitated, and the development of positive attitudes and self-regulatory study skills associated with academic achievement may even be supported (Hoover-Dempsey & Sandler, 1995). However, Cooper et al. (2000) found that two thirds of parents reported engaging in some inappropriate form of involvement not expected to have a positive effect, including simply giving correct answers or completing assignments themselves. Consequently, parents engaging in these inappropriate involvement behaviors may impede learning during homework study and hinder the development of self-regulatory skills if students come to rely on their parents for correct answers or external regulation and motivation. Further, parent involvement in homework may have detrimental effects if it is self-initiated on the part of the parent without request from the child or is perceived as intrusive or controlling by the child. In analyses assessing the effect of mother’s daily self-initiated involvement on success and failure the following day, Pomerantz and Eaton (2001) found that although self-initiated monitoring and help with homework did promote school success the following day, it also fostered school failure for low-achieving (but not high-achieving) children.
Finally, some educators and researchers have suggested that parent involvement may exacerbate differences between high- and low-achieving students when the achievement difference is associated with economic differences. It might be more difficult for families to be involved in homework if they are of limited economic means, if there is a single parent, or if both factors are the case (Grolnick, Benjet, Kurowski, & Apostoleris, 1997; Kronholtz, 1997; McDermott et al., 1984; Odum, 1994; Scott-Jones, 1984). Still, parents who participated in the National Education Longitudinal Study in 1988 reported providing similar amounts of homework supervision, regardless of socioeconomic status (Sui-Chu & Willms, 1996).
In sum, then, research and theory suggest the relationship between parent involvement in homework and achievement or achievement-related outcomes is complex. Further, it seems reasonable to suggest that the positive and negative consequences can both occur and can even occur together. For instance, parent involvement might improve homework completion rates at the same time that it creates tension between parent and child. In this article, we use meta-analysis to examine the cumulative evidence on the relationship between parent involvement in homework and achievement, as well as a number of related outcomes. Given conflicting findings within the literature on parent involvement in homework, a meta-analysis of the literature would help answer the most basic question of whether parent involvement in homework does indeed have a beneficial effect and for what particular achievement-related outcomes. Further, factors such as the type of involvement, the subject matter of the homework, the achievement outcome assessed, and the grade or ability level of the student may all be variables that affect whether parent involvement in homework will positively influence achievement. A meta-analysis would help identify what moderators might explain the conflicting findings.
Past Syntheses of Parent Involvement in Homework
Cooper (1989) reviewed eight studies examining the relationship between achievement and parent involvement in homework. No study was included that directly manipulated the presence versus the absence of involvement, but two studies did partially address the question of a causal impact. Cooper concluded, “There is as yet no reliable evidence on whether parent involvement in homework affects student achievement” (p. 140). Furthermore, five studies that related the amount of parent involvement to student achievement produced correlations ranging from −.22 to +.44.
Hoover-Dempsey et al. (2001) provided a narrative review of the research on parent involvement in homework. After reviewing 59 studies related to parent involvement, they concluded that although research examining the effect of parent involvement in homework on achievement produced mixed findings, homework involvement seemed to support improved achievement, as well as student attention to homework, homework completion, homework performance, attitudes toward homework and school, perceptions of competence, and self-regulatory skills.
Three recent meta-analyses by Jeynes (2003, 2005, 2007) examined the relationship between numerous types of parent involvement and achievement, including the unique effect of parent involvement in homework. Jeynes’s 2003 meta-analysis of 21 studies examined the effect of parent involvement on minority children’s academic achievement. It revealed a significant positive effect of involvement in homework on achievement for African American students (β = .72; p. 211). Similarly, Jeynes’s 2007 meta-analysis of 52 studies examined the relationships between various forms of parent involvement and urban secondary school student academic achievement. It revealed that involvement in homework had a significant positive relationship with achievement for urban secondary school students, but only when variables such as socioeconomic status, gender, race, and prior achievement were not controlled for (β = .38 and β = .13, ns; pp. 95–96). However, a different pattern emerged in Jeynes’s 2005 meta-analysis of 41 studies assessing the relationship between parent involvement and urban elementary school student academic achievement. Specifically, a null relationship between parent involvement in homework and achievement was found (β = −.08; p. 254).
In sum, what we know about the effectiveness of parent involvement in homework from previous syntheses is rather limited. Cooper’s (1989) meta-analysis was based on very few studies and uncovered little causal evidence. Although Hoover-Dempsey et al. (2001) provided a useful description of studies that have addressed this question, it remains difficult to determine the impact of parent involvement in homework on the basis of the narrative approach of this review. Finally, Jeynes’s syntheses suggest that there may be positive effect of parent involvement in homework, but these meta-analyses are limited to particular populations and produced results somewhat inconsistent with each other. Taken together, the most basic question of whether parent involvement in homework has a positive effect on achievement remains unsettled. However, findings across these syntheses do suggest that involvement is effective under certain conditions for particular groups of individuals.
The Need for a New Synthesis of the Literature on Parent Involvement in Homework
In this article, we synthesize research on parent involvement in homework conducted since 1987 to address the following questions: (a) What is the overall effect of interventions designed to improve the frequency and quality of such parent involvement, (b) what is the overall relationship between achievement and parent involvement in homework found in studies that did not manipulate involvement, and (c) what moderators might explain variation in study findings. Although previous syntheses have provided useful information regarding the impact of parent involvement in homework, this synthesis attempts to address some of their limitations. Thus, there are four reasons for conducting a new synthesis of research on parent involvement in homework: (a) to apply quantitative techniques to all evidence accumulated since 1987 on the effects of parent involvement in homework and determine whether the conclusions of earlier research syntheses need modification; (b) to examine the literature with the explicit purpose of looking for evidence of a causal relationship, as opposed to a simple association; (c) to look for variables that might moderate the effect of parent involvement in homework and potentially explain contradictory findings; and (d) to apply new research synthesis techniques.
Regarding the exhaustiveness of studies covered in syntheses, although Jeynes’s meta-analyses provide recent updates on the relationship between parent involvement in homework and academic achievement, the primary purpose of these reviews was to examine the effect of a variety of different forms of parent involvement more broadly categorized (i.e., parental attendance and participation at school functions, communicating with one’s child, or having household rules), not just involvement in homework. Consequently, the examination of the effect of parent involvement in homework in particular was based on a small number of studies, all of which are included in the current synthesis. Further, Jeynes’s syntheses were restricted to studies with specific populations, specifically minority and urban students, so it is unclear whether the effects found would generalize to a broader population of students. Our synthesis examined the effects of parent involvement in homework by using the findings of a comprehensive set of studies.
Regarding causal relationships, with only two experimental studies related to the effect of parent involvement in homework on achievement, Cooper’s (1989) synthesis was not able to draw any conclusions concerning a causal relationship between parent involvement in homework and achievement. Also, Cooper’s synthesis revealed inconsistent correlations between parent involvement in homework and achievement. Neither Hoover-Dempsey et al. (2001) nor Jeynes (2003) formally distinguished between studies based on their research design. Although Jeynes (2005, 2007) distinguished between school programs meant to improve parent involvement and naturally occurring parent involvement in analyses collapsing various forms of involvement, a distinction between experimental and correlational data was not made for homework involvement in particular. In addition, recent studies have employed multivariate approaches to data analysis, including structural equation modeling, to test complex networks of influence on student achievement. Parent involvement in homework has been used as a variable in many of these models. The earlier syntheses did not include these designs as a distinct subtype. This synthesis has included these multivariate approaches and has conducted separate syntheses for each type of design.
Regarding the search for moderators of the relationship, the Cooper (1989) synthesis, as well as 2003 and 2005 syntheses conducted by Jeynes, did not examine moderator variables because of a limited number of studies. In the 2007 meta-analysis, Jeynes did examine whether the relationship between parent involvement in homework and achievement varied as a function of the type of achievement measure, though no other moderator variables were examined. This synthesis has extended the findings of previous reviews by using quantitative methods to shed light on the particular conditions under which parent involvement in homework is most effective. In particular, the literature has suggested that it may be important that we examine whether the relationship between parent involvement in homework and achievement varies as a function of the type of involvement that parents engage in. Further, Jeynes revealed a different pattern of findings for various populations. Specifically, parent involvement in homework had a significant relationship with achievement for African American students and urban secondary students but a null relationship for urban elementary school students. These results point to the possibility that parent involvement in homework has a different relationship with achievement for students with different characteristics. This meta-analysis will examine whether individual differences, such as age or socioeconomic status,1 may affect the involvement–achievement relationship. In addition, a number of other relevant variables were examined, including outcome measure, subject matter, and the type and length of an intervention.
Finally, the past two decades have introduced new techniques and refinements in the practice of research synthesis. Among others, these include (a) a greater understanding of meta-analytic error models (i.e., the use of fixed- and random-error assumptions), (b) new tests to estimate the impact of missing data on research synthesis findings, and (c) new techniques for performing data diagnostics and display. We use these techniques in the synthesis that follows.
Method
Literature Search Procedures
The first procedures used to locate studies on parent involvement in homework involved a computer search of the ERIC, PsycINFO, Sociological Abstracts, and Dissertation Abstracts electronic databases for documents catalogued between January 1, 1987, and December 31, 2004. The single keyword homework was used for all searches. In addition, Science Citation Index Expanded and the Social Sciences Citation Index databases were searched for documents catalogued between 1989 and 2004 that had cited Cooper (1989). These searches located a total of 4,400 nonduplicate, potentially relevant documents.
To supplement searches of electronic databases and obtain any research that might not be obtained through computer searches, the reference sections of relevant documents were examined for cited works that also might be relevant to the topic.
Three direct contact strategies also were employed to ensure that we tapped sources that might have access to homework-related research that would not be included in the reference and citation databases. First, we contacted the dean, associate dean, or chair of 77 colleges, schools, or departments of education at Research 1 institutions of higher education and requested that they ask their faculty to share with us any research they had conducted that related to homework. Second, we sent similar requests to 21 researchers who our reference database search revealed had published two or more articles on homework and academic achievement between 1987 and the end of 2003. Finally, we sent similar requests to the directors of research or evaluation in more than 100 school districts listed as members of the National Association of Test Directors.
The first and third authors then examined each title, abstract, or document. If either author felt that the document might contain data relevant to the relationship between parent involvement in homework and an achievement-related outcome, we obtained the full document (in those cases in which the judgment was made on the basis of the title or abstract).
Criteria for Including Studies
For a study to be included in the research synthesis, several criteria had to be met. Most obviously, the study had to have estimated in some way the relationship between a measure of parent involvement in homework and a measure of achievement or an achievement-related outcome.
Two sampling restrictions were placed on included studies. The study had to assess students in kindergarten through 12th grade. We excluded studies conducted on preschool-age children or postsecondary students because the purpose and causal structure underlying the parent involvement–achievement relationship would be very different for these populations. For similar reasons, the studies used in the synthesis were restricted to those conducted in the United States and Canada.
Three categories of design, some with subtypes, were included in the synthesis. First, studies could employ manipulations of parent training for involvement in homework and compare treatment and control conditions. This meant that parents in a treatment condition received some type of training geared toward encouraging or improving skills for parent involvement with homework expressly for purposes of the study. Parents and students in a control condition received no such intervention. Within such studies, the experimenters could introduce the manipulation (a) at either the student or classroom level and (b) by either randomly assigning parents to training and no-training conditions or by some nonrandom process. If a nonrandom process was used, the experimenter then might or might not employ a priori matching or post hoc statistical procedures to equate the training and no-training groups. If random assignment was not used, the variables used to enhance the equivalence of the groups could differ from study to study.
The second type of design included studies that took naturalistic, cross-sectional measures of the amount of involvement parents had in their children’s homework without intervention on the part of the researchers and related these to an achievement-related measure. This second type of design also included an attempt to statistically equate students on other variables that might be confounded with parent involvement in homework and therefore might account for the parent involvement–achievement relationship.
The third type of design involved the calculation of a simple bivariate correlation coefficient between the amount of parent involvement in homework and the measure of achievement. In these studies, no attempt was made to equate students on other variables that might be confounded with parent involvement in homework. Studies that utilized a one-group posttest-only or a one-group pre-posttest design were not included.
Finally, the report had to contain enough information to permit the calculation of an estimate of the relationship between involvement and achievement. Sometimes, these estimates were reported by the researchers; other times they were calculated from other reported statistics (see Rosenthal, 1994).
Information Retrieved From Studies
Numerous characteristics of each study were included in the database. These characteristics encompassed six broad distinctions among studies: (a) the research report, (b) the research design, (c) the characteristics of parent involvement in homework, (d) the characteristics of the sample of students, (e) the measure of achievement, and (f) the estimate of the relationship between parent involvement in homework and achievement. Table 2 presents the characteristics coded for each type of design.
Effect Size Estimation
For studies with manipulations of training for parent involvement in homework, we used the standardized mean difference, known as the d index (Cohen, 1988), to estimate the effect of parent training on measures of student achievement. Calculating the d index for any comparison involves dividing the difference between the two group means by either their average standard deviation or by the standard deviation of the control group. In the synthesis, we subtracted the no-parent-training condition mean from the parent-training condition mean and divided the difference by their average standard deviation. Thus, positive effect sizes indicate that students whose parents received training had better achievement-related outcomes than did students whose parents received no training. Effect sizes that were adjusted for preexisting individual differences were obtained by subtracting a pretest effect size from a posttest effect size (Shadish & Ragsdale, 1996).
For studies that involved naturalistic, cross-sectional measures of the amount of parent involvement in homework and related these to achievement but also included an attempt to statistically equate students on other variables, our preferred measure of relationship strength was the standardized beta weight, β. These parameters were derived from either the output of multiple regression analyses or from path coefficients in structural equation models. In a few instances, beta weights could not be obtained from study reports, so the most similar measures of effect (e.g., unstandardized regression weights) were retrieved. There were no instances in which we calculated beta weights from other statistics.
For studies that involved naturalistic, cross-sectional measures but included no attempt to statistically equate students on third variables, we used simple bivariate correlation coefficients, r, as measures of the direction and magnitude of the relationship. In some instances these were calculated from other inferential statistics.
Coder Reliability
Two graduate students extracted information from reports selected for inclusion. Discrepancies were noted and discussed by the coders, and if agreement was not reached, the second author was consulted. Because all studies were independently coded twice and all disagreements resolved by a third independent coder, we did not calculate a reliability for this process. Evidence suggests that the process used results in high reliability (Rosenthal, 1987).
Methods of Data Integration
Before conducting any statistical integration of the effect sizes, we first counted the number of positive and negative effects. Next, we calculated the range of estimated relationships. Also, we examined the distribution of sample sizes and effect sizes to determine whether any studies contained statistical outliers. Grubbs’s (1950) test was applied (see also Barnett & Lewis, 1994), and if outliers were identified, these values were set at the value of their next nearest neighbor.
Both published and unpublished studies were included in the synthesis. However, there still exists the possibility that we did not obtain all studies that have investigated the relationship between parent involvement in homework and achievement. Therefore, we employed Duval and Tweedie’s (2000a, 2000b) trim-and-fill procedure to test whether the distribution of effect sizes used in the analyses was consistent with that expected if the estimates were normally distributed.
Calculating Average Effect Sizes
We used a weighting procedure to calculate average effect sizes across all comparisons. In this procedure, each independent effect size was first multiplied by the inverse of its variance. The sum of these products was then divided by the sum of the inverses. Also, 95% confidence intervals (CIs) were calculated. If the CI did not contain zero, then the null hypothesis that parent involvement had no relation to the achievement-related outcome was rejected.
Identifying Independent Hypothesis Tests
One problem that arises in calculating effect sizes involves deciding what constitutes an independent estimate of effect. We used a shifting unit of analysis approach (Cooper, 1998). In this procedure, each effect size associated with a study is first coded as if it were an independent estimate of the relationship. For example, if a single sample permitted comparisons of the effect of parent involvement on both reading and math outcomes, two separate effect sizes were calculated. For estimating the overall effect of involvement, these two effect sizes were averaged prior to entry into the analysis so that the sample contributed only one effect size. However, in an analysis that examined the effect of involvement on reading and math separately, this sample would contribute one effect size to each estimate of a category mean effect size.
Statistical Integration
We conducted all our analyses twice, once employing fixed-error assumptions and once employing random-error assumptions. Through conducting these sensitivity analyses (Greenhouse & Iyengar, 1994), we could examine the effects of different assumptions on the outcomes of the synthesis. The interested reader should refer to Hedges and Vevea (1998) for a discussion of fixed and random effects. Possible moderators of the parent involvement–achievement relationship were tested via homogeneity analyses (Cooper & Hedges, 1994; Hedges & Olkin, 1985). All statistical analyses were conducted with the Comprehensive Meta-Analysis statistical software package (Borenstein, Hedges, Higgins, & Rothstein, 2005).
Results
The strength of the causal inferences that can be drawn from a study will vary with the study’s design. Therefore, each category of design described in the Method section is discussed separately in the presentation of the results. We begin with studies that manipulated parent homework involvement by providing training. These studies include (a) randomized experiments with the student as the unit of assignment and analysis, (b) randomized experiments with the classroom as the unit of assignment but the student as the unit of analysis, and (c) quasi-experimental studies. For each design, the effect of training on achievement is discussed, as well as a number of achievement-related outcomes. In addition, for randomized experiments with the student as the unit of assignment and analysis, the question of whether a number of variables moderated the effect of parent training on achievement is examined. Next, the pattern of results is examined for cross-sectional studies in which third variables were controlled by either multiple regression or structural equation modeling. This analysis includes studies using data from the National Educational Longitudinal Study (1988, 1990, or 1992). Finally, the association between parent involvement in homework and achievement is examined in studies in which the simple bivariate correlation was calculated.
Studies With Manipulations of Parent Training for Homework Involvement
The literature search located 14 studies that employed a procedure in which parent training for homework involvement was made available to families explicitly for the purpose of studying training effects. Of these studies, 11 were unpublished; 6 used random assignment of individual families to conditions, 3 employed random assignment of classrooms to conditions, and the remaining 5 studies employed a quasi-experimental design without random assignment. The characteristics of these studies are listed in Tables 3, 4, and 5.
Randomized Experiments With Student or Parent as the Unit of Assignment and Analysis
The six studies using random assignment of individual families to conditions appeared between 1992 and 2003. Five of the six studies were unpublished dissertations (Doering, 1993; Kiesner, 1997; Meteyer, 1998; Nesbitt, 1993; Tamayo, 1992) and one study was a journal article (Toney, Kelley, & Lanclos, 2003). Taken together, five of the six studies that provided parent training for homework involvement and randomly assigned participants to conditions tested the effect of parent training on achievement. Four of these studies also tested the effect of parent training on homework completion rates, and three tested the effect on homework problems, defined as the frequency with which the student demonstrated problematic behavior that might result in reduced completion or quality of the homework, or the frequency with which students experienced complications with homework. Examples of homework problems included the frequency with which students forgot to bring assignments home, refused to do homework, denied having homework, argued with their parents about homework, complained about homework, were frustrated about homework, rushed through homework, or were sent to the office at school for homework behavior. Finally, just one study revealed a negative effect for attitudes toward the academic subject.
Achievement
Table 6 presents the results of the analyses of the effect sizes with achievement and related outcomes as the dependent variable. There were no significant outliers among the effect sizes, so all were retained for analysis as reported. The unadjusted effects of parent training for homework involvement on achievement varied between d = −.40 and .78. Using fixed-error assumptions, the weighted mean d index was .11 but was not significantly different from zero (95% CI = −.06/.27). With a random-error model, the weighted average d index was .09 (95% CI = −.16/.34). The tests of the distribution of d indexes revealed that we could not reject the hypothesis that the effects were estimating the same underlying population value, Q(4) = 8.24, p = .08.
We looked at the overall effect of parent training for homework involvement on achievement excluding two studies, Doering (1993) and Nesbitt (1993), in which the training intervention was less specifically geared toward teaching parents to be involved with their children’s homework. Without those two studies, the effect of parent training on achievement was smaller and still not significantly different from zero under either fixed- or random-effect assumptions.
Next, trim-and-fill analyses were conducted in several different ways. We performed the analyses looking for asymmetry with both fixed- and random-error models (see Borenstein et al., 2005) while searching for possible missing effect sizes on the left side of the distribution (those that would reduce the size of the positive effect). We conducted these analyses both with Doering (1993) and Nesbitt (1993) retained and with these two studies excluded. For all analyses, under both the fixed- or random-effects models, no additional effect sizes were imputed.
Achievement moderator analyses
Despite the apparent lack of heterogeneity among the study results, we decided to examine whether the magnitude of the effect of parent training for homework involvement on achievement was moderated by the type of achievement measure, grade level of the students, the subject matter of the outcome, the method of intervention used to train parents, or the number of workshops or individual meetings in which training took place. Table 7 presents the results of these analyses.
We found that only grade level significantly moderated the effect of parent training for homework involvement on achievement under either fixed or random-error assumptions. Effect sizes were grouped into those involving elementary school students, Grades 2 through 5, and middle school students, Grades 6 through 8. None of the studies included tested the effect of parent training on achievement in high school students. Under fixed-error assumptions, the weighted mean d index was significantly higher for elementary school students, d = .22 (95% CI = .02/.42), than for middle school students, d = −.18 (95% CI = −.49/.14), Q(1) = 4.28, p < .05. As indicated by the CIs, with the fixed-error model, the effect of parent training for achievement was significantly different from zero for elementary school students but not significantly different from zero for middle school students. Under random-error assumptions, there was no significant difference between elementary and middle school students, Q(1) = 3.39, p = .07.
Homework completion rate
The unadjusted effects of parent training for homework involvement on homework completion rate varied between d = −.13 and d = .69. Using fixed-error assumptions, the weighted mean d index was .28 and was significantly different from zero (95% CI = .00/.55), indicating that students whose parents were trained to be involved in homework completed more homework than did students whose parents did not receive training. Using a random-error model, the weighted average d index was also .28 (95% CI = −.01/.56) but fell just short of statistical significance, p < .06. In addition, the tests of the distribution of d indexes revealed that we could not reject the hypothesis that the effects were estimating the same underlying population value, Q(3) = 3.09, p = .38.
Frequency of homework problems
The unadjusted effects of parent training for homework involvement on the frequency of homework problems varied between d = −2.51 and d = −.03. With fixed-error assumptions, the weighted mean d index was −.84 and was significantly different from zero (95% CI = −1.10/−.57), indicating that students whose parents were trained to be involved in homework had fewer problems with homework than did students whose parents did not receive training. With a random-error model, the weighted average d index was −1.20 (95% CI = −2.33/−.06) and was significantly different from zero.
Attitudes
Just one study tested the effect of parent training for homework involvement on attitudes toward academics (Doering, 1993). It revealed a significant negative effect of involvement training, d = −.49 (95% CI = −.92/−.06), indicating poorer attitudes toward homework for students whose parents received training.
Randomized Experiments With Classroom as Unit of Assignment and Student or Parent as Unit of Analysis
The three studies that used random assignment of classrooms to condition but analyzed data with the student as the unit of analysis appeared between 1995 and 2004. Two studies (Albright, 2004; Rillero & Helgeson, 1995) were unpublished, and one was a journal article (Balli et al., 1998). The unadjusted effects of parent training for homework involvement on achievement varied between d = −.37 and d = .65. With fixed-error assumptions, the weighted mean d index was .00 (95% CI = −.16/.16). With a random-error model, the weighted average d index was .01 (95% CI = −.24/.26). Thus, even though these studies used the incorrect unit for statistical analysis, one that would improve the chance of finding a significant effect, a nonsignificant effect on achievement outcomes was found. The tests of the distribution of d indexes revealed that we could not reject the hypothesis that the effects were estimating the same underlying population value, Q(2) = 3.73, p = .16. The small number of studies precluded their use in any formal analyses investigating possible influences on the effect of the parent training for homework involvement. Only one study tested the effect of parent training for both homework completion (Albright, 2004) and attitude toward the academic subject (Rillero & Helgeson, 1995). The effect on both homework completion rate, d = −.01 (95% CI = −.46/.45), and attitudes, d = −.11 (95% CI = −.39/.16), was negative.
Quasi-Experimental Studies Without Random Assignment
The five quasi-experimental studies appeared between 1987 and 2003. Four of the five studies were unpublished (Austin, 1988; Pigford, 1987; Warrick, 2000; White, 1996), and one study was a journal article (Van Voorhis, 2003). With fixed-error assumptions, the weighted mean d index adjusted for individual differences including prior achievement was .23 and was significantly different from zero (95% CI = .14/.32). With a random-error model, the weighted average d index was .22 (95% CI = .01/.43). The tests of the distribution of d indexes revealed that we could reject the hypothesis that the effects were estimating the same underlying population value, Q(4) = 19.86, p < .001. However, these results need to be interpreted with caution because of a number of methodological considerations. Namely, a variety of methods were used to equate groups across studies, and all but one study involved a training intervention that was introduced at the classroom level but then analyzed data with the student as the unit of analysis. Also, although the standard formula for the d index with means and a pooled standard deviation was often used to calculate the effect size, the variety of information provided in this set of studies required the use of an assortment of formulas to obtain an estimate of the effect of parent training. Furthermore, the studies varied along multiple dimensions, and few studies were contained in each combination of multiple design features. Because of the small number of studies and their variety of methods and contexts, we decided that moderator analyses would not be conducted on this set of studies.
Just one study tested the effect of parent training on homework completion (Van Voorhis, 2003). It revealed a null effect, d = .00 (95% CI = −.25/.25).
Summary
With regard to achievement, findings from studies with experimental manipulations provide fairly consistent evidence that the effect of training parents for homework involvement has at best a slightly positive overall impact on achievement. Although studies in which parent training was randomly assigned at either the student or classroom level did not produce a significant effect of parent training, quasi-experimental studies demonstrated a significant positive effect of parent training on achievement. Results of experiments using random assignment at the student level suggest that parent training had a significant positive impact on homework completion rates and the frequency of problems with homework, although an individual experiment using random assignment at the classroom level and an individual quasi-experiment did not support this finding for homework completion. Finally, moderator analyses of randomized experiments may explain the overall null finding for achievement, suggesting that training parents to be involved in homework may result in a significant positive effect on achievement for elementary school students but not for middle school students.
Studies With Cross-Sectional Data and Control of Third Variables
Studies Using Data From the National Educational Longitudinal Study (1988, 1990, or 1992)
The literature search located three reports that contained multivariate analyses of data collected as part of the National Education Longitudinal Study of 1988 (NELS) or one of the NELS follow-ups on the same students in 1990, 1992, 1994, or 2000. These studies are described in Table 8. All studies using the NELS data sampled students from the NELS itself. Muller (1995) was interested in how parent involvement mediated the relationship between maternal employment and math achievement. Peng and Wright (1994) were interested in studying differences in relationships between predictors of achievement across ethnic groups, with a focus on Asian Americans. Simon (2000) restricted her sample to students who participated at all time points and had parent and administrator surveys completed as well.
Examined together, the three studies and four models using NELS sample data employed a wide variety of outcome measure configurations and different sets of predictor variables. Still, every regression coefficient estimating the relationship between parent involvement in homework and achievement was negative, and 7 of these 10 negative coefficients were statistically different from zero. Furthermore, Simon (2000) also examined the effect of parent involvement in homework on student conduct, finding a positive relationship between both parent rules about and help with homework and the frequency of problem behavior.
Studies Using Data Other Than the National Educational Longitudinal Study and Performing Multivariate Analyses
We found 17 additional studies, with 20 independent samples, that performed multivariate analyses on cross-sectional data to examine the relationship between homework and achievement, with other variables controlled. The characteristics of these studies are described in Table 9. Shumow and Miller (2001) used the Longitudinal Study of American Youth database containing a national probability sample of approximately 6,000 seventh and eighth graders in 1987, when the study began, stratified by geographic area and degree of urban development. Zick, Bryant, and Osterbacka (2001) used the National Survey of Families and Households database, containing a sample of more than 13,000 households and more than 5,600 households with one or more children under the age of 18. The rest of the studies used data collected by the researchers for the specific purpose of studying variables related to achievement.
Of the 20 samples, 10 reported a positive relationship between parent involvement in homework and achievement across all types of involvement and achievement measures, 8 reported negative relationships, and 2 reported mixed results that varied depending on the type of involvement or achievement measure examined. Two studies (Cooper et al., 2000; Van Voorhis, 2003) examined the link between parent involvement in homework on homework completion rates, and both found a positive relationship. Two of the five estimated relationships between involvement and completion were significantly different from zero. Finally, 3 studies examined the effect of involvement on student attitudes toward school and homework. Two studies found a positive relationship (Cooper et al., 2000; Shumow & Miller, 2001), 1 of which reported a beta coefficient significantly different from zero (Shumow & Miller, 2001), and 1 study (Strauss, 2000) found that parent involvement was negatively related to school enjoyment and school-related anxiety, although neither relationship was significantly different from zero.
If we omit those studies that reported nonstandardized regression weights or another product of regression analyses (such as semipartial or partial correlations; Anderson, 2000; Cooper et al., 2000; Pezdek, Berry, & Renno, 2002; Sheldon & Epstein, 2005), then the reported beta weights for the relation between parent involvement in homework and achievement range from −.33 to .34. We examined the range of effect sizes for different achievement outcomes and types of parent involvement. No clear pattern of moderation emerged. Finally, it is important to note that consistent with randomized experiments, the two studies that reported results involving homework completion rates both revealed positive results.
Structural Equation Modeling Studies
Three studies used structural equation models to test the relationship between parent involvement in homework and achievement. The characteristics of those studies are described in Table 10. All three studies used original data collected by the researchers. Two studies (Cooper, Jackson, Nye, & Lindsay, 2001; Pomerantz & Eaton, 2001) revealed a positive relationship between parent involvement in homework and achievement, and one (Schultz, 1999) found a negative relationship.
Summary
Studies using cross-sectional data and control of third variables found inconsistent effects of parent involvement in homework, with a fairly even split between those studies producing positive effects of involvement and those producing negative or null effects. In addition, findings of cross-sectional studies controlling for third variables provided evidence consistent with studies in which parent training for involvement was manipulated; the effect of involvement on homework completion was positive in those studies in which homework completion rates were examined.
Studies Correlating Time on Homework and Academic Achievement
The literature search uncovered 20 studies that estimated the correlation between parent involvement in homework, as reported by either the student or a parent, and a measure of academic achievement or an achievement-related measure. The 20 studies reported 63 separate correlations based on 22 separate samples of students and their parents. Of those correlations, 58 measured achievement, 2 measured homework behaviors, 2 measured student attitudes toward homework, and 1 measured the frequency of problem behavior. The characteristics of these studies are listed in Table 11.
The 20 studies appeared between the years 1987 and 2003. The sample sizes ranged from 20 to approximately 9,685, with a median size of 179.5. The mean sample size was 704, with a standard deviation of 2,027, suggesting a nonnormal distribution. The Grubbs test revealed a significant outlier, p < .05. This sample was the largest in the data set, reported by Peng and Wright (1994) for one correlation obtained from the NELS study. As a result, we replaced this sample size with the next largest sample size in the data set, 1,021. The mean sample size for the adjusted data set was 310, with a standard deviation of 329. There were no significant outliers among the correlations, so all were retained for analysis as reported.
Analysis of All Correlations
The weighted average correlation was r = .04 (95% CI = .02/.05) with a fixed-error model and r = .07 (95% CI = −.08/.21) with a random-error model. As revealed by the CIs, the hypothesis that the relationship between parent involvement in homework and achievement is r = 0 can be rejected under the fixed-error model but not the random-error model. In addition, the tests of the distribution of correlations revealed that we could reject the hypothesis that the correlations were estimating the same underlying population value, Q(21) = 1418.79, p < .0001.
Trim-and-fill analyses were conducted in the same fashion as for experimental studies. With a fixed-effects model, there was evidence that eight effect sizes might have been missing. Imputing these values would change the mean correlation to r = −.12 (95% CI = −.13/−.10). Under the random-effects model, no additional correlations were imputed.
Next, we carried out a moderator analysis examining the association between the magnitude of correlations and the publication status of the study report. Eleven of the samples had been published, and their results were compared to the 11 samples that had appeared in dissertations, conference papers, or unpublished research reports. Under the fixed-error model, correlations from published reports, r = −.08 (95% CI = −.10/− .05), were significantly different from those from unpublished sources, r = .12 (95% CI = .10/.15), Q(1) = 140.35, p < .0001. Under the random-error model, there was no difference between published and unpublished reports, Q(1) = 3.32, p = .07.
Moderator Analyses
Next, we conducted moderator analyses of the relationship between parent involvement in homework and academic achievement using six moderators: outcome measure, grade level, type of parent involvement, subject matter, respondent, and socioeconomic status. Information was collected on several additional variables, including at-risk status, student ability, gender, and ethnicity. However, because of a lack of reporting or a lack of variation across categories, no analyses were conducted on these variables.
All six moderator analyses were significant under fixed-effects assumptions, but only three moderators remained significant when a random-effects model was implemented. Here we present only these three most robust moderators: grade level, type of parent involvement, and subject matter. Table 12 presents the results of analyses examining all seven different moderators.
Grade level
Correlations were grouped into those involving students in Grades 1 through 6, Grades 7 and 8, and Grades 9 through 12. One study (Smock & McCormick, 1995) was omitted from the analysis because it included students in Grades K through 12. Xu and Corno (2003) was also omitted from the analysis because it included students in Grades 6 through 8 and the correlation for the sixth graders could not be separated from that for the seventh and eighth graders.
The overall moderator test revealed that the effect of parent involvement on achievement varied by grade level under both fixed-effect, Q(2) = 106.00, p < .0001, and random-effect assumptions, Q(2) = 20.65, p < .0001. We then proceeded to conduct pairwise comparisons. Using fixed-error assumptions, the significant average weighted correlation for elementary school students, r = .06 (95% CI = .04/.08), was different from the significant average correlation for middle school students, r = −.17 (95% CI = −.21/−.13), Q(1) = 88.26, p < .0001. Using random-error assumptions, the average correlation for elementary students was not significant, r = .05 (95% CI = −.16/.25), but remained significantly different from the average correlation for middle school students, r = −.17 (95% CI = −.26/−.09), Q(1) = 3.80, p < .05.
Using fixed-error assumptions, the average correlation for elementary school students was significantly different from the significant average correlation for high school students, r = .17 (95% CI = .11/.24), Q(1) = 10.30, p < .001. Under the random-error assumptions, the average correlation for high school students was also significant, r = .22 (95% CI = .07/.36). However, there was no difference between elementary and high school students in the effect of parent involvement under random-effects assumptions, Q(1) = 1.80, ns.
Finally, the average correlation for middle school students was significantly different from that for high school students under fixed-effect assumptions, Q(1) = 71.35, p < .0001, and under random-effects assumptions, Q(1) = 19.39, p < .001.
Type of parent involvement
Next, we examined whether the correlation between parent involvement in homework and achievement was moderated by the type of involvement. The correlations were grouped into those involving parents’ monitoring of homework, setting and enforcing rules about homework, and providing direct homework aid and those measuring the relationship between achievement and the use of all three types of involvement. Monitoring homework involved checking that homework was completed or supervising so that children remained on task. Setting rules about homework involved deciding when and where children were to complete homework. Direct homework aid involved the parent’s giving feedback to the student on homework accuracy or tutoring and giving instruction about the content of the homework. Seven studies and one correlation from each of two studies (Broxie, 1987; Clark, 1993) were omitted from the analysis because details about the type of involvement parents had in homework were not specified. Finally, one correlation from Broxie (1987), between providing homework resources and achievement, and one correlation from Clark (1993), between parents’ talking to children about homework and achievement, were omitted because these were the only studies in which each of these types of involvement were examined.
The overall moderator analysis revealed that the effect of parent involvement varied with the type of involvement under both fixed-error assumptions, Q(3) = 754.56, p < .0001, and under random-error assumptions, Q(3) = 12.86, p < .01. We then proceeded to conduct pairwise comparisons. Under fixed-error assumptions, the significant average correlation for parent monitoring, r = −.09 (95% CI = −.14/−.03), was significantly different from the significant average correlation for rule setting, r = .54 (95% CI = .51/.57), Q(1) = 393.07, p < .0001. Under random-error assumptions, there was no difference between monitoring and rule setting, Q(1) = .57, ns.
Using fixed-error assumptions, the significant average correlation for direct aid, r = .10 (95% CI = .07/.13), was significantly different from that for monitoring, Q(1) = 33.81, p < .0001. Under the random-error assumptions, there was no difference between direct aid and monitoring, Q(1) = .02, ns. Finally, the average correlation for rule setting was significantly different from that for direct aid under fixed-effect assumptions, Q(1) = 331.09, p < .0001, but not under-random effects assumptions, Q(1) = .49, ns.
Subject matter
Eleven studies were omitted from this analysis because they involved multiple subjects, and correlations for each individual subject could not be distinguished. Thus, the moderator analysis compared only studies involving language arts, reading, and mathematics achievement.
First, we compared correlations involving language arts with correlations involving reading. Under fixed-error assumptions, the significant average weighted correlation for language arts, r = .12 (95% CI = .05/.20) was significantly different from the significant average correlation for reading outcomes, r = .20 (95% CI = .18/.23), Q(1) = 3.88, p < .05. Under random-error assumptions, the effect of parent involvement was not different for reading compared with language arts subjects, Q(1) = .002, ns. Because of these results, we chose not to combine the language arts and reading data sets.
The overall moderator analysis revealed that the effect of parent involvement varied by subject matter under fixed-error assumptions, Q(2) = 217.03, and under random-error assumptions, Q(2) = 52.14, p < .0001. We then proceeded to conduct pairwise comparisons.
Under fixed-error assumptions, the correlation between parent involvement in homework and achievement was significantly higher for reading, r = .20 (95% CI = .18/.23), than for math, r = −.19 (95% CI = −.24/−.15), Q(1) = 216.82, p < .0001. As indicated by the CIs, both were significantly different from zero under fixed-error assumptions. Under random-error assumptions, there was not a significant difference in the effect of parent involvement for reading compared with math outcomes, Q(1) = 2.70, p < .10, and only the correlation for math outcomes remained significantly different from zero.
The correlation between parent involvement in homework and achievement was significantly higher for language arts than for math under both fixed-error assumptions, Q(1) = 50.69, p < .0001, and random-error assumptions, Q(1) = 50.69, p < .0001.
Nonachievement Measures
Two studies provided estimates of the relationship between parent involvement and three other achievement-related outcomes: homework behaviors, conduct, and student attitudes toward homework. Characteristics of these studies can be found in Table 11. Epstein (1988) found a significant negative relationship between minutes parents spent helping with homework and teacher nominations of students as “homework stars,” r = −.10, and a nonsignificant positive correlation with teacher nominations of students who have problems with homework, r = .07. Epstein (1988) also found a significant positive correlation, r = .13, between parent involvement in homework and teacher reports of student discipline problems at school. Cooper et al. (2001) found a nonsignificant negative relationship between parent homework rule setting and student attitudes toward homework, r = −.08; however, the relationship between parent-provided direct homework aid and student attitudes toward homework was significant and positive, r = .23.
Summary
As with the results from other study designs, studies correlating time on homework and academic achievement suggested that the positive effect of involvement was small, if it was different from zero at all. However, also consistent with other study designs was the finding that the relationship between involvement in homework and achievement was positive and significant for elementary school students, as well as high school students, but negative for middle school students. Finally, results from moderator analyses of correlational studies pointed to several other potentially important moderators, including type of homework involvement (the establishment of homework rules produced the strongest relationship with achievement) and subject matter (homework involvement in verbal subjects showed a positive relationship with achievement, and mathematics involvement showed a negative relationship).
Discussion
In this synthesis, we meta-analyzed research examining the relationship between parent involvement in homework and achievement across three discreet study designs: manipulations of parent training for homework involvement, cross-sectional data collections in which third variables were controlled, and cross-sectional data collections without controlling for third variables. Although the results varied slightly across each group of studies, several consistent findings emerged. First, consistent with the conclusions of previous reviews (Cooper, 1989; Hoover-Dempsey et al., 2001; Jeynes, 2005), across all designs, the overall effect of parent involvement in homework was small and often not significant. However, evidence also suggested that this small overall relationship between involvement and achievement may be qualified by the fact that homework involvement was not equal across all circumstances. Specifically, both causal and correlational evidence suggested that the effect of parent involvement varied with the student’s age. Further, correlational evidence suggested that other variables may be important moderators, including the type of homework involvement provided by parents and the subject matter of the assignments. Further, causal evidence suggested that homework involvement may have an effect on the most proximal achievement-related outcomes, including homework completion rates and the frequency of homework problems, both of which may lead to gains in achievement over the long term.
Summary of Studies on Causal Relationships
Studies that randomly assigned parents to training or no-training conditions provided the strongest evidence from which to infer the causal effects of parent involvement in homework. The results of these randomized experiments suggest that parent training had a significant positive impact on (a) homework completion rates and (b) the frequency of problems with homework. With Cohen’s (1988) measure of distribution overlap (U3) as a means to express the size of these effects, the average student who had a parent trained to be involved with homework had a higher homework completion rate than did about 61% of students without a trained parent. Students of trained parents also had fewer homework problems, such as refusing to do homework, being frustrated by homework, complaining about homework, or being sent to the office for poor homework behavior. Even more striking than the effect of training for involvement on completion, the average student whose parents were trained to be involved with homework had fewer homework problems than did about 80% of students whose parents were not trained. Further, studies using multivariate techniques to control for confounded variables provided additional support for the positive impact of parent involvement on homework completion rates. Five regression coefficients derived from two studies were all positive, and two reached statistical significance. These results are consistent with conclusions drawn from a previous review of the relationship between parent involvement in homework and achievement from Hoover-Dempsey et al. (2001), in which it was suggested that involvement may have a greater effect on proximal measures of achievement.
Across training programs, parents were trained to be involved in homework through a variety of involvement strategies. However, because of the small number of studies that examined homework completion rate and homework problems, we could not conduct moderator analyses for these outcomes. Consequently, it remains difficult to determine what specific characteristics of training programs led to greater homework completion rates and fewer homework problems. Still, it is interesting to note that the training program that primarily emphasized monitoring homework completion was the only one to have a negative impact on homework completion rates (Tamayo, 1992). Other programs included an emphasis on improving the learning environment, helping students improve homework habits, and supervising the homework process (Doering, 1993; Kiesner, 1997; Meteyer, 1998). This pattern is similar to that found for correlational evidence, in which monitoring was the only form of involvement to have a negative relationship with achievement, whereas all other forms had a positive relationship, and in particular, setting rules about homework had the largest association with achievement.
With regard to direct measures of achievement, findings from studies that randomly assigned parents to training or no-training conditions suggested little or no positive effect of training. Findings from all manipulated study designs provided fairly consistent evidence that the effect of training parents for homework involvement has at best a slightly positive overall impact on achievement. Quasi-experimental studies demonstrated the greatest positive effect of parent training on achievement; in fact, this was the only set of studies in which parent training was manipulated to show a significant overall effect. However, methodological flaws render this set of studies less trustworthy than the studies described earlier. Similarly, the range of estimated coefficients derived from studies using multiple regression, path analysis, or structural equation modeling revealed inconsistency in the direction of their links involving parent involvement.
There is some encouraging news with regard to achievement measures, however. When we examined subgroups of students, the randomized experiments did reveal a reliable positive effect of parent training in homework involvement on the achievement outcomes for students in elementary school, Grades 2 through 5. The size of the effect for elementary school students suggests that the average elementary school student whose parents were trained to be involved in homework performed better than about 59% of students whose parents were not trained. This effect was similar in magnitude to the effect found when random-error assumptions were used, but in this analysis the effect was not significantly different from zero. Thus, the generality of any inference regarding this positive causal effect is restricted to experiments similar to those included in the meta-analysis.
Summary of Studies on Correlations
The strength of the overall correlation between parent involvement in homework and achievement was very small and significant only under fixed effects. However, it would be ill advised to interpret this generally weak overall relationship as indicating no causal effect of parent involvement on achievement. Rather, it seems plausible to assume that the simple correlation might be conflating two causal relationships with opposite effects. That is, generally speaking, the experimental results suggest a possible positive effect of appropriate parent involvement on homework achievement for young students. However, the simple correlation would also be influenced—but in an opposite direction—if poor achievement caused more parent involvement in homework, either through self-initiated behavior by the parent or through teacher requests. This conflation of effects illustrates why the interpretation of correlations as indicators of causal effects can be misleading.
Where simple associational studies may hold the most value is in the tests of variables that moderate the size of correlations. The moderator tests on correlations (Table 12) revealed three robust moderators—the grade level of the student, the type of parent involvement, and the subject matter—and these analyses contained five of the six subgroup correlations that also proved different from zero under both fixed- and random-effect assumptions. We will limit our speculations to only these three moderators that permitted the most robust generalization.
Grade Level
Consistent evidence from both experimental and correlational studies suggests that parent involvement in homework had desirable effects for elementary school students. Because younger students appear to have less developed study habits (Dufresne & Kobasigawa, 1989), parent involvement may serve as an opportunity not only to learn academic content but also to internalize and develop study skills and effective forms of self-management modeled by parents. In addition, involvement in homework may be effective for achievement in elementary school in part because parents have greater mastery of the subject matter covered in the early grades (Cooper, 2001).
The analyses using both correlational and experimental studies showed that middle school students generally may not benefit from parent involvement in homework. Parents may need to be aware of the developmental stage of their child when engaging in involvement in homework and adjust their type of involvement accordingly. For example, adolescence is a time when young people attempt to develop some level of independence and autonomy from their parents in many domains (Erikson, 1968; Gutman & Midgley, 2000; Hill & Holmbeck, 1986). Further, the transition to middle school is an especially difficult period for many children (Carnegie Council on Adolescent Development, 1995; Gutman & Midgley, 2000), during which an increase in parent–child conflict often occurs (Laursen, Coy, & Collins, 1998). Thus, although providing guidelines for homework behavior or providing direct help with homework may be an effective form of involvement for elementary students, as students reach adolescence, it may be important that parents gradually withdraw from the homework process and shift their involvement more to support of the child’s own autonomous efforts.
It is also important again to consider the possibility that the negative correlation in middle school is actually caused by a strong achievement-to-involvement link. Evidence suggests that many children experience a general decline in school performance during the transition to middle school (Simmons & Blyth, 1987; see also Eccles et al., 1993, for review). Consequently, the observed negative association may be due to a decline in achievement among middle school students causing more parent involvement, differentially for those students experiencing the greatest decline.
One surprising finding was the significant positive correlation between parent involvement in homework and achievement for high school students. We might speculate that parent involvement during the high school years may be effective because parents’ provision of assistance becomes highly specialized. That is, parents may become directly involved in homework only when they have particular expertise to share with their child. Thus, although homework involvement occurs less frequently as students grow older (Eccles & Harold, 1996; Epstein & Lee, 1995; Hoover-Dempsey & Sandler, 1997; Stevenson & Baker, 1987), it is likely to be more effective when it does occur. This pattern of results was partially inconsistent with previous reviews by Jeynes (2005, 2007), in which parent involvement in homework had no relationship with achievement among urban elementary school students but had a significant positive relationship with achievement among urban secondary school students. The inconsistency between the results of this meta-analysis and those of Jeynes may be explained by the fact that Jeynes included only studies using urban student populations and combined middle school and high school students in his analyses of secondary students, whereas this meta-analysis distinguished between middle school and high school students.
Type of Parent Involvement
Consistent with hypotheses were the findings that although the overall effect of parental homework involvement on achievement was generally small or nonexistent, various forms of involvement were differentially effective. Setting rules about when and where homework should be done had the strongest positive relationship with achievement. Setting rules entails clearly communicating expectations, providing guidelines, and reinforcing behavior when rules are followed. Consequently, this strategy may be a particularly effective way to increase the time students attend to homework tasks, the effectiveness of how that time is used, or both. Further, this involvement strategy may have long-term achievement benefits to the extent that homework rules may become internalized by the child over time and help the student develop self-regulation skills. Although this finding is based on correlational evidence, it seems implausible that the student’s achievement would be causing this type of parent involvement, that is, higher academic achievement causing parents to set homework rules. Of course, it is still possible that the relationship is spurious—perhaps, for example, as family socioeconomic status goes up, so does the likelihood both that parents set rules and that children achieve better in school.
The other two specific forms of parent involvement produced significant relationships with achievement (under fixed-effect assumptions) in opposite directions. Parent monitoring of homework was negatively related to achievement whereas direct aid was positively related. In line with the time-on-task perspective, parents’ provision of instruction may make homework study more effective by facilitating greater understanding. In contrast, monitoring strategies used in isolation, without clear guidelines, reinforcement, or instruction during which desired behaviors can be modeled, may be experienced as controlling rather than informative.
Subject Matter
Finally, in experimental studies, parent involvement in homework had positive effects on both mathematics and verbal subject matter (though nonsignificantly so), and in correlational studies, involvement had a positive relationship with achievement in verbal subject matter but a negative relationship with achievement in mathematics. Although this latter relation is again causally ambiguous, it serves to highlight the potential importance of considering limitations in parent mentoring skills. Inadequate helping skills and use of instructional strategies that conflict with those the teacher uses in class might attenuate the relationship between parent involvement in homework and achievement. That is, involvement in mathematics homework may be more difficult for parents because it is less likely that they have received recent exposure to material covered in mathematics homework and therefore may encounter difficulty when attempting to assist their children with math homework. Second, the greater variety of instructional strategies used to teach mathematics may provide for more opportunities for parents to interfere with learning when the instructional techniques parents use differ from those being used at school. Reading and language arts, on the other hand, may require less specific subject matter skill. Still, we must also entertain the possibility of a reverse causal process to explain the pattern of results among correlational studies: Children who struggle more with mathematics may seek the intervention of their parents (or parents may be asked by teachers to intervene) more often than occurs in language-related skill areas.
Limitations to Generalizability
Among experimental studies manipulating parent training for homework involvement, analyses testing for the effect of publication bias and data censoring revealed little evidence to suggest that the strategies we used to locate studies for the synthesis produced a biased representation of results. Still, other limitations to the generalizability of the synthesis findings need to be made clear.
First, the duration of interventions meant to enhance the quality and frequency of parent involvement in homework was generally very brief, and the effects of such interventions have been tested preponderantly on measures of immediate achievement outcomes. Therefore, it is not possible to make claims about the effect of parent training during an extended time or on long-term achievement. Similarly, it is not possible to make claims regarding the long-term effect of involvement on other achievement-related outcomes, including homework completion rates and frequency of homework problems. However, the studies using naturally occurring parent involvement did use longer time frames, and these were consistent with the experimental findings for the general outcomes that involvement improved completion rates and reduced problems. It is important that future research create interventions that persist for an extended time and provide multiple opportunities to instruct and support parents so that training fidelity is high and sustained.
With regard to subject matter, our database contained too few correlations involving subjects other than math, reading, and language arts, such as science and social studies, to include these in the meta-analysis. Therefore, the effect of parent involvement remains unknown for these subject areas. In addition, few studies examined the effect of parent involvement on achievement with homework grades or nonstandardized (unit) test scores as the outcome measure. This is particularly problematic because we would expect that involvement might have its greatest impact on these more proximal measures of achievement rather than with class grades or standardized test scores. Consequently, our understanding is limited about the effects of involvement and how they might spread from specific and proximal measures to general and distal measures.
Further, few studies exist that examine the effectiveness of parent involvement for students of various cultural backgrounds, socioeconomic statuses, or levels of ability. Most studies failed to report these characteristics of their sample, and those that examined broad samples of students rarely looked for moderating effects of student individual differences. However, results from previous meta-analyses (Jeynes, 2003, 2005, 2007) suggest that parent involvement may be particularly effective for minority students. Research should continue to explore this possibility.
Finally, it is also important to keep in mind that synthesis-generated evidence should not be misinterpreted as supporting statements about causality (see Cooper, 1998). When groups of effect sizes are compared within a research synthesis, regardless of whether they come from simple correlational analyses or from experiments using random assignment, the synthesis can only establish an association between a moderator variable and the outcomes of studies, not a causal connection. Thus, when different study characteristics are found to be associated with the effects of an intervention or the size of a correlation, these findings should be used to direct future research to examine these factors by means of a more controlled design, so that their causal impact can be appraised.
Future Research
Most important, we want to emphasize that the continued gathering of correlational estimates of the involvement–achievement relationship is of limited value. Because the likelihood exists that both causal directions exist in nature, such studies will only add more ambiguity to an already ambiguous database. Instead, more useful will be (a) carefully conducted ethnographic studies that provide detailed longitudinal descriptions of what happens in the parent–child interactions surrounding help with homework and (b) studies that introduce parent involvement in homework as an exogenous intervention, randomly assign parents to conditions, and analyze the data at the same unit of analysis as the manipulation. In particular, future studies should make attempts to isolate the effect of involvement in homework by grade level and explore potential explanations for differential effects. A first step may be qualitative studies that compare the kinds of help given and the reasons parents become involved at each grade level. Future studies should explore whether the dissimilar nature of the involvement across developmental stages may explain the differential effect of homework involvement between students at different grade levels. Further, the finding that parent involvement in homework had a negative relationship to achievement of middle school students presents a particular dilemma. It is critical that future research explore the motivation behind and nature of parent involvement during this developmental stage.
This synthesis also pointed to the distinct impact of different forms of parent involvement in homework on academic achievement. Clearly, future research should employ interventions that distinguish between types of parent involvement, not simply encourage involvement per se or present multiple forms of involvement as an undifferentiated package of interventions. One of the most striking relationships revealed by this synthesis was the positive correlation found between parents’ use of homework rule setting and academic achievement. A closer look at individual studies shows that rule setting has generally been defined as a multiple component strategy, including deciding when and where children are to complete homework. Future research should make attempts to refine the rule-setting construct, as well as other broad forms of involvement, to better identify what particular elements of involvement in homework may be most effective for what types of students, and when.
Attempts should also be made to assess nonachievement measures that have been largely neglected in the literature, including student attitudes, motivation, study habits, and homework problems. For example, theoretically, one way parent involvement in homework may affect achievement is by first affecting motivation. Little or no research has been conducted to assess motivation as either an outcome of involvement or mediator of the relationship between involvement and achievement. Similarly, future research should directly assess whether the more abstract mediators suggested throughout this article—such as time on task, attentiveness, communication of expectations and guidelines, reinforcement, motivation, and improved self-regulation—account for the relationship between parent involvement in homework and achievement.
Evidence of Policy Statements in Light of New Evidence
Based on the results of this research synthesis, we might suggest several guidelines for educators and education policy makers regarding parent involvement in homework. First, on the basis of the strongest evidence, we might suggest the following:
Schools and teachers should encourage parent involvement through Grade 5 for the purposes of improving achievement and homework completion rates and decreasing problems with homework.
Parent involvement in homework should be encouraged for high school students when parents have a particular and relevant expertise, for the purpose of improving achievement.
Parent involvement in homework may also be encouraged for middle school grades for the purposes of improving homework completion rates and decreasing problems with homework. However, this recommendation should be considered with caution in light of the lack of an involvement effect on achievement among middle school students and until we have a better understanding of the dynamics of assistance for students in these grade levels.
Schools and teachers may want to encourage particular forms of parent involvement. Specifically, setting rules about when and where homework should be done may have the most beneficial effects for students.
When involving parents directly in mathematics homework, special attention should be paid to ensuring that parents have the requisite skills to be effective mentors.
Finally, in districts and schools that wish to emphasize parent involvement in homework, it may be necessary to consider resource-intense efforts to increase not only the amount, but also the quality of parent involvement in homework. For example, districts and schools might consider providing workshops to teach parents how to be involved in homework and to provide support for parent questions about involvement. Districts and schools might also consider adopting a contract system in which parents are asked to agree to be involved in homework. Districts, schools, and teachers should consider utilizing the Internet, developing Web pages that provide general information to parents about how to be involved in homework as well as specific information about nightly homework assignments. And these efforts may need to be sustained for long periods, following students as they progress through grade levels, for the efforts to have their desired effect.
Conclusion
Students who do homework have better school outcomes than students who do not (Cooper, 1989; Cooper et al., 2006). Our findings suggest that parent involvement in homework could affect student success by having a positive impact on homework completion and by reducing student problems with homework. Yet the effect of parent involvement on achievement was negligible to nonexistent, except among the youngest students. Clearly, it is important to consider the developmental stage of the student when parents become involved in homework. Further, different forms of parent involvement in homework may have different effects. Finally, parent’s skills in the subject area may be an important mediator of the effect of helping. Each of these moderating influences suggests it is time for the study of parent involvement in homework to begin testing hypotheses grounded in psychological and educational theory. The findings of this research synthesis provide some guidance for future investigations that will be both useful and illuminating of the underlying social and psychological dynamics of parent help with homework.
Tables
|
TABLE 1 Potential effects of parent involvement in homework
| Positive Effects |
|---|
| Accelerates learning (cf. Epstein et al., 1997) |
| Increases time spent studying |
| Makes homework study more efficient, effective, and focused |
| Enhances proximal achievement-related outcomes |
| Improves homework completion (cf. Cooper et al., 2000) |
| Improves homework performance (cf. Callahan et al., 1998) |
| Promotes positive affect |
| Enhances positive mood and attentiveness during homework (cf. Leone & Richards, 1989) |
| Enhances enjoyment during homework (cf. Shumow, 1998) |
| Improves attitudes toward homework and school (cf. Cooper et al., 1998) |
| Facilitates communication between parent and child (cf. Hoover-Dempsey et al., 2001) |
| Enhanced expression of parent beliefs and expectations about school |
| Enhances feedback, reinforcement, or both for desired homework behavior |
| Facilitates communication between parent and teacher (cf. Epstein & Van Voorhis, 2001) |
| Improves behavior during homework and school (cf. Sanders, 1998) |
| Enhances development of self-regulation and study skills (cf. Xu & Corno, 1998) |
| Negative Effects |
| Interference with learning (cf. Epstein, 1988) |
| Confusion of instructional techniques (cf. Cooper et al., 2000) |
| Help beyond tutoring (cf. Cooper et al., 2000) |
| Emotional costs and tension (cf. Levin et al., 1997) |
| Increased fatigue, frustration, disappointment |
| Increased tension between mother and child |
| Increased pressure on student to perform well (cf. Cooper et al., 2000) |
| Increased differences between high and low achievers (cf. McDermott et al., 1984) |
|
TABLE 2 Complete list of information retrieved from studies
| Report characteristics |
| 1. Author name |
| 2. Type of research report (journal article, book chapter, book, dissertation, master’s thesis, private report, government report, school or district report, conference paper, other type of report) |
| 3. Source of the data (for naturalistic cross-sectional designs) |
| Research design |
| 1. Manipulations of parent training for homework involvement |
| a. Random assignment |
| b. Quasi-experiment |
| i. Analytic strategy used to equate students |
| 2. Naturalistic, cross-sectional designs with statistical controls |
| a. Analytic strategy used to equate students |
| 3. Naturalistic, cross-sectional designs using correlation |
| Parent involvement variable |
| For studies using a manipulation of parent involvement in homework: |
| 1. Total number of training sessions |
| 2. Length of training sessions |
| 3. Frequency of training sessions |
| 4. Duration of training sessions |
| 5. Subject matter (reading, other language arts, math, science, social studies, foreign language, other, multiple subjects) |
| 6. Method of training |
| For studies using naturalistic, cross-sectional designs: |
| 1. Type of involvement (monitoring, rule setting, direct aid) |
| 2. Parent/guardian involved in homework |
| 3. Subject matter of homework |
| Sample |
| 1. Number of parents/students in each condition |
| 2. Ability label (gifted, average, at-risk, underachieving, learning deficit, behaviorally disordered, general) |
| 3. Socioeconomic status (low, low-middle, middle, upper-middle, upper, mixed, no socioeconomic status information) |
| 4. Grade level |
| 5. Sex |
| 6. Ethnicity (White, African American, Asian American, Hispanic, Native American, other, not specified) |
| 7. Classroom type (regular, special, gifted) |
| 8. Setting (school, home, nonschool/nonhome) |
| Outcome measure |
| 1. Subject matter (reading, other language arts, math, science, social studies, foreign language, other, multiple subjects) |
| 2. Type of achievement or achievement-related outcome (standardized achievement test; nonstandardized achievement test; class grades; homework grades; homework completion rate; student study habits/skills; student attitudes toward school, homework, or subject; parent attitudes toward school, homework, or subject; parent competency beliefs; or student conduct) |
| 3. When outcome was measured relative to the end of training |
| Estimate of the effect |
| 1. Effect size unit of analysis |
| 2. Direction of the effect |
| 3. Magnitude of the effect |
|
TABLE 3 Randomized experiments of parent training for homework involvement, with the student as the unit of assignment and analysis
| Author (year) | Type of document | Sample size: Number of classes training group/control group | Intervention type | Intervention: Duration in weeks/total number of sessions/minutes per session | Grade level | Subject matter | Type of outcome measure | Effect size (d) |
|---|---|---|---|---|---|---|---|---|
| Doering (1993) | Dissertation | NR/22/21 | Workshop or videotape | 35/6/60 | 4 | Reading | California Achievement Test in Reading | +.35 |
| Newsletter | 35/35/NA | Attitude Toward Reading Scale (adapted from Attitude Toward Mathematics Scale; Peterson, 1978) | −.27 | |||||
| Math | California Achievement Test in Math Attitude Toward Mathematics Scale (Peterson, 1978) | +.78 | ||||||
| −.73 | ||||||||
| Homework completion rate | +.69 | |||||||
| Kiesner (1997) | Dissertation | NR/11/12 | Individual meetings | 8/2/60 | 6–8 | Math | Math achievement test developed by researcher | .00 |
| Videotape | 8/1/60 | Class grades | −.40 | |||||
| Weekly phone calls | 8/8/NA | Cumulative score on math quiz developed by researcher | −.26 | |||||
| Homework accuracy | −.11 | |||||||
| Homework completion rate | +.26 | |||||||
| Meteyer (1998) | Dissertation | 13/25/27 | Workshop | 6/1/90 | 4–5 | Math | Class grades | +.16a |
| Bi-weekly phone calls | 6/3/NA | Curriculum-Based Measurement Test | +.24b | |||||
| Homework completion rate | +.26c | |||||||
| Homework problems Checklist (parent perceptions of homework problems; Anesko et al., 1987) | −1.06d | |||||||
| Nesbitt (1993) | Dissertation | NR/36/32 | Newsletter | 16/4/NA | 2–3 | Reading | Treatment A | |
| Iowa Test of Basic Skills (1993) Homework habits (No. times student sent to office for homework infractions) | −.04 | |||||||
| −.03 | ||||||||
| NR/34/32 | Newsletter | 16/4/NA | Treatment B | |||||
| Invitation to attend class | 16/1/NA | Iowa Test of Basic Skills (1993) Homework habits | +.13 | |||||
| −.24 | ||||||||
| NR/33/32 | Individual meetings | 16/4/NR | Treatment C | |||||
| Iowa Test of Basic Skills (1993) Homework habits | −.18 | |||||||
| −.19 | ||||||||
| Tamayo (1992) | Dissertation | NR/16/15 | Individual meetings | 9/1/60 | 7 | Math | Score on computation section of the mathematics subtest, level 2 of the Stanford Achievement Test | −.17 |
| Weekly phone calls | 9/9/NA | |||||||
| Class grades (quarter 3) | −.14 | |||||||
| NR/17/16 | Homework completion rate (quarter 3) | −.13 | ||||||
| Toney, Kelley, & Lanclos (2003) | Journal article | NR/13/12 | Individual meetings | 6/1/90 | 6–8 | English, social studies, math, science | Homework problem checklist (parent perceptions of homework problems; Anesko et al., 1987) | −2.51e |
Note. All unadjusted standardized d coefficients were computed using the means, standard deviations, and sample size. NA = not applicable; NR = not reported.
aThis effect size was collapsed across two effects sizes for class grades at different times: at 24 days following completion of the intervention (d =+.18) and at 48 days following completion of the intervention (d =+.13).
bThis effect size was collapsed across two effects sizes for the Curriculum-Based Measurement Achievement Test at different times: at 24 days after completion of the intervention (d =+.25) and at 48 days after completion of the intervention (d =+.23).
cThis effect size was collapsed across two effects sizes for homework completion rate at different times: at 24 days following completion of the intervention (d =+.17) and at 48 days following completion of the intervention (d =+.34).
dThis effect size was collapsed across two effects sizes for homework problems at different times: at 24 days after completion of the intervention (d =−.95) and at 48 days after completion of the intervention (d =−1.18).
eThis effect size was collapsed across two effects sizes for homework problems at different times: immediately after completion of the intervention (d =−2.45) and after a delay (d =−2.57).
|
TABLE 4 Randomized experiments of parent training for homework involvement, with class as the unit of assignment and student as the unit of analysis
| Author (year) | Type of document | Sample size: Number of classes training group/control group | Intervention type | Intervention: Duration in weeks/total number of sessions/minutes per session | Grade level | Subject matter | Type of outcome measure | Effect size (d) |
|---|---|---|---|---|---|---|---|---|
| Albright (2004) | Unpublished manuscript | 5/51/22 | Newsletter | 4/12/NA | 2 | Spelling | Spelling achievement test created by the researcher | −.06 |
| 5/42/20 | Homework Grades | −.37 | ||||||
| 5/55/28 | Homework completion rate | −.01 | ||||||
| 5/32/13 | The Parent Perceptions of Efficacy Regarding Homework Scale (Hoover-Dempsey et al., 1998) | −.41 | ||||||
| Balli, Demo, & Wedman (1998) | Journal article | 2/22/25 | Newsletter | 12/20/NA | 6 | Math | Math achievement test assessed for content validity by mathematics educator | +.02 |
| Homework grades | +.65 | |||||||
| Rillero & Helgeson (1995) | Conference paper | 8/101/99 | Newsletter | 14/NR/NA | Middleschool | Science | The Science Process Assessment | −.03 |
| 8/54/48 | Science Lab Practical Exam A | +.18 | ||||||
| 8/45/44 | Science Lab Practical Exam B | −.23 | ||||||
| 8/101/99 | Attitudes Toward Science in School Assessment | −.11 |
Note. All unadjusted standardized d coefficients were computed using the means, standard deviations, and sample size.
|
TABLE 5 Quasi-experimental studies with a parent training for homework involvement manipulation
| Author (year) | Type of document | Sample Size: Number of classes/training group/control group | Intervention type | Intervention: Duration in weeks/total number of sessions/minutes per session | Grade level measure | Subject matter | Type of outcome measure | Covariates | Effect size (d) |
|---|---|---|---|---|---|---|---|---|---|
| Austin (1988) | Dissertation | 4/27/22 | Workshop | 24/2/NR | 1 | Reading | Georgia Criterion Reference Test Reading Score | California Achievement Test Reading Score | −.05 |
| 4/45/31 | Macmillan Series R Reading Readiness Score | +.04 | |||||||
| 4/27/22 | Math | Georgia Criterion Reference Test Math Score | California Achievement Test Math score | −.04 | |||||
| 4/40/31 | Dekalb County Math End-of-Readiness Math Test score | +.93 | |||||||
| Pigford (1987) | Dissertation | 8/78/90 | Newsletter | 32/1/NA | 1 | Math | Progress and Review in Mathematics Education (PRIME) | Pretest on PRIME | +.96a |
| Workshop | 32/1/NR | ||||||||
| Comprehensive Test of Basic Skills | Pretest on PRIME | +.17b | |||||||
| Van Voorhis (2003) | Journal article | 10/233d | Newsletter | 18/NR/NR | 6 & 8 | Science | Homework grades Class grade (science) | Prior science achievement, gender, ethnicity, mother's education level, ability, teacher, homework completion rate (and family involvement for class grades) | +.43e |
| 10/240d | +.31e | ||||||||
| 10/233d | Homework completion rate | Prior achievement, gender, ethnicity, mother's education level, ability, teacher | .00e | ||||||
| Warrick (2000) | Dissertation | NR/30/35 | Newsletter | 9/9/NA | 5 | Math | Iowa Test of Basic Skills (ITBS) 11 Mathematics Battery | Pretest on ITBS | +.02 |
| Workshop | 9/9/60 | ||||||||
| Class grades | Baseline grades | −.18 | |||||||
| White (1996) | Dissertation | 8/92/74 | Newsletter | 11/1/NA | 4 | Math | Mathematics achievement test developed by the researcher | Pretest on achievement test | −.02c |
| Workshop | 11/1/NR | Classroom grades | Baseline grades | −.09 | |||||
| ITBS 1996 score | ITBS 1995 score | +.11 |
Notes. Unless otherwise noted, all adjusted standardized d coefficients were calculated by computing the effect size for the pretest and subtracting that number from the effect size calculated for the posttest.
aThis effect size for the PRIME was collapsed across underachieving below-kindergarten-level students, (n = 30, d +.16), underachieving at-kindergarten-level students, (n = 65, d +.33), and average-achieving students, (n = 72, d +2.87).
bThis effect size for the CTBS was collapsed across underachieving below-kindergarten-level students, (n = 30, d +.87), underachieving at-kindergarten-level students, (n = 65, d +.63), and average-achieving students, (n = 73, d +.42).
cThis effect size was calculated using the F statistic and sample sizes from an ANCOVA.
dTotal sample size is reported because the number of participants in each group for each outcome remained unclear.
eThis effect size was calculated using the t statistic and sample sizes from a multiple regression analysis.
|
TABLE 6 Results of analyses examining the overall effect of parent training for homework involvement on outcomes
| Outcome | k | d | 95% Confidence Interval | Q | |
|---|---|---|---|---|---|
| Low estimate | High estimate | ||||
| Randomized experiments | |||||
| Academic achievement (unadjusted) | 5 | .11 (.09) | −.06 (−.16) | .27 (.34) | 8.24† |
| Academic achievement (unadjusted without Nesbitt, 1993, and Doering, 1993) | 3 | .04 (.00) | −.17 (−.27) | .24 (.27) | 3.11 |
| Homework completion rate | 4 | .28* (.28)† | .00 (−.01) | .55 (.56) | 3.09 |
| Frequency of homework problems | 3 | −.84** (−1.20)* | −1.10 (−2.33) | −.57 (−.06) | 32.35** |
| Experiments randomized by class, analyzed by student | |||||
| Academic achievement (unadjusted) | 3 | .00 (.01) | −.16 (−.24) | .16 (.26) | 3.73 |
| Quasi-experiments | |||||
| Academic achievement (adjusted) | 5 | .23** (.22)* | .14 (.01) | .32 (.43) | 19.86** |
Note. Random effects estimates are presented in parentheses.
†p <.10.
*p <.05.
**p <.01.
|
TABLE 7 Results of moderator analyses examining the effect of parent training for homework involvement on academic achievement
| Moderator | k | d | 95% Confidence Interval | Qb | |
|---|---|---|---|---|---|
| Low estimate | High estimate | ||||
| Outcome measure | .88 (.52) | ||||
| Homework grades | 1 | −.11 (−.11) | −.93 (−.93) | .71 (.71) | |
| Class grades | 3 | .02 (.02) | −.29 (−.29) | .33 (.33) | |
| Nonstandardized test | 2 | .13 (.12) | −.19 (−.22) | .45 (.46) | |
| Standardized test | 3 | .18 (.16) | −.09 (−.29) | .45 (.61) | |
| Grade level | 4.28* (3.39)† | ||||
| Elementary (2–5) | 3 | .22* (.23) | .02 (−.06) | .42 (.52) | |
| Middle school (6–8) | 2 | −.18 (−.18) | −.49 (−.49) | .14 (.14) | |
| Subject matter | .02 (.01) | ||||
| Math | 4 | .11 (.12) | −.08 (−.23) | .31 (.47) | |
| Reading | 2 | .09 (.09) | −.25 (−.26) | .42 (.44) | |
| Intervention type | .40 (.15) | ||||
| Indirect | 2 | .04 (.04) | −.30 (−.30) | .38 (.38) | |
| Direct | 3 | .06 (.03) | −.16 (−.24) | .27 (.29) | |
| Both | 2 | .17 (.18) | −.13 (−.56) | .46 (.92) | |
| Number of workshops or meetings | .99 (.24) | ||||
| <4 sessions | 3 | .04 (.00) | −.17 (−.27) | .24 (.27) | |
| ≥4 sessions | 2 | .23 (.20) | −.09 (−.53) | .56 (.93) | |
Note. Only studies using a randomized experimental design with student as the unit of assignment and analysis were tested for moderator effects. Random-effects Q values and point estimates are presented in parentheses. Qb is an index of the heterogeneity between the group mean effect sizes. If Qb is significant, it indicates that the mean effect sizes across categories differ by more than sampling error; that is, there is a statistical difference between groups.
†p <.10.
*p <.05.
**p <.01.
|
TABLE 8 Characteristics of studies using data from the National Educational Longitudinal Study (NELS; 1988 or 1992) and performing multivariate analyses
| Author (year) document type | Sample characteristics | Modeling technique | Outcome variable(s) | Homework involvement variable/number additional predictors | Regression coefficient | Size and significance |
|---|---|---|---|---|---|---|
| Muller (1995) Journal article | N = 13,329 | Hierarchical multiple regression | Item response theory score on mathematics achievement test (8th-grade score) | Parent checks homework/28 | B = −1.05 | p <.001 |
| NELS 1988 | ||||||
| Grade 8 | ||||||
| Item response theory score on mathematics achievement test (10th-grade gain score) | Parent checks homework/29 | B = −.16 | p < .01 | |||
| Peng & Wright (1994) Journal article | N = 9,685a | Multiple regression | Achievement test composite (reading and mathematics) | Parental assistance with homework/9 | β=−.17 | p < .01 |
| Grade 8 | Parental assistance with homework/10 | β=−.17 | p <.01 | |||
| Simon (2000) Dissertation | N = 11,348b | |||||
| NELS 1992 | ||||||
| Grade 12 | Hierarchical multiple regression | Average English grades | Family rule about doing homework/21 | β=−.02 | t = −2.57, p <.01 | |
| Average math grades | β=−.04 | t = −5.25, p <.001 | ||||
| Achievement test composite (reading and mathematics) | β=−.01 | t = −1.63, ns | ||||
| Conduct | β=−.03 | t = 3.61, p <.001 | ||||
| Average English grades | Parent works on homework with teen/21 | β=−.01 | t = −1.47, ns | |||
| β=−.03 | t = −3.43, p <.001 | |||||
| Achievement test composite (reading and mathematics) | β=−.01 | t = −1.13, ns | ||||
| Conduct | β=−.06 | t = 5.89, p <.001 |
aThe authors write that this number represents the “effective” sample size, which was the actual sample size adjusted by the design effect, citing Ingels, Abraham, Karr, Spencer, & Frankel (1990).
bThis study used pairwise deletion. Sample sizes were N = 10,583 for Average English Grades, N = 10,549 for Average Math Grades, N = 9,921 for Achievement Test Composite, and N = 11,223 for Conduct
|
TABLE 9 Characteristics of studies using data other than the National Educational Longitudinal Study (1988, 1990, or 1992) and performing multivariate analyses
| Author (year) document type | Sample characteristics | Modeling technique | Outcome variable(s) | Homework involvement variable/number additional predictors | Regression coefficient | Size and significance |
|---|---|---|---|---|---|---|
| Anderson (2000) Dissertation | N = 91 | Logistic regressiona | Score > 315 on 4th-grade Michigan Assessment Program (MEAP) reading test | Help with homework/5 | B = 1.965 | p <.00 |
| Original data | ||||||
| Grade 4–5 | ||||||
| Broxie (1987) Dissertation | N = 58 | Hierarchical multiple regression | Class Grades (English, reading, spelling, social studies, & math) | Parent supplies materials and resources required for homework completion/0 | β = .11 | F = 1.35, p = .20 |
| Original data | ||||||
| Grade 4–6 | ||||||
| Parent directs student to do homework/1 | β = −.11 | F = 1.10, p = .30 | ||||
| Parent checks homework for completion/ 2 | β = .34 | F = 9.10, p = .05 | ||||
| Parent homework instruction/4 | β = .33 | F = 10.79, p = .05 | ||||
| Cooper, Lindsay, & Nye (2000)b Journal article | N = 709 | Multiple regression | Tennessee Comprehensive Assessment Program | Parent support for homework autonomy/3 | sr = .15 | F = 15.36, p <.0001 |
| Original data | ||||||
| Grades 2, 4, 6–8, 10–12 | Class grades (multiple subjects) | sr = .13 | F = 11.22, p <.0009 | |||
| Homework completion | sr = .21 | F = 31.43, p <.0001 | ||||
| Student attitude | sr = .02 | ns | ||||
| Tennessee Comprehensive Assessment Program | Parent direct involvement with homework/3 | sr = − .18 | F = 21.80, p < .0001 | |||
| Class grades (multiple subjects) | sr = − .11 | F = 7.18, p < .008 | ||||
| Homework completion | sr = .03 | ns | ||||
| Student attitude | sr = .05 | ns | ||||
| Tennessee Comprehensive Assessment Program | Parent elimination of homework environment distractors/3 | sr = − .01 | ns | |||
| Class grades (multiple subjects) | sr = .03 | ns | ||||
| Homework completion | sr = .07 | ns | ||||
| Student attitude | sr = .02 | ns | ||||
| Tennessee Comprehensive Assessment Program | Parent interference in homework/3 | sr = .01 | ns | |||
| Class grades (multiple subjects) | sr = .06 | ns | ||||
| Homework completion Student attitude | sr = .06 | ns | ||||
| sr = .01 | ns | |||||
| Deslandes, Royers, Potvin, & LeClerc (1999) Sample 1 Journal article | N = 525 | Hierarchical multiple regressionc | Grades in French | Parent help with homework/6 | β = .11 | F = 20.71, p < .05 |
| general education students | ||||||
| Original data | ||||||
| Grade 9 | ||||||
| Deslandes, Royers, Potvin, & LeClerc (1999) Sample 2 Journal article | N = 65 | Hierarchical multiple regressionc | Grades in French | Parent help with homework/5 | β = .32 | F = 1.56d, p < .05 |
| special education students | ||||||
| Original data | ||||||
| Grade 9 | ||||||
| Esptein, Simon, & Salinas (1997) Journal article | N = 255 | Multiple regression | Writing sample score (winter) | Parent participation in Teachers Involve | β = .18 | t = 2.91, p < .01 |
| Original data | ||||||
| Grade 6 & 8 | ||||||
| Writing sample score (spring) | Parents in School work interactive homework/6 | β = .23 | t = 3.40, p < .001 | |||
| Ginsburg & Bronstein (1993) Journal article | N = 93 | Multiple regression | Grade point average (GPA; reading, math, language, social studies, and science) | Mother's surveillance of homework/2 | β = − .33 | t = 3.70, p < .001 |
| Original data | ||||||
| Grade 5 | ||||||
| Stanford Achievement Test | β = − .18 | t = − 1.98, p < .01 | ||||
| GPA (reading, math, language, social studies, and science) | Father’s surveillance of homework/2 | β = − .24 | t = − 1.94, p < .01 | |||
| Stanford Achievement Test | β = − .21 | t = − 1.59, p > .05 | ||||
| Kim (2002) Journal article | N = 209 | Hierarchical multiple regression | GPA (English, mathematics, science, and social studies) | Parent homework checking/18 | β = .11 | p < .05 |
| Original data | ||||||
| Grades 7 & 8 | ||||||
| Multiple regression | Parent homework checking/12 | β = .11 | p < .05 | |||
| Okagaki & Frensch (1998) Journal article | N = 275 | Multiple regression | GPA (mathematics, science, language, and reading) | Parent help with homework/5 | European American | |
| Original data | ||||||
| Grades 4 & 5 | β = −.20 | p > .05 | ||||
| Latino | ||||||
| β = −.03 | p > .05 | |||||
| Asian American | ||||||
| β = −.25 | p < .05 | |||||
| Olson (1988) Dissertation | N = 191 | Multiple regression | California Achievement Test in math | Parent involvement in homework/7 | β = −.17 | F = 8.10, p < .01 |
| Original data | ||||||
| Grades 3–6 | ||||||
| California Achievement Test in reading | β = −.14 | F = 5.25, p < .05 | ||||
| Stepwise multiple regression | California Achievement Test in math | Parent involvement in homework/3 | β = −.12 | F = 7.83, p < .01 | ||
| Stepwise multiple regression | California Achievement Test in math | Parent involvement in homework/3 | β = −.15 | F = 6.23, p < .05 | ||
| Pezdek, Berry, & Renno (2002) Study 1 Journal article | N = 52 | Multiple regression | Mathematics achievement test developed by the researcher | Hours parents help with math homework/1 | pr = − .14 | ns |
| Original data | ||||||
| Grades 4–6 | ||||||
| Pezdek, Berry, & Renno (2002) Study 2 Journal article | N = 215 | Multiple regression | Mathematics achievement test developed by the researcher | Hours parents help with math homework/1 | pr = .06 | ns |
| Original data | ||||||
| Grades 4–6 | ||||||
| Sheldon & Epstein (2005) Journal article | N = 18 schools | Partial correlation | Percentage of students scoring satisfactorily on a standardized mathematics achievement test in 1998 | School practice of assigning mathematics homework requiring parent involvement/1 | pr = .60 | p < .05 |
| Original data | ||||||
| Grades K–12 | ||||||
| School practice of assigning mathematics homework requiring parent involvement/1 | pr = .60 | p < .05 | ||||
| Shumow & Miller (2001) Journal article | N = 1670 | Hierarchical multiple regression | GPA (language arts/English, mathematics, science, and social studies) | At-home parental involvement/5 | β = − .10 | p < .001 |
| Longitudinal Study of American Youth | ||||||
| Achievement test composite (science and mathematics) | β = − .14 | p < .001 | ||||
| Grade 7–8 | ||||||
| School Values | β = .13 | |||||
| Strauss (2000) Dissertation | N = 91 | Hierarchical multiple regression | Teacher or parent reported grades (English, language, mathematics, science, social studies) | Father's homework involvement/8 | β = − .04 | p < .01 |
| Original data | p > .05 | |||||
| Grades K-5 | ||||||
| Parent report of child's school enjoyment | β = − .01 | p > .05 | ||||
| Parent report of child's school-related anxiety | β = −.13 | p > .05 | ||||
| Van Voorhis (2003) Journal article | N = 233/240e | Hierarchical multiple regression | Science homework completion | Student report of family involvement/7 | β = .19 | t = 2.68, p < .01 |
| Original data | ||||||
| Grades 6 & 8 | Student report of family involvement/6 | β = .07 | t = 1.15, p > .05 | |||
| Wilkinson (1998) Dissertation | N = 40 | Multiple regression | Cumulative GPA | Help with homework/6 | β = .11 | p > .05 |
| Original data | ||||||
| Grades 9–12 | ||||||
| Zick, Bryant, & Osterbacka (2001) Journal article | N = 230 | Multiple regression | Mother’s assessment of school grades | Predicted frequency of parents' reading/helping with homework/14 | β = .09 | t = 2.31, p < .05 |
| National Survey of Families and Households | ||||||
| Father’s assessment of school grades | β = .06 | t = 1.45, ns | ||||
| Elementary |
aAuthor indicates that multinomial logistic regression was used. However, no evidence was found to suggest a case in which the dependent variable had more than two classes.
bData from this study are also used in Cooper, Jackson, Nye, & Lindsay (2001).
cAuthor indicates in text that a two-step hierarchical regression was conducted using covariates as block in Step 1 and six types of parent involvement in Step 2. Table indicates that parent involvement types were actually entered hierarchically in multiple steps. Betas reflect what authors reported in the tables.
dAuthor indicates that although beta was significant, F statistic was not.
eAn N of 233 was used in the analysis examining the relationship between family involvement in homework and homework completion with seven additional predictor variables in the model. An N of 240 was used in the analysis examining the relationship between family involvement in homework and homework completion with six additional predictor variables in the model.
|
TABLE 10 Characteristics of studies performing structural equation modeling
| Author (year) document type | Sample characteristics | Program used | Outcome variable(s) | Homework involvement variable/number of additional predictors | Regression coefficient | Size and significance |
|---|---|---|---|---|---|---|
| Cooper, Jackson, Nye, & Lindsay (2001)a Journal article | N = 214 Original data Grades 2 & 4 | MPlus | Class grade (multiple subjects) | Parent facilitation of homework/4 | β = .45 | p < .01 |
| Pomerantz & Eaton (2001) Journal article | N = 166 Original data Grades 4, 5, & 6 | AMOS | GPA (English, math, reading, science, social studies and spelling) | Maternal monitoring or helping with homework/3 | β = .12 | p < .05 |
| Schultz (1999) Dissertation | N = 255 Original data Grades 3–5 | EQS | Stanford Achievement Test in Reading | Parent help with homework/5 | NR | NR |
| Phonological Awareness score (Woodcock Word Attack) | NR | NR | ||||
| Stanford Achievement Test in reading | Parent checking of homework/5 | β = −.33 | NR | |||
| Phonological Awareness score (Woodcock Word Attack) | β = −.23 | NR |
Note. NR = not reported.
aData from this study are also used in Cooper, Lindsay, & Nye (2000).
|
TABLE 11 Characteristics of studies correlating parent involvement in homework and academic achievement
| Author (year) | Type of document | Sample size | Grade level | Type of parent help | Respondent | Subject matter | Outcome measure | Correlation |
|---|---|---|---|---|---|---|---|---|
| Anderson (2000) | Dissertation | 91 | 4–5 | General | Students | Reading | Michigan Educational Assessment Program Reading Test | +.43 |
| Balli, Wedman, & Demo (1997) | Journal Article | 47 | 6 | General | Parent/guardian | Math | Math achievement test assessed for content validity by mathematics educator | −.24 |
| Broxie (1987) | Dissertation | 56a | 4–6 | General | Parent/guardian | Multiple subjects | Class grades (English, reading, spelling, science, social studies, mathematics) | +.64 |
| 58a | Provides HW resources | +.22 | ||||||
| Rules | +.20 | |||||||
| Monitoring | +.51 | |||||||
| Direct aid | +.51 | |||||||
| Clark (1993) | Book chapter | 460 | 3 | General | Parent/guardian | Multiple subjects | Comprehensive Test of Basic Skills | +.09 |
| Direct aid | −.14 | |||||||
| Talks with child about homework | .00 | |||||||
| Monitoring | −.04 | |||||||
| Cooper, Jackson, Nye, & Lindsay (2001)b | Journal article | 214 | 2 & 4 | Rules | Parent/guardian | Multiple subjects | Tennessee Comprehensive Assessment Program Test of Student Ability | +.07 |
| Class grades | −.04 | |||||||
| Student attitudes toward homework | −.08 | |||||||
| Direct aid | Tennessee Comprehensive Assessment Program Test of Student Ability | −.26 | ||||||
| Class grades | −.29 | |||||||
| Student attitudes toward homework | +.23 | |||||||
| Deslandes, Royer, Potvin, & Leclerc (1999) | Journal article | 525 general education students | 9 | Direct aid | Student | French as a first language | Class grades | +.12 |
| 65 special education students | +.15 | |||||||
| Epstein (1988) | Government report | 1,021 | 1, 3, & 5 | General | Student | Reading, math | Reading achievement not specified | −.18 |
| Math achievement not specified | −.20 | |||||||
| Teacher report of discipline problems | +.13 | |||||||
| Homework Star | −.01 | |||||||
| Homework problems | +.07 | |||||||
| Ginsburg & Bronstein (1993) | Journal article | 60 fathersc | 5 | Monitoring, rules, & direct aid | Parent guardian | Reading, math, language, social studies, science | Class grades | −.44 |
| Stanford Achievement Test total battery score | −.33 | |||||||
| 93 mothersc | Class grades | −.28 | ||||||
| Stanford Achievement Test total battery score | −.25 | |||||||
| McGrath (1992) | PhD dissertation | 56 | 12 | Monitoring | Parent/guardian | English | Class grades | +.17 |
| Achievement test in English literature | +.09 | |||||||
| Okagaki, Frensch, & Gordon (1995) | Journal article | 82 | 4–5 | Direct aid | Parent/guardian | Reading, math, language | Comprehensive Test of Basic Skills | +.12 |
| Olson (1988) | PhD dissertation | 191 | 3–5 | Monitoring, rules, & direct aid | Parent/ guardian | Reading | California Achievement Test in Reading | −.25 |
| Homework grades | −.15 | |||||||
| Math | California Achievement Test in Math | −.27 | ||||||
| Homework grades | −.12 | |||||||
| Student | Reading | California Achievement Test in Reading | −.33 | |||||
| Homework grades | −.12 | |||||||
| Math | California Achievement Test in Math | −.20 | ||||||
| Homework grades | −.14 | |||||||
| Peng & Wright (1994) | Journal article | 9,685d | 8 | General | Parent | Reading, math | NELS Achievement Test Reading and Math Composite Score | −.13e |
| Rozevink (1995) | PhD dissertation | 176 | 9 | General | Student | Multiple subjects | Iowa Test of Basic Skills | +.19 |
| 183 | 12 | +.51 | ||||||
| Schultz (1999) | PhD dissertation | 255f | 3–5 | Monitoring | Student | Reading | Stanford Achievement Test Reading Comprehension subtest | −.20 |
| Woodcock Word Attack | −.18 | |||||||
| Direct aid | Stanford Achievement Test Reading Comprehension subtest | +.10 | ||||||
| Woodcock Word Attack | +.12 | |||||||
| 162f | Monitoring | Parent/guardian | Woodcock Word Attack | −.35 | ||||
| Direct aid | Woodcock Word Attack | −.20 | ||||||
| Shumow & Miller (2001) | Journal article | 879 | 7–8 | General | Parent/guardian | Multiple subjects | Self-report of GPA and behavior | −.22 |
| Smith (1998) | PhD dissertation | 86 | 4 | Monitoring | Parent/guardian | Reading | Iowa Test of Basic Skills reading comprehension | +.18 |
| Smock & McCormick (1995) | Journal article | 397 | K–12 | Direct aid | Parent/guardian | Multiple subjects | Parent perception of child school performance | −.11 |
| Walker, Hoover-Dempsey, Reed, & Jones (2000) | Conference paper | 20 | 4 | Monitoring, rules, & direct aid | Student | Reading, language, math | Class grades | −.40 |
| Tennessee Comprehensive Assessment Program total battery score | −.12 | |||||||
| Interview with student | Class grades | +.22 | ||||||
| Tennessee Comprehensive Assessment Program total battery score | +.22 | |||||||
| Parent/guardian | Class grades | −.52 | ||||||
| Tennessee Comprehensive | −.21 | |||||||
| Assessment Program total battery score | ||||||||
| Interview with parent/guardian | Class grades | +24 | ||||||
| Tennessee Comprehensive Assessment Program total battery score | +30 | |||||||
| Xu & Corno (2003) | Journal article | 119 | 6–8 | General | Student | Reading, math | Combined Score on California Test of Basic Skills in Reading and Math | −.03 |
| Yap (1987) | Conference Paper | 781 | 1–6 | Rules about time | Parent/guardian | Reading | Metropolitan Achievement Test in reading | +.30 |
| Rules about space | +.85 | |||||||
| Direct aid | +.42 |
aAlthough the sample size is distinct for different analyses, the same sample is being analyzed. Sample size is 56 for the correlation between general involvement and class grades and 58 for all other analyses examining the relationship between achievement and specific forms of parent involvement in homework.
bData from this study are also used in Cooper, Lindsay, & Nye (2000).
cThese sample sizes do not represent distinct samples. Rather, both mothers and fathers from the same family are reporting on the same 93 students.
dThe authors write that this number represents the “effective” sample size, which was the actual sample size adjusted by the design effect, citing Ingels, Abraham, Karr, Spencer, & Frankel (1990).
eThis effect size is collapsed across five effect sizes representing the effect of parent involvement for five different ethnic groups, Asian (−.08), Hispanic (−.03), Black (−.12), White (−.15), and Native American (−.04).
fNote that the difference in sample size for these analyses is not meant to indicate distinct samples. Rather, analyses were conducted separately for data based on student versus parent responses on parent involvement in homework. Students (N = 255) reported on their parents' involvement in homework; however, 162 of these students’ parents provided a report of their own involvement in homework.
|
TABLE 12 Results of analyses examining the correlation between parent involvement and academic achievement
| 95% Confidence Interval | |||||
|---|---|---|---|---|---|
| Moderators | k | r | Low estimate | High estimate | Q |
| Overall | 22 | .04** (.07) | .02 (−.08) | .05 (.21) | 1,418.79** Qb |
| Publication type | 140.35** (3.32)† | ||||
| Published | 11 | −.08** (−.06) | −.10 (−.14) | −.05 (.02) | |
| Unpublished | 11 | .12** (.19) | .10 (−.07) | .15 (.42) | |
| Outcome measure | 8.90** (.13) | ||||
| Class grades | 7 | .05† (.03) | .00 (−.18) | .10 (.25) | |
| Standardized test | 14 | .13** (.09) | .11 (−.12) | .15 (.29) | |
| Grade level | 106.00** (20.65)** | ||||
| Elementary (1–6) | 13 | .06** (.05) | .04 (−.16) | .08 (.25) | |
| Middle school (7–8) | 2 | −.17** (−.17)** | −.21 (−.26) | −.13 (−.09) | |
| High school (9–12) | 5 | .17** (.22)** | .11 (.07) | .24 (.36) | |
| Type of parent involvement | 754.56** (12.86)** | ||||
| Monitoring | 5 | −.90** (.09) | −.14 (−.12) | −.3 (.30) | |
| Rules | 3 | .54** (.33) | .51 (−.25) | .57 (.73) | |
| Direct aid | 9 | .10** (.11) | .07 (−.06) | .13 (.28) | |
| Mixed | 3 | −.21** (−.20)** | −.25 (−.31) | −.16 (−.08) | |
| Subject matter | 217.03** (52.14)** | ||||
| Math | 3 | −.19** (−.19)** | −.24 (−.24) | −.15 (−.15) | |
| Language arts | 3 | .12** (.12)** | .05 (.05) | .20 (.20) | |
| Reading | 6 | .20** (.13) | .18 (−.25) | .23 (.48) | |
| Respondent | 109.73** (0.40) | ||||
| Parent or guardian | 15 | .10** (.00) | .08) (−.19) | .12 (.20) | |
| Student | 10 | −.08** (.08) | −.11 (−.04) | −.06 (.20) | |
| Socioeconomic status | 374.86** (1.70) | ||||
| Low | 4 | .36** (.18) | .33 (−.32) | .39 (.60) | |
| Middle | 3 | −.19** (−.16)** | −.23 (−.26) | −.14 (−.06) | |
Note. Random-effects Q values and point estimates are presented in parentheses. Qb is an index of the heterogeneity between the group mean effect sizes. If Qb is significant, it indicates that the mean effect sizes across categories differ by more than sampling error; that is, there is a statistical difference between groups.
†p < .10.
*p < .05.
**p < .01.
1
Information was collected on several additional variables, including at-risk status, student ability, gender, and ethnicity. However, because of a lack of reporting or a lack of variation across categories, no analyses were conducted on these variables.
References
References marked with an asterisk indicate studies included in the meta-analysis. | |
| * | Albright, MI Parent involvement in spelling homework and children’s achievement: A multi-level analysis2004Manuscript submitted for publication Google Scholar |
| * | Anderson, JR An exploratory study of the relationship between selected children’s perceptions of maternal acceptance and their reading achievement. (Doctoral dissertation, Michigan State University, 2000)Dissertation Abstracts International200062343 Google Scholar |
| * | Austin, CA Homework as a parental involvement strategy to improve achievement of first grade childrenDissertation Abstracts International19885003A0622(UMI No. 8911432) Google Scholar |
| Balli, SJ When mom and dad help: Student reflections on parent involvement with homeworkJournal of Research and Development in Education199831142146 Google Scholar | |
| * | Balli, SJ, Demo, DH, Wedman, JF Family involvement with children’s homework: An intervention in the middle gradesFamily Relations199847149157 Google Scholar | ISI |
| * | Balli, SJ, Wedman, JF, Demo, DH Family involvement with middle-grades homework: Effects of differential promptingJournal of Experimental Education1997663148 Google Scholar | ISI |
| Barnett, V, Lewis, T Outliers in statistical data analysis19943rd edNew YorkJohn Wiley & Sons Google Scholar | |
| Borenstein, M, Hedges, L, Higgins, J, Rothstein, H Comprehensive Meta Analysis (Version 2.1) [Computer software]2005Englewood, NJBioStat Google Scholar | |
| * | Broxie, GJB Parental participation in homework completion as a predictor of academic success of students in the single parent family (Doctoral dissertation, Virginia Polytechnic Institute and State University, 1987)Dissertation Abstracts International1987512806 Google Scholar |
| Callahan, K, Rademacher, JA, Hildreth, BL The effect of parent participation in strategies to improve the homework performance of students who are at riskRemedial and Special Education199819131141 Google Scholar | SAGE Journals | ISI | |
| Carnegie Council on Adolescent Development Great transitions: Preparing adolescents for the new century1995New YorkCarnegie Corporation of New York Google Scholar | |
| * | Clark, R Chavk, NF Homework parenting practices that positively affect student achievementFamilies and schools in a pluralistic society1993 Albany, NYState University of New York Press5371 Google Scholar |
| Cohen, J Statistical power analysis for the behavioral sciences1988Hillsdale, NJErlbaum Google Scholar | |
| Cooper, H Homework1989New YorkLongman Google Scholar | |
| Cooper, H Synthesizing research: A guide for literature reviews19983rd edThousand Oaks, CASage Google Scholar | |
| Cooper, H Homework for all—in moderationEducational Leadership2001583438 Google Scholar | ISI | |
| Cooper, H, Hedges, LB Handbook of research synthesis1994New YorkRussell Sage Foundation Google Scholar | |
| * | Cooper, H, Jackson, K, Nye, B, Lindsay, JJ A model of homework’s influence on the performance of elementary school studentsJournal of Experimental Education200169181199 Google Scholar | ISI |
| * | Cooper, H, Lindsay, JJ, Nye, B Homework in the home: How student, family, and parenting-style differences relate to the homework processContemporary Educational Psychology200025464487 Google Scholar | Medline | ISI |
| Cooper, H, Lindsay, JJ, Nye, B, Greathouse, S Relationships among attitudes about homework, amount of homework assigned and completed, and student achievementJournal of Educational Psychology1998907083 Google Scholar | ISI | |
| Cooper, H, Robinson, JC, Patall, EA Does homework improve academic achievement—A synthesis of research, 1987–2003Review of Educational Research200676162 Google Scholar | SAGE Journals | ISI | |
| Corno, L Homework is a complicated thingEducational Researcher1996252730 Google Scholar | SAGE Journals | |
| Dauber, SL, Epstein, JL Chavk, NF Parents’ attitudes and practices of involvement in inner-city elementary and middle schoolsFamilies and schools in a pluralistic society1993 AlbanyState University of New York Press5371 Google Scholar | |
| * | Deslandes, R, Royer, E, Potvin, P, Leclerc, D Patterns of home and school partnership for general and special education students at the secondary levelExceptional Children199965496506 Google Scholar | SAGE Journals | ISI |
| * | Doering, WG The effects of a parent education program on fourth-grade student academic achievement: Homework, attitude, and attendance. (Doctoral dissertation, Miami University, 1993)Dissertation Abstracts International1993540806 Google Scholar |
| Dufresne, A, Kobasigawa, A Children’s spontaneous allocation of study time: Differential and sufficient aspectsJournal of Experimental Child Psychology198942274296 Google Scholar | ISI | |
| Duval, S, Tweedie, R A nonparametric “trim and fill” method of accounting for publication bias in meta-analysisJournal of the American Statistical Association2000a958998 Google Scholar | ISI | |
| Duval, S, Tweedie, R Trim and fill: A simple funnel plot-based method of testing and adjusting for publication bias in meta-analysisBiometrics2000b56276284 Google Scholar | ISI | |
| Eccles, JS, Harold, RD Booth, A, Dunn, JF Family involvement in children’s and adolescents’ schoolingFamily-school links: How do they affect educational outcomes?1996Hillsdale, NJErlbaum334 Google Scholar | |
| Eccles, JS, Midgley, C, Buchanan, CM, Flanagan, C, Mac Iver, D, Reuman, D Development during adolescence: The impact of stage/environment fitAmerican Psychologist19934890101 Google Scholar | Medline | ISI | |
| Epstein, JL Parents’ reactions to teacher practices of parent involvementElementary School Journal198686277294 Google Scholar | ISI | |
| * | Epstein, JL Homework practices, achievements, and behaviors of elementary school students. Baltimore: Center for Research on Elementary and Middle Schools1988(ERIC Document Reproduction Service No. ED301322) Google Scholar |
| Epstein, JL School/family/community partnershipsPhi Delta Kappan199576701712 Google Scholar | ISI | |
| Epstein, JL, Lee, S Ryan, BA, Adams, GR, Gullota, TO, Weissberg, RO, Hampton, RL National patterns of school and family connections in the middle gradesThe family-school connection: Theory, research, and practice1995 Thousand Oaks, CASage108154 Google Scholar | |
| Epstein, JL, Simon, BS, Salinas, KC Involving parents in homework in the middle gradesResearch Bulletin1997September18Bloomington, INPhi Delta Kappa, CEDR Google Scholar | |
| Epstein, JL, Van Voorhis, FL More than minutes: Teachers’ roles in designing homeworkEducational Psychologist200136181193 Google Scholar | ISI | |
| Erikson, E Identity: Youth and crisis1968New YorkNorton Google Scholar | |
| Forgatch, MS, Ramsey, E Boosting homework: A videotape link between families and schoolsSchool Psychology Review199423472484 Google Scholar | ISI | |
| Frome, PM, Eccles, JS Parents’ influence on children’s achievement-related perceptionsJournal of Personality and Social Psychology199874435452 Google Scholar | Medline | ISI | |
| * | Ginsburg, GS, Bronstein, P Family factors related to children’s intrinsic/extrinsic motivational orientation and academic performanceChild Development19936414611474 Google Scholar | Medline | ISI |
| Greenhouse, JB, Iyengar, S Cooper, H, Hedges, LV Sensitivity analysis and diagnosticsThe handbook of research synthesis1994 New YorkRussell Sage Foundation383398 Google Scholar | |
| Grolnick, WS, Benjet, C, Kurowski, CO, Apostoleris, NH Predictors of parent involvement in children’s schoolingJournal of Educational Psychology199789538548 Google Scholar | ISI | |
| Grolnick, WS, Ryan, RM Parent styles associated with children’s self-regulation and competence in schoolJournal of Educational Psychology198981143154 Google Scholar | ISI | |
| Grubbs, FE Sample criteria for testing outlying observationsJournal of the American Statistical Association1950212758 Google Scholar | |
| Gutman, LM, Midgley, C The role of protective factors in supporting the academic achievement of poor African American students during the middle school transitionJournal of Youth and Adolescence200029223248 Google Scholar | ISI | |
| Hedges, LB, Olkin, I Statistical methods for meta-analysis1985Orlando, FLAcademic Press Google Scholar | |
| Hedges, LV, Vevea, JL Fixed and random effects models in meta-analysisPsychological Methods19983486504 Google Scholar | ISI | |
| Hill, JP, Holmbeck, GN Whitehurst, GJ Attachment and autonomy during adolescenceAnnals of child development19863Greenwich, CTJAI Press145189 Google Scholar | |
| Hoover-Dempsey, KV, Bassler, OC, Burow, R Parents’ reported involvement in student homework: Strategies and practicesElementary School Journal199595435450 Google Scholar | ISI | |
| Hoover-Dempsey, KV, Battiato, AC, Walker, JMT, Reed, RP, DeJong, JM, Jones, KP Parental involvement in homeworkEducational Psychologist200136195209 Google Scholar | ISI | |
| Hoover-Dempsey, KV, Sandler, HM Parental involvement in children’s education: Why does it make a differenceTeachers College Record199597310331 Google Scholar | ISI | |
| Hoover-Dempsey, KV, Sandler, HM Why do parents become involved in their children’s education?Review of Educational Research199767342 Google Scholar | SAGE Journals | ISI | |
| Hutsinger, CS, Jose, PE, Larson, SL Do parent practices to encourage academic competence influence the social adjustment of young European American and Chinese American children?Developmental Psychology199834747756 Google Scholar | Medline | ISI | |
| Jeynes, WH A meta-analysis: The effects of parental involvement on minority children’s academic achievementEducation and Urban Society200335202218 Google Scholar | SAGE Journals | ISI | |
| Jeynes, WH A meta-analysis of the relation of parental involvement to urban elementary school student academic achievementUrban Education200540237269 Google Scholar | SAGE Journals | ISI | |
| Jeynes, WH The relationship between parent involvement and urban school student academic achievement: A meta-analysisUrban Education20074282110 Google Scholar | SAGE Journals | ISI | |
| Kay, PJ, Fitzgerald, M, Paradee, C, Mellencamp, A Making homework work at home: The parent’s perspectiveJournal of Learning Disabilities199427550561 Google Scholar | SAGE Journals | ISI | |
| * | Kiesner, JW The effects of a parental homework monitoring intervention on school engagement of high risk middle school students. (Doctoral dissertation, University of Oregon, 1997)Doctoral Abstracts International1997580980 Google Scholar |
| * | Kim, E The relationship between parental involvement and children’s educational achievement in the Korean immigrant familyJournal of Comparative Family Studies200233529540 Google Scholar | ISI |
| Kronholtz, J Homework becomes a critical assignment for moms and dadsThe Wall Street Journal1997June6A1A8 Google Scholar | |
| Laursen, B, Coy, KC, Collins, WA Reconsidering changes in parent-child conflict across adolescence: A meta-analysisChild Development199869817832 Google Scholar | Medline | ISI | |
| Leone, CM, Richards, MH Classwork and homework in early adolescence: The ecology of achievementJournal of Youth and Adolescence198918531548 Google Scholar | Medline | ISI | |
| Levin, I, Levy-Shiff, R, Appelbaum-Peled, T, Katz, I, Komar, M, Meiran, N Antecedents and consequences of maternal involvement in children’s homework: A longitudinal analysisJournal of Applied Developmental Psychology199718207227 Google Scholar | ISI | |
| Maertens, NW, Johnston, J The effects of arithmetic homework upon the attitudes and achievement of fourth, fifth, and sixth grade pupilsSchool Science and Mathematics197272117126 Google Scholar | |
| McDermott, RP, Goldman, SV, Varenne, H When school goes home: Some problems in the organization of homeworkTeachers College Record198485391409 Google Scholar | ISI | |
| * | McGrath, JB Student and parental homework practices and the effect of English homework on student test scoresDissertation Abstracts International19925310A3490(UMI No. 9231359) Google Scholar |
| * | Meteyer, BV The effects of manual-based training program on parent perceptions of homework problems, on mathematics homework, and on mathematics achievement. (Doctoral dissertation, Alfred University, 1998)Dissertation Abstracts International1998590725 Google Scholar |
| * | Muller, C Maternal employment, parent involvement, and mathematics achievement among adolescentsJournal of Marriage and the Family1995585100 Google Scholar | ISI |
| Natriello, G, McDill, EL Performance standards, student effort on homework, and academic achievementSociology of Education1986591831 Google Scholar | ISI | |
| * | Nesbitt, GK The effects of three school-to-home parental involvement communication programs on reading achievement conduct, homework habits, attendance, and parent-student-school interaction. (Doctoral dissertation, Georgia State University, 1993)Dissertation Abstracts International1993543993 Google Scholar |
| Odum, ME A tough assignment for teachers: Getting students to do homework is getting harderThe Washington Post1994May2A1A10 Google Scholar | |
| * | Okagaki, L, Frensch, PA Parenting and children’s school achievement: A multiethnic approachAmerican Educational Research Journal199835123144 Google Scholar | SAGE Journals | ISI |
| * | Okagaki, L, Frensch, PA, Gordon, EW Encouraging school achievement in Mexican American childrenHispanic Journal of Behavioral Sciences199517160179 Google Scholar | SAGE Journals | ISI |
| * | Olson, M Homework variables and academic achievement: An integrated studyDissertation Abstracts International19884908A2104(UMI No. 8821110) Google Scholar |
| * | Peng, SS, Wright, D Explanation of academic achievement of Asian American studentsJournal of Educational Research199487346352 Google Scholar | ISI |
| * | Pezdek, K, Berry, T, Renno, PA Children’s mathematics achievement: The role of parents’ perceptions and their involvement in homeworkJournal of Educational Psychology200294771777 Google Scholar | ISI |
| * | Pigford, T A study to investigate the effects of parent involvement on the mathematical achievement of first-grade pupils using a structured parent and teacher homework plan. (Doctoral dissertation, University of South Carolina, 1987)Dissertation Abstracts International1987481980 Google Scholar |
| * | Pomerantz, EM, Eaton, MM Maternal intrusive support in the academic context: Transactional socialization processesDevelopmental Psychology200137174186 Google Scholar | Medline | ISI |
| Pomerantz, EM, Grolnick, WS, Price, CE Elliot, AJ, Dweck, CS The role of parents in how children approach achievementHandbook of competence and motivation2005 New YorkGuilford Press259278 Google Scholar | |
| Reetz, LJ Parental perceptions of homeworkRural Educator1990121419 Google Scholar | |
| * | Rillero, P, Helgeson, SL An evaluation of the use of hands-on science homework assignments by sixth grade students and their parents1995AprilPaper presented at the annual meeting of the National Association for Research in Science TeachingSan Francisco, CA(ERIC Document Reproduction Service No. ED382478) Google Scholar |
| Roderique, TW, Polloway, EA, Cumblad, C, Epstein, MH, Bursuck, WH Homework: A survey of policies in the United StatesJournal of Learning Disabilities199427481487 Google Scholar | SAGE Journals | ISI | |
| Rosenthal, R Judgment studies: Design, analysis, and meta-analysis1987New YorkCambridge Google Scholar | |
| Rosenthal, R Cooper, H, Hedges, LV Parametric measures of effect sizeHandbook of research synthesis1994 New YorkRussell Sage231244 Google Scholar | |
| * | Rozevink, ML Selected small school motivational factors regarding homework completion in grades nine and twelve (Doctoral dissertation, Bowling Green State University, 1995)Dissertation Abstracts International19955759 Google Scholar |
| Sanders, MG The effects of school, family, and community support on the academic achievement of African American adolescentsUrban Education199833385409 Google Scholar | SAGE Journals | ISI | |
| Sanders, MG, Epstein, JL, Connors-Tadros, L Family partnerships with high schools: The parents’ perspective1999BaltimoreCenter for Research on the Education of Students Placed at Risk(Report No. 32) Google Scholar | |
| * | Schultz, MD Parental involvement in low-income and minority children’s education. (Doctoral Dissertation, State University of New York at Stony Brook, 1999)Dissertation Abstracts International199961549 Google Scholar |
| Scott-Jones, D Gordon, E Family influences on cognitive development and school achievementReview of research in education198411 Washington, DCAmerican Educational Research Association259304 Google Scholar | Abstract | |
| Shadish, WR, Ragsdale, K Random versus nonrandom assignment in controlled experiments: Do you get the same answer?Journal of Consulting and Clinical Psychology19966412901305 Google Scholar | Medline | ISI | |
| * | Sheldon, SB, Epstein, JL Involvement counts: Family and community partnerships and mathematics achievementJournal of Educational Research200598196206 Google Scholar | ISI |
| Shumow, L Promoting parental attunement to children’s mathematical reasoning through parent educationJournal of Applied Developmental Psychology199819109127 Google Scholar | ISI | |
| * | Shumow, L, Miller, JD Parents’ at-home and at-school involvement with young adolescentsJournal of Early Adolescence2001216891 Google Scholar | SAGE Journals | ISI |
| Simmons, RG, Blyth, DA Moving into adolescence: The impact of pubertal change and school context1987Hawthorn, NYAldine de Gruyter Google Scholar | |
| * | Simon, BS Predictors of high school and family partnerships and the influence of partnerships on student successDissertation Abstracts International20006110A3949(UMI No. 9993187) Google Scholar |
| * | Smith, BB Effects of home-school collaboration and different forms of parent involvement on reading achievement. (Doctoral Dissertation, Virginia Polytechnic Institute and State University, 1998)Dissertation Abstracts International1998593738 Google Scholar |
| * | Smock, SM, McCormick, SM Assessing parents’ involvement in their children’s schoolingJournal of Urban Affairs199517395411 Google Scholar | ISI |
| Stevenson, DL, Baker, DP The family-school relation and the child’s school performanceChild Development19875813481357 Google Scholar | Medline | ISI | |
| * | Strauss, RM A contextual approach to understanding fathers’ involvement in their children’s schoolingDissertation Abstracts International20006110B5601(UMI No. 9990428) Google Scholar |
| Sui-Chu, EH, Willms, JD Effects of parental involvement on eighth-grade achievementSociology of Education199669126141 Google Scholar | ISI | |
| * | Tamayo, LF Hispanic parent monitoring of seventh grade mathematics homework assignments and relationship with achievement and self-esteem. (Doctoral dissertation, University of Massachusetts, 1992)Dissertation Abstracts International1992531805 Google Scholar |
| * | Toney, LP, Kelley, M, Lanclos, NF Self- and parental monitoring of homework in adolescents: Comparative effects on parents’ perceptions of homework behavior problemsChild & Family Behavior Therapy2003253551 Google Scholar | ISI |
| U.S. Department of Education Parent and family involvement in education: 2002–20032005May(Indicator No. 43). National Center for Educational Statistics. Retrieved September 10, 2008, from http://nces.ed.gov/pubs2005/2005043.pdf Google Scholar | |
| * | Van Voorhis, F Interactive homework in middle school: Effects on family involvement and science achievementJournal of Educational Research200396323338 Google Scholar | ISI |
| Voelkl, KE Academic achievement and expectations among African-American studentsJournal of Research and Development in Education1993274155 Google Scholar | |
| * | Walker, JM, Hoover-Dempsey, KV, Reed, RP, Jones, KP “Can you help me with my homework?” Elementary school children’s invitations and perspectives on parental involvement2000AprilPaper presented at the annual meeting of the American Educational Research AssociationNew Orleans, LA(ERIC Document Reproduction Service No. ED443581) Google Scholar |
| Wallis, C How to make a better student: Their eight secrets of successTime1998October8085 Google Scholar | Medline | |
| * | Warrick, PL An intervention program for parental assistance with mathematics homework and the relationship with increased student achievementDissertation Abstracts International20006107A2638(UMI No. 9978212) Google Scholar |
| * | White, MS The effects of TIPS math activities on fourth grade student achievement in mathematics and on primary caregivers’ and teachers’ opinions about primary caregivers’ involvement in schoolwork. (Doctoral dissertation, University of Maryland at College Park, 1996)Dissertation Abstracts International1996580116 Google Scholar |
| * | Wilkinson, SC Influences on the academic performance of twice-exceptional high school students: An exploratory study of risk, resilience, and achievementDissertation Abstracts International19985907A2449(UMI No. 9993187) Google Scholar |
| Xu, J, Corno, L Case studies of families doing third-grade homeworkTeachers College Record1998100402436 Google Scholar | ISI | |
| * | Xu, J, Corno, L Family help and homework management reported by middle school studentsElementary School Journal2003103503517 Google Scholar | ISI |
| * | Yap, KO Improving chapter 1 through parents: A family goal program1987AprilPaper presented at the annual meeting of the American Educational Research AssociationWashington, DC Google Scholar |
| * | Zick, CD, Bryant, WK, Osterbacka, E Mothers’ employment, parental involvement and the implications for intermediate child outcomesSocial Science Research2001302549 Google Scholar | ISI |
| Zimmerman, BJ Boekaerts, M, Pintrich, P, Zeidner, M Attaining self-regulation: A social cognitive perspectiveHandbook of self-regulation2000 San Diego, CAAcademic Press1339 Google Scholar |
