Honor on Death Row

The Southern United States is described as having a culture of honor, an argument that has been used to explain higher crime rates in the Southern United States than in the rest of the country. This research explored whether the combination of honor-related violence and traditional southern politeness norms is related to regional differences in the degree of remorse expressed by those who have committed violent crimes. It was proposed that different social norms regarding politeness and apologies in the Southern United States would be reflected in the narratives provided by offenders. The data came from the final statements that offenders on death row made before they were executed. Results showed that, compared with offenders executed in the non-Southern United States, offenders executed in the South more often apologized for their crimes in their final statements, but they were not necessarily more remorseful.

cultures of honor have strict codes of both politeness and defense of honor, raises interesting questions about how apologies might be used in these cultures. If individuals are more likely to respond to insults with aggression, how do they reconcile this with competing norms regarding politeness and kindness (and, hence, apologies)? Would this make apologies more or less likely in a culture of honor?
In those regions of the United States that support the death penalty, offenders who are about to be executed are permitted to make a final statement in front of a group of witnesses, which may include members of their family and their victim's family. These narratives, which are usually made available to the public, provide a unique opportunity to examine remorse for an extreme, usually unambiguous event (capital murder) in a situation where the transgressor arguably has nothing left to lose. When most of the factors that might influence an apology for a criminal transgression (e.g., the threat of harsher punishment or a hope for leniency, the threat of retaliation) are stripped away, is southern politeness still apparent in offenders from the U.S. South? This research examines the use of apologies in this national sample to discover if there is a distinct pattern of remorse in the Southern honor states.

The Culture of Honor
Historically, more violence has been reported in the U.S. South than in the rest of the country. Research has shown, for example, that compared with the rest of the United States, the South has higher rates of homicide and other types of violent crime (Gastil, 1989), more of its citizens own weapons (Cohen & Nisbett, 1994), and it executes more prisoners under the death penalty (Snell, 2011). In examining these statistics more closely, however, both sociologists and psychologists have noted that homicide rates in the South are only higher than the rest of the country for argument-or insult-related conflict, and that regional differences disappear when this is taken into account (Ellison, 1991;Lee, Bankston, Hayes, & Thomas, 2007;Nisbett, 1993). The same is true of harsher Southern attitudes toward criminal offenders: There are no regional differences in punitiveness for crimes not resulting from arguments (Cohen & Nisbett, 1994).
Nisbett and colleagues (Cohen & Nisbett, 1997;Cohen, Nisbett, Bowdle, & Schwarz, 1996;Nisbett, 1993) propose that this is due to a distinct culture of honor in the Southern United States. According to this theory, the early settlers in the Southern United States, from Scotland and Ireland, had a long-standing herding culture. These settlers, whose livelihoods depended on the safety of their herds, became hypervigilant against any perceived threat to their herd, and thus they developed a reputation for toughness, and were known to defend their honor with extreme vigilance. In a series of archival, correlational, and experimental studies, Nisbett and his colleagues provide compelling evidence that Southerners, in general, respond more violently to insults and are more likely to resolve conflict with violence after being insulted Nisbett, 1993). This is also reflected in Southern social policies and laws: Cohen (1996) found that, compared with the Northern U.S. states, the U.S. South has less stringent gun laws and fewer policies regarding mandatory arrest for domestic violence. Recent research shows that those in honor states in the South are more approving of defensive violence (Hayes & Lee, 2005), that there are more weapons in schools and more school shootings in honor states (Brown, Osterman, & Barnes, 2009), and, in a novel study examining violence against the self, that antidepressant use is lower, and suicide rates higher, in honor states compared with non-honor states (Osterman & Brown, 2011).

Honor and Apologies
Individuals in honor cultures also tend to have a reputation for kindness, generosity, politeness, and hospitality (Abbott, 1847;Cohen & Vandello, 2004;Kierner, 1996), and there is empirical evidence of this kindness. Levine, Martinez, Brase, and Sorenson (1994) examined United Way contributions per capita in the United States, along with various friendly and helpful behaviors, and concluded that Southerners were more generous and helpful than those from any other region in the country. Cohen and Vandello (2004) reanalyzed these data with statistics on homicides in the same regions, and found that in comparison with the North, those counties in the South with higher friendliness ratings also had higher rates of argument-related (but not felony-related) homicides. Cohen, Vandello, Puente, and Rantilla (1999) argue that one reason for Southern politeness is self-protective: As Southerners are quick to use violence to defend their honor, politeness is used to prevent or dispel conflict. Cohen et al. (1999) suggest that because of their established social scripts when it comes to conflict, Southerners do not signal their displeasure clearly (e.g., by displaying anger). This can result in sudden and violent responses to insult. In a series of experiments, Cohen and his colleagues (1999) found that Southerners, as compared with Northerners, not only displayed fewer anger cues and were more prone to unsignaled anger following an annoying stimulus but were also less able to recognize anger cues in others. Thus, being polite can prevent one from unintentionally raising the ire of a fellow Southerner. In a similar way, Southerners have been known to be vigilant about repaying social debts (Kierner, 1996), presumably to avoid future conflict.
It is unlikely that politeness when one is facing execution stems from a fear of retaliation for oneself, but it may be expressed out of concern for one's family. Those in honor cultures do not only seek to defend their own honor but are also concerned about the honor of their family (Wyatt- Brown, 1982). Research shows that family reputation is important in cultures of honor. For example, parents from honor cultures have been found to be less accepting of their children's relationships with those from ethnic outgroups, in part due to family reputation vulnerability (Munniksma, Flache, Verkuyten, & Veenstra, 2012), and male children from honor cultures in the U.S. South are more likely to be named after their father, which may serve, among other things, to strengthen kinship bonds (Brown, Carvallo, & Imura, 2014). Thus, politeness may serve as a form of impression management, not only for oneself but also for one's kin.
There is little research that addresses how these politeness norms might be manifested when it comes to an offender from a culture of honor having the opportunity to make a public statement after committing a transgression. On one hand, given that they are motivated to dispel conflict quickly to avoid retaliation, and given that they might also be concerned about tarnishing the reputation of their family, they might be more likely to apologize after committing a transgression. Indeed, certain cultures of honor, such as Japan, have well-defined rules regarding apologies (Haley, 1998;Wagatsuma & Rosett, 1986). On the other hand, to apologize is to admit to a mistake, which would be at odds with a desire to protect and maintain one's honor. Hayes and Lee (2005) describe honor is "an inner conviction of self-worth" (p. 601). Because such importance is placed on self-regard and pride in honor cultures, to apologize might be seen as a weakness. Ironically, even though being a member of an honor culture may be associated with having a strong selfimage, the maintenance of that self-image can be costly for the individual (Baumeister, Smart, & Boden, 1996). When those in honor cultures are faced with a threat to their honor, they tend to respond with increased anger and shame (Rodriguez Mosquera, Manstead, & Fischer, 2002). Directing too many resources at defending their self-worth may leave individuals unable to admit to and apologize for their mistakes. Some anthropological accounts support this notion, and have noted the rarity of apologies in certain honor cultures (e.g., Campbell, 1964).
There is also little research that suggests how sincere these apologies, if they are offered, might be. True repentance involves more than just making an apology. Researchers generally agree that additional indicators of remorse must be present, such as accepting responsibility for one's actions, attempting to repair the damage, if possible, and expressing concern for the victim (Darby & Schlenker, 1982;Exline & Baumeister, 2000;Schmitt, Gollwitzer, Förster, & Montada, 2004). If those from a culture of honor are simply following established cultural scripts when they apologize, then their apologies would not necessarily be expected to be more sincere. In fact, some researchers have pointed out that Southern politeness may veil hostility. Johnstone (1992) suggests that Southerners are particularly prone to "linguistic politeness" (p. 5), such as using indirectness to place doubt on their true intentions or to mitigate their responsibility for perceived misdeeds. Johnstone (1992) argues that the combined politeness and belligerence present in Southern dialect may signal underlying hostility rather than a sincere form of politeness. This lack of sincerity is also suggested by Wyatt-Brown (1982), who claims that certain gestures of hospitality were more about showing oneself to be better than one's neighbors than about a genuine concern for helping others. Some historians suggest that the overt generosity of Southerners traditionally served to place them in a position of power over the recipients of their kindness (Kierner, 1996;Wyatt-Brown, 1982), and that in addition to making their guest indebted to them, public displays of hospitality could bolster their social image. Thus, even if they do apologize, Southerners' apologies might not be accompanied by true remorse.

Honor on Death Row
When an offender is on death row and is about to be executed, it is unlikely that the fear of retaliation would influence his or her decision to apologize. It may be more likely that his or her response would be guided by cultural norms. If Southerners have well-defined politeness scripts, then it is predicted that they would be more likely to apologize to their victims than those from other regions in the United States. Thus, the first hypothesis is that offenders executed in the South will be more likely to apologize than those from other regions. The second hypothesis is that, although they may be more likely to apologize, Southerners will not necessarily be more remorseful than those from non-honor states. If Southerners are simply following established cultural scripts regarding apologies, then these apologies would not necessarily be expected to be sincere.

The Data Set
In the United States, offenders on death row are given the opportunity to make a last statement immediately before they are executed. These final words, often made in the presence of the offender's family, the victim's family, and the media, are made available to the public through justice department websites, personal websites, and the news media. During the past decade, there has been growing interest in death row last statements and how they can expand our understanding of how offenders feel about their crimes, their victims, and their own impending death. One common finding in the research is that repentance and seeking forgiveness are key themes (Eaton & Theuer, 2009;Heflick, 2005;Rice, Dirks, & Exline, 2009;Schuck & Ward, 2008;Vollum, 2010), with approximately one third of offenders offering an apology in their last statement (Eaton & Theuer, 2009;Rice et al., 2009).
Most of the research on last statements has used data from Texas exclusively, in part because Texas executes more people than any other state and hence is a rich source of data, and in part because the TDCJ publishes on its website the full text of the last statements of all offenders executed since 1982; this information is not as freely available in other states with the death penalty. This study seeks to examine the last statements of all offenders executed throughout the entire United States. This not only provides an opportunity to replicate the Texas-based research but also makes it possible to test the current hypothesis regarding regional differences in apology and remorse. Although most executions have taken place in the Southern United States (Snell, 2011), there is enough geographical diversity to test the hypothesis that those from honor states will show different degrees of remorse than those from non-honor states.

Participants
A total of 679 people were executed in the United States between January 2000 and December 2011. There were 670 men and 9 women, with a mean age of 28 years (SD = 8.29) at the time of the offense. They had been on death row for an average of 14 years (SD = 5.88) before being executed. The ethnic background of the offenders was 57% White, 34% Black, 8% Hispanic, and 1% Other.
Before conducting the analysis, some individuals were removed from the data set. First, there were three offenders (Louis Jones, Juan Garza, and Timothy McVeigh) who were executed by the U.S. Federal Government (and not a specific U.S. state), and thus were not included in the analyses. Second, because the Southern culture of honor hypothesis relates only to White males from honor states (Nisbett, 1993), non-Whites and females were excluded from the main analyses; however, parallel analyses were conducted with the non-White males as a further test of the culture of honor hypothesis with this data set. This resulted in a sample size of 379 White males and 288 non-White males. Third, because the key variables in this study come from the last statements, only those who made a last statement were included in the analyses. Of the 379 White offenders, 279 (74%) chose to make a last statement, 92 (24%) declined, and the remaining 8 (2%) were not found. Of the 288 non-White offenders, 231 (80%) chose to make a last statement, 52 (18%) declined, and the remaining 5 (2%) were not found. For the final White sample (n = 279), 220 were from honor states and 59 were not from honor states. The mean age was 29.9 years (SD = 8.96) at the time of the offense and they were on death row for an average of 14.1 years (SD = 5.94) before they were executed. For the final non-White sample (n = 231), 206 were from honor states and 25 were not from honor states. The mean age was 25.5 years (SD = 6.94) at the time of the offense and they were on death row for an average of 13.3 years (SD = 4.92) before they were executed.

Procedure
The last statements were collected primarily from the website of the Prosecuting Attorney of Clark County, Indiana (www.clarkprosecutor.org). This website provides statistics on all U.S. executions, including the last statements, where available. In instances where the last statement was not available from this source, attempts were made to find it elsewhere, by checking the website of the relevant department of justice, conducting an Internet search of the offender's name, and searching news articles published within a week following the execution.
The website of the Prosecuting Attorney of Clark County, Indiana, while extensive, only includes last statements for offenders executed after 1999. In seeking out other sources for last statements, it soon became apparent that the only other consistently reliable source was the website of the TDCJ, which only publishes information on executions in Texas. Rather than risk skewing the data by oversampling from Texas for executions prior to 2000, the decision was made to only include executions that had taken place from January 2000 to December 2011.
Information about each offender was recorded, including the state in which they had been executed, their age, gender, how long they had spent on death row, and details about their crime and the victims. The last statements were compiled in a separate file, identified only by the offender's name.
Content analysis was used to code the last statements for remorse-related variables. Two independent researchers read each statement and coded for the presence (coded as 1) or absence (coded as 0) of each variable, using a predetermined coding scheme. Inter-rater reliability (Cohen's κ) was substantial or almost perfect, ranging from .61 to .96 (see Table 1), and any disagreements were resolved with discussion.
Participants were classified as being from an honor state or not from an honor state. Data on where the offenders were from originally was not available; thus, an alternative (albeit flawed) method was to classify them in terms of the state in which they were executed. Some research on the culture of honor has included the Western states as honor states (e.g., Brown et al., 2009;Cohen, 1998;. The total number of offenders from the Western states who fit the criteria for this study (i.e., they were male, White, and made a last statement) was 11, which was too low to conduct a separate analysis for the West. For this reason, the main analyses were conducted with the Western states removed, but additional analyses were conducted with the Western states combined with the Southern states and with the Western states combined with the non-Southern states. Any differences in the results of these analyses are noted in the Results section.

Measures
Apology. Offenders were recorded as having apologized if they specifically said they were sorry to the victim or the victim's family (e.g., "I do apologize to the Surace family . . . I am sorry for what I did," Ryan Dickson, executed April 26, 2007, Texas).
Responsibility. Offenders were recorded as having accepted responsibility if they explicitly said they took responsibility, or if they acknowledged what they had done without making excuses, for example, To the families of Leslie Roark, Pam Jones, Charles Smith and Sandra Wilson, I am profoundly sorrowful for taking their lives. I know that I have caused irreparable damage to each surviving Ask for forgiveness. If an offender specifically asked for forgiveness or expressed a desire to be forgiven in the future, they were coded as asking for forgiveness (e.g., "I wish to apologize to the people who I have hurt and I ask for their forgiveness. I don't deserve it, but I do ask for it," Jack Trawick, executed June 11, 2009, Alabama).
Regret. Offenders were considered to have expressed regret if they specifically said they felt remorseful or if they mentioned regretting or wishing to reverse their actions (e.g., "If there were any way I could change things and bring them back I would," Richard Dinkins, executed January 29, 2003, Texas).
Earnestness. If the offender expressed love or gratitude toward the victim's family, or if they appeared to be sincere in their statements, they were coded as being earnest (e.g., "I know I took someone very precious to you . Remorse. Remorse was calculated by adding the dichotomous scores on responsibility, ask for forgiveness, regret, and earnestness, for a total possible score of 4. The mean score on remorse was .93 (SD = 1.17). Frequencies of the variables included in the content analysis of the last statements are shown in Table 1.

Demographic Variables
Before examining the variables related to the last statements (and before filtering out those who did not make a last statement), analyses were conducted to determine whether there were differences between Southerners (n = 299) and non-Southerners (n = 60) in demographic and offense-related variables. A logistic regression was performed with region (South; non-South) as the response variable and last statement (i.e., whether the offender chose to make one or not), the type of crime committed (i.e., whether it consisted of capital murder alone, or whether the crime also involved kidnapping or abduction, robbery, drugs, sex offense, or nonsexual assault), whether the victim was known to the offender or not, the number of victims, whether members of the victim's family were present at the execution, and the length of time the offender spent on death row before being executed as the explanatory variables.
A test of the full model against a constant-only model was not statistically significant, χ 2 (11, n = 359) = 12.45, p = .33, indicating that the set of predictors did not accurately distinguish between those from the South and those not from the South. Individually, none of the explanatory variables was significant except for the length of time the offender spent on death row before being executed, β = −.056, p = .03, odds ratio = .946. Those in the South spent an average of 13.98 years on death row (SD = 6.167), while those not in the South spent an average of 15.82 years on death row (SD = 6.11).

Main Analyses
Regression analyses were conducted to determine whether being executed in an honor state predicted apology and remorse in the last statements of those on death row. In each analysis, region (honor state, non-honor state) was included as an explanatory variable, with variables related to the crime and offender (whether the victim was known to the offender, the number of victims, whether the victim's family was present at the execution, the length of time the offender had been on death row, and whether the crime was murder alone, kidnapping, robbery, drugs, a sex offense, or nonsexual assault) included as possible covariates. The covariates were entered first, in two blocks: First, the variables related to the offender and victim, and second, the variables related to the crime. Region was entered last.
Apology. Because apology was a dichotomous variable (the offender either apologized or he did not), logistic regression was used to test the extent to which the explanatory variables predicted whether or not the offender apologized in his last statement. A test of the full model against a constant-only model was statistically reliable, χ 2 (11, N = 279) = 24.083, p = .012, indicating that the set of predictors accurately distinguished between those who apologized and those who did not. Overall predictive accuracy was 66%, with Nagelkerke's R 2 = .14. Table 2 shows regression coefficients, Wald statistics, and odds ratios for each of the predictors. Region was a significant predictor of apology. The odds ratio of 2.123 indicates that, holding other variables constant, offenders were approximately 2 times more likely to apologize in their last statement if they were executed in an honor state. Another way to examine this is to look at the frequencies: Almost half (46%) of offenders from honor states apologized, whereas only 29% of offenders from the non-honor states apologized. The only covariate that accurately distinguished whether an offender apologized or not was the length of time the offender had been on death row, whereby the longer the offender had been on death row, the less likely he was to apologize.
As explained previously, these analyses were conducted without the Western states. The analyses were repeated with (a) the Western states included as honor states and (b) the Western states included as non-honor states. The overall findings were very similar, with only marginal differences. When the Western states were included as honor states, the explanatory power of the model was slightly lower, with Nagelkerke's R 2 = .12, and although the odds ratio for region still indicated that offenders were approximately 2 times more likely to apologize in their last statement if they were executed in an honor state, this relationship was only marginally significant (p = .07). When the Western states were included as non-honor states, the explanatory power of the model was similar, with Nagelkerke's R 2 = .13. The odds ratio for region indicated that offenders were approximately 2.3 times more likely to apologize in their last statement if they were executed in an honor state, p = .02.
As a further test of the applicability of the culture of honor theory to these data, identical analyses were conducted with the non-White offenders. Given that the theory only purports to explain the retaliatory behavior of White males from the South, differences in rates of apology and expressions of remorse would not be expected in the non-White sample. In line with this prediction, the model, when tested with the non-White sample, was not statistically reliable, and region did not significantly predict apology, regardless of how the Western states were categorized (ps > .70). None of the covariates were significant.
Remorse. Because remorse was a continuous variable, linear regression was used to test the extent to which the region in which the offender was executed predicted whether or not he appeared to be remorseful in his last statement. The same covariates were entered as the previous analysis, with remorse as the outcome variable. Table 3 shows the results of the regression analysis. Region did not account for unique variance in remorse, regardless of whether the Western states were excluded, included as honor states, or included with the non-honor states (ps > .15). However, the length of time the offender had spent on death row did significantly predict remorse, whereby the longer the offender had been on death row, the less likely he was to express remorse. Also, whether the victim's family was present at the execution predicted remorse, in that remorse was higher when the victim's family was present (M = 1.08, SD = 1.21) than when the victim's family was not present (M = .57, SD = .96). None of the other covariates significantly predicted remorse.
When these analyses were repeated with the non-White sample, region did not significantly predict remorse, regardless of how the Western states were categorized (ps > .70), and none of the covariates were significant.

Discussion
These data indicate that there are regional differences in apologies offered in the last statements made by individuals on death row. In line with predictions, apologies were more likely from Southerners than from non-Southerners. This does not necessarily mean that Southerners were more remorseful, however. The analysis revealed that they were not more likely than non-Southerners to express remorse, defined as the extent to which they accepted responsibility, asked for forgiveness, expressed regret, and appeared to be earnest. Researchers generally agree that true remorse includes more than simply saying "I'm sorry"; it also must include, at a minimum, an acceptance of responsibility for the offense and an offer to make amends (Darby & Schlenker, 1982;Exline & Baumeister, 2000;Schmitt et al., 2004). Thus, the absence of any other indicators of remorse in the Southern sample suggests that their apologies, while more frequent than those of non-Southerners, were not necessarily more heartfelt.
One explanation for the higher likelihood of Southern apologies is that Southerners are more wedded to particular norms regarding politeness and kindness than those from the rest of the United States. The apparent lack of true remorse in their apologies supports the arguments of theorists who suggest that in the South politeness can be used to mask hostility (Johnstone, 1992) and/or to deflect anger (Cohen et al., 1999). It may also be that the Southern offender is more motivated than the non-Southern offender to protect, both publically and privately, his reputation as an honorable person. If, as noted by David Gibbs (the executed offender discussed at the beginning of this article), the act of murder creates dissonance with the Southern offender's views of himself as a "southern country gentleman," then a deathbed apology might serve to protect his private image (i.e., "I did a bad thing, but I apologized") and, through impression management, his public image. It may also serve to protect the reputation of his family and/or his culture. The significant difference between apologies in Southerners and non-Southerners only held for White males. Nisbett (1993) argued that Southern violence comes from the original herding culture of the U.S. South, which was mostly composed of Scottish or Irish settlers. This culture of honor, then, should only be evident in their descendants, which would consist mainly of Whites. The fact that non-White Southerners were not more likely to apologize than the rest of the United States offers further support for culture of honor theory.
It should be noted that the data set is not without limitations. Whether offenders apologized to their victims' families was based solely on the last statements of offenders. There may have been offenders who apologized to their victims' families at some point before their execution (e.g., at their trial or by letter while they were incarcerated) and thus did not feel the need to do so in their last statement. While the number of these cases is likely quite small, it cannot be ruled out that there may be regional differences in apologies made prior to (and not included in) the last statement. Also, this data set relies on prison officials and journalists for written records of the last statements. While every attempt was made to cross-check the accuracy of the statements, some could not be verified.
Because the non-Southern states have fewer executions, in part because fewer of them endorse the death penalty but also because the top four states for number of executions from 1977 to 2010 are all Southern states-Texas, Virginia, Oklahoma, and Florida accounted for 60% of all executions (Snell, 2011)-it may be that the non-Southern offenders who were executed were in some way worse than those who were executed in the South. The argument could be made that one needs to be particularly hardened or considered to be beyond any hope of redemption to be executed in a non-Southern state. If this is true, then perhaps the non-Southerners were less likely to apologize simply because they consist only those extreme offenders who are incapable of remorse. Assessing the severity of the crime is challenging with this sample. Because all of the crimes involved capital murder, there would be a high degree of subjectivity in determining what constituted a more "severe" murder. In this study, the number of victims and the type of crime committed were used as covariates in an attempt to assess severity objectively, and they did not significantly predict apology or remorse. Future research might consider a qualitative analysis of the actual crimes to further test this hypothesis.
Data on where the offenders were originally from were not available, and thus offenders were classified as "Southern" or "non-Southern" based on the state in which they were executed. Clearly, this is a flawed technique, as it does not account for the fact that an offender who was raised in the South may have committed the crime in the North, and thus would be misclassified as "non-Southern." This is a limitation of using archival data. Given that the data supported the theoretically derived predictions, the argument could be made that if all offenders were classified correctly the significant difference between Southerners and non-Southerners would be even more pronounced. However, the findings should be interpreted with caution due to the potential misclassification of some offenders.
These findings are notable in that they provide further evidence of the Southern culture of honor using a nontraditional sample. While the data set does have limitations, it is unique in that it includes actual narratives from offenders convicted of extreme crimes, in a setting that could not be reproduced in an experimental study. This study extends the work on the southern culture of honor by examining a relatively untested aspect of honor (at least in North America): how the "paradox of politeness" (Cohen & Vandello, 2004) influences Southern apologies. These findings also add to the growing literature on the use of apologies in the criminal justice system, replicating previous research that has found that offenders are motivated to apologize (e.g., Sherman et al., 2005;Umbreit, Vos, Coates, & Brown, 2003), but also suggesting that there are cultural differences in offenders' motivations for apologizing. This may have implications for restorative justice programs, in that their effectiveness may be dependent, in part, on the cultural backgrounds of the offender and victim. Finally, the findings reported here suggest that regional differences in apologies may not indicate real differences in remorse. However, it is important to note that these apologies may still serve a valuable function for victims. If Southern victims (or their families) adhere to the same politeness norms and social scripts as offenders, it may be that an apology from an offender, regardless of whether it is sincere, may make victims' families feel better.