A woman’s place is in the ‘home’? Gender-specific hiring patterns in academia in gender-equal Norway

We investigated women’s hiring patterns in academia, bridging the gap between two streams of literature typically not brought in conversation with one another: (1) access to tenured professorship by gender; (2) academic mobility and the disadvantage it can create for women. We studied all recruitments of new permanent associate and full professors by open call from 2007 to 2017 at one faculty of the largest university in Norway (N = 1009 applicants). We found that women were not reluctant to apply, nor systematically dispreferred in the selection process, nor deterred by the prospect of relocating. About 40% of applicants were women and the percentage was roughly maintained among winners. About 80% of women applicants were external candidates, nearly matching their male counterparts. However, women had an advantage when they were internal applicants and a disadvantage when external, and this disadvantage was a function of the distance. Underneath an overall gender-equal picture, there are gender-specific hiring patterns that are not the consequence of women’s application choices.

seeing a rapid growth of casual employment at the expense of job security (Brown et al., 2010;Jessop, 2018;Larson et al., 2014;Wånggren, 2018). There is evidence that casualization undermines women's career progression more than men's (Ackers and Oliver, 2007;Barrett and Barrett, 2011;Ceci et al., 2014;Winchester and Browning, 2015). The sizeable growth of the pool of qualified women in the past decades has not led to a proportionate share of women rising to professorship (European Commission, 2019; Monroe and Chiu, 2010). Gender balance is not just a question of time, it does not materialize spontaneously: the dynamics are more complex (D'Amico et al., 2011;Valian, 1999).
One of the main points of contention in the literature is the role that women's own application patterns play in generating unequal outcomes. This debate reflects a more general controversy on the role of women's own learned behavioural schemas in perpetuating gender inequality, as well as a feature of the current scholarly and public debate on gender equality that 'tends to emphasise choices rather than constraints' (Baker, 2010b: 215;Morley, 1994). A number of recent studies conclude that, compared to men, women academics have a lower propensity to compete, lower career confidence and a higher tendency to withdraw and forego future opportunities following rejection (Baker, 2010a;Buser et al., 2014;Chesterman et al., 2005;London et al., 2012). This strand of literature explains the gender gap with women's lower proclivity to apply for faculty positions, promotion, professorship or habilitation (Bagues et al., 2017;Bosquet et al., 2019;De Paola et al., 2018;Nielsen, 2016A;Pautasso, 2015;Shaw and Stanton, 2012).
According to other contributions, women's purported tendency to self-select is a myth, a gender stereotype. Winchester et al. (2006) interviewed staff holding central roles in the promotion process to senior faculty positions in 17 Australian universities (such as chairs of promotion committees, deputy chancellors and directors of HRM [human resources management]). While the interviewees assumed that women tended to be reluctant to apply for promotion, the analysis of promotions data showed that application rates for women were similar to men's. The different outcomes across genders could therefore not be attributed to unequal application patterns, but rather to the selection process itself where women applicants are systematically dispreferred ('bottleneck problem'). Several recent studies point to the existence of a bottleneck for women applicants to faculty positions (De Paola and Scoppa, 2015;De Paola et al., 2018;Hart, 2011;Husu, 2000;Monroe et al., 2008;Nielsen, 2016a;Weisshaar, 2017;Winkler, 2004) habilitation (Filandri and Pasqua, 2019) and promotion (Bonawitz and Andel, 2009;Marini and Meschitti, 2018;Sabatier et al., 2006;van den Brink et al., 2006). These results resonate with other recent findings on gender and access to academic employment in general (Bagilhole, 2006;Moss-Racusin et al., 2012;Sheltzer and Smith, 2014). The vast majority of studies on gender and academic hiring point to a bottleneck for women; exceptions are very rare Lutter and Schröder, 2016;Williams and Ceci, 2015). Several works within this strand of literature question the objectivity of ostensibly gender-neutral and merit-based selection processes (Nielsen, 2016B;Roos and Gatta, 2009;van den Brink and Benschop, 2012).

Literature review
Our study stems from the need to investigate the role of mobility in driving application and selection patterns by gender. This is a question that has so far received insufficient attention in the literature. Women's propensity to apply and likelihood to win have typically been investigated without breaking down the pool of applicants into local and external candidates. Factoring in mobility is important because there is solid evidence linking academic mobility and gender inequality. There is a gap between two streams of literature that are usually not brought in conversation with each other: the literature on gender and access to faculty positions, and that on academic mobility and gender.
Internationally, there is increasing emphasis on academic mobility in academic policies, understood not only as student mobility but also as mobility of knowledge workers (Ferencz and Wächter, 2012). Academic mobility is portrayed as a strategy to accumulate symbolic, social and cultural capital -a determinant of scholarly and professional success in academia, and a precondition for career advancement (Bourdieu, 1988;Kim, 2017;van Balen et al., 2012). The emphasis on mobility in academic policies has been such that some authors speak of 'fetishisation of mobility' (Bauder et al., 2017: 1), treated as an asset irrespective of the context in which it occurs, or 'an outcome in its own right' simplistically equated with upward social mobility (Ackers, 2008: 411; see also Leung, 2017). There is a wealth of research showing that the emphasis on academic mobility tends to put women at a systematic disadvantage because it is poorly compatible with family commitments, particularly for women who have children (Børing et al., 2015;Fritsch, 2015;Leemann, 2010;Scheibelhofer, 2010;Shauman and Xie, 1996;Vohlídalová, 2014). Women academics are typically less internationally mobile than their male colleagues; mobile women academics are particularly disadvantaged in dual-career families and vulnerable to major conflicts between working life and family life that put their relationships at risk (Ackers, 2004;Bauder, 2015;Melville et al., 2019;Probert, 2005;Smith et al., 2016). Some authors have taken a more radical stance and problematized the success models in academia (including the emphasis on mobility) as tilted in favour of men (Knights and Richards, 2003;Lund, 2015). Gender disparities in mobility patterns seem a common issue across several academic systems, interwoven with matters of job security and self-selection. Brechelmacher et al. (2015) interviewed academics in eight European countries and reported a recurring notion among interviewees that women are over-represented in less attractive academic positions (such as externally funded projects), and this lack of job security exacerbates the tension between professional and private life for women who, 'in critical phases decide against their careers, for example by refusing to go abroad' (as one of their interviewees puts it; our italics). Are women self-selecting by not applying for jobs, when the job would involve relocating?
It has been argued that women academics are typically tied to a specific spatial context because of their partner's career and their own care responsibilities (Ackers and Gill, 2009). Employees may develop their career in a single academic institution, and that does not necessarily mean that they are not marketable beyond that institution (Sagaria and Dickens, 1990). Bryson (2004) reported data from a Women in Higher Education survey, based on interviews with about 600 academics on fixed-term contracts. Compared to men, women were significantly less willing to take up a new post that involved relocating, but significantly more willing to be mobile locally; and the willingness to move was a function of the distance. The smaller the distance, the more willing women were to be mobile, the larger the gap with men's preferences. Local mobility (within a 1-hour commute) was an option for 55% of the women and 27% of the men; relocation within the UK for 20% of the women (and 27% of the men); and expatriation for 20% of the women and 44% of the men. The same study showed that women academics, as an occupational group, did not wish to have less commitment and obligations than their male colleagues, and did not have a preference for fixedterm and generally casual employment situations. Gendered patterns of mobility have been found also among (aspiring) academic managers, with women being less likely than their male counterparts to apply for leadership jobs in other universities (Chesterman et al., 2005;Shepherd, 2017). However, what happens to women academics who are not deterred by the prospect of relocating?
Gender can be a major factor affecting how women external applicants are received, particularly if they come from a different academic system. Part of the literature has called attention to the existence of a discrepancy between the prevalent portrayal of academic mobility as a career asset and its concrete impact on particular groups of academics (Morano-Foadi, 2005). In some contexts, it is actually non-mobile academics who get a permanent position earlier (Cruz-Castro and Sanz-Menéndez, 2010). Academic mobility may even lead to deskilling when the value of experience acquired outside of the institution is not recognized (Bilecen and van Mol, 2017;Weiss, 2005Weiss, , 2016. Most of these contributors stress that academics are not only knowledge producers, they are also workers embedded in particular employment relations and social and cultural contexts, but they are hardly ever described as such in the literature on academic mobility (Bauder, 2015;Shinozaki, 2017). Other expressions that erase the individual worker's perspective are typically preferred, such as 'circulation of human capital', 'knowledge exchange' or 'brain drain' when the movement of skilled workers systematically impoverishes particular countries to the advantage of others, reflecting and amplifying global inequalities (Altbach, 2013;Blachford and Zhang, 2014;Nerdrum and Sarpebakken, 2006). Women can be doubly disadvantaged by their gender and outsider status when relocating, and highly skilled women are no exception (Isaakyan and Triandafyllidou, 2016). Seeber et al. (2016) found that in the Flemish university system, men, native-born and internal candidates have better chances of getting hired than women, foreigner and mobile peers, and the effects are stronger for faculty positions. The intersection between gender and international academic mobility to take up faculty jobs is an area where inequality dynamics are particularly visible. In her study on academic hiring in Germany, Shinozaki (2017) shows that, throughout all career stages, German men predominate; the higher up the career ladder (professorship), the stronger the predominance. The effect is driven by gender as well as nationality: the proportion of foreign women (but not foreign men) declines at the more senior levels. In essence, mobility does not in itself 'serve as a ticket for career progression' for women (Shinozaki, 2017(Shinozaki, : 1325; this could in turn reflect an underlying 'feeling that the jobs . . . should be protected from foreign competition' (Skachkova, 2007: 730; see also Canagarajah, 1999). There are demonstrated inequalities in access to mobility; in addition to that, 'mobility itself can create new inequalities or reinforce existing ones', including gender inequality (Bilecen andvan Mol, 2017: 1245). So much for gender and external hiring; however, is there evidence that gender can play a role in internal hiring as well?
Academia differs from other types of organizations (Chan, 2006) in that its success models give prominence to mobility and typically encourage external hiring at all levels. The tendency to emphasize the value of local experience in academic recruitment and prefer internal candidates is depicted as unquestionably negative and it is referred to as 'inbreeding', a potent metaphor as biological inbreeding may lead to disability in the new generation (Gorelova and Yudkevich, 2015;Inanc and Tuncer, 2011). Altbach et al. (2015) argue that faculty inbreeding is associated with preferential hiring standards for internal candidates and leads to the exclusion of women and other groups traditionally marginalized in academia. With this approach, the problem then becomes how to counter inbreeding; for instance by means of policies, including 'mobility throughout scientific and academic careers' (Horta, 2013: 487). At the same time, even critics admit to some advantages in internal hiring, perceived as lower risk (Horta and Yudkevich, 2016). With the increasing casualization of the academic workforce, the number of faculty positions shrinks and the perceived cost of a sub-optimal hire increases. Appointment panels may use local experience as a quality indicator, and 'there is no one better able to demonstrate experience than an individual already undertaking the role' (Shepherd, 2017: 85; see also Abramo et al., 2015). Despite its claim to uniqueness and the emphasis on external hiring, academia sometimes follows the same recruitment logics that characterize other types of organizations, including the avoidance of risk, the need to maintain an incentive for employees already working at the organization, and the effort to capitalize on competency and social ties accumulated within the institution (Bidwell, 2011;Bradley, 2006;Chan, 2006;Floyd and Dimmock, 2011;Jones and Stout, 2015). One of the academic leaders interviewed by Shepherd (2017: 85) stated: 'You increasingly have to do the job before you get it.' Relatedly, there is a strand of research investigating the job tasks of externally as opposed to internally hired faculty. These vary depending on the institutional context and academic system. The two groups may significantly differ in research productivity (Horta, 2013;Inanc and Tuncer, 2011) or not (Alipova and Lovakov, 2018;Smyth and Mishra, 2014;Tavares et al., 2019). However, a pattern of findings can be noticed in the literature with regard to teaching: internally hired faculty members seem to contribute disproportionately more to teaching activities, enabling colleagues to spend more time on research (Altbach et al., 2015;Gorelova and Yudkevitch, 2015;Horta et al., 2010). This in itself can provide an organizational rationale for internal recruitment in academia, and it intersects interestingly with gender. There are structural mechanisms in place that systematically drive women on fixed-term contracts away from research and towards teaching, and the pattern seems to recur across different academic systems (Aiston and Jung, 2015;Barrett and Barrett, 2011;O'Brien and Hapgood, 2012). This creates a dual academic labour-market model and disadvantages women in contexts where research performance is assessed with a metric that does not account for their teaching load. Lecturer jobs, in particular, have been called 'a female ghetto' (O'Brien andHapgood, 2012: 1002). In short, the literature on academic inbreeding does not quite do justice to the complexity of the considerations involved in the choice to hire internal candidates, nor to the role that gender can play in it; the matter has been investigated to an extent in organizations other than academia, with mixed findings (Gorman and Kmec, 2009;Lyness and Schrader, 2006).
Against the backdrop of this rather complex scenario, we will now present our data. Table 1 below summarizes our model and research questions.

Setting
Our study was carried out in Norway, a social democracy with a solid welfare state concretely supporting mothers' employment and fathers' caring role in the family. Norway is ranked among the most gender-equal countries in the world, particularly for education and employment (World Economic Forum, 2019). The main part of the Norwegian university system is public, academics are civil servants and their recruitment is tightly regulated by statutory law and institutional regulations affording ample protection to gender equality (Frølich et al., 2018). In Norwegian academia, gender balance defined as 'at least 40% of each gender' in each academic rank has been an explicit, concretely pursued political goal since the 1980s (Nielsen, 2014(Nielsen, , 2017. Indeed the gender balance figures for Norwegian academia seem relatively reassuring, compared to the European Union average. Our study was carried out at one Faculty of the Norwegian University of Science and Technology (NTNU), the largest university in the country with about 40,000 registered students. Table 2 gives an impression of the gender balance at the EU, national and local level.
The setting of the present study seems reasonably gender equal if one looks at the aggregate figures. Since 2004, the percentage of women among tenured associate professors at the faculty considered has been over 40%; in the past decade it has swung between 44% and 53% (Norwegian Centre for Research Data, 2020).
All of the above does not mean that Norwegian academia is a gender equality paradise; recent contributions have brought evidence to the contrary (Bergman, 2013;Husu, 2015;Piñheiro et al., 2015;Rasmussen, 2015;Seierstad and Healy, 2012). Like academia internationally, Norway is experiencing casualization of the academic workforce, growing competition for faculty jobs, increased emphasis on mobility and gender-specific patterns of academic mobility, particularly following relatively recent higher education governance reforms (Christensen et al., 2014;Frølich, 2005Frølich, , 2006Maassen et al., 2011;Nerdrum and Sarpebakken, 2006;Thun, 2019;Vabø et al., 2014). Also in Norway, the division of labour in academia is gendered in that most temporary senior lecturer (teaching-intensive) positions are today held by women. By 'senior', we refer to posts reserved to candidates with a doctorate. The national trend is reflected in situation at the faculty we investigated: locally, 53% of temporary lecturers were women in 2002, 64% in 2010, a peak of 86% in 2014 and 76% in 2019 (data for earlier years not available). The distribution of research-intensive senior temporary contracts across genders is instead more equal at all levels (Norwegian Centre for Research Data, 2020).

Method
We accessed the full documentation pertaining to all appointments of new permanent professors and associate professors at the faculty considered in the decade 2007 to 2017. Nearly all these documents are confidential. Our research project was granted advance approval by the Data Protection Service at the Norwegian Centre for Research Data (NSD) on condition of maintaining strict anonymity, and access to the documentation was approved by the competent administration unit of the faculty. We did not transfer information that could render candidates, departments and evaluators recognizable from the material made available to us to the raw data-sheets on which we performed our analysis. Once the approval was in, the administration sent us electronic documents organized into folders, pertaining to a total of 88 open job calls and 4 'direct invitations'; the latter refers to the appointment of one particular candidate to a position created for him or her, without advertisement and competition between applicants. Unlike the US, in Norway there is no tenure-track hiring system: academics work on fixed-term contracts until they are recruited as permanent faculty (Frølich et al., 2018;Kyvik, 2015). We did not include the direct invitations in our sample, because the lack of an applicants' pool renders them not comparable with the open calls (the direct invitations are discussed in Moratti 2020). We were only interested in permanent professorships. We therefore excluded from our analysis 11 open calls (7 lectureships, 3 temporary chairs, and one position that was planned but never got advertised). We retained part-time appointments of at least 0.5 FTE (full-time employment). This left us with a sample of N = 77 open calls, the near totality of which were for associate professorships, reflecting the fact that in Norway the rank of full professor is most often achieved through promotion (Frølich et al., 2019). From our material, we extracted manually information on applicants and winners. An open call usually leads to the appointment of one new full-time professor. Exceptionally, one call can lead to the appointment of two full-time professors (n = 4), one (n = 1) or two (n = 5) part-time professors, or to no appointment at all (n = 7). We counted part-time professors as a fraction of 1, depending on their appointment (for example, 0.5% FTE professors were counted as 0.5). In some instances, the winning candidate declined the job offer and the chair was offered to someone else. However, this is a study on decision-making and academic selection so we did not focus on who eventually filled the vacancy but on who received the offer first. We then proceeded to code each applicant, based on his or her affiliation at the time of application. The coding was as follows: D (Department) for candidates affiliated with the department that opened the vacancy, U (University) for applicants affiliated with a different NTNU department, N (Norway) for candidates with Norwegian affiliations other than NTNU, and I (International) for applicants with non-Norwegian affiliations. 'Guest researchers' were coded as affiliated with their host department, if that was their sole affiliation; conversely, if they had another Norwegian or foreign affiliation, we assumed that to be their primary affiliation and coded them accordingly. For 'job seekers', we looked at their recent employment history. The candidates' affiliations are succinctly stated in the list of applicants that is sent to all candidates soon after the deadline, but these can be inaccurate; therefore, we triplechecked affiliations against the reports from the various committees involved in the selection procedure, then against the applicants' CV when included in the material made available to us. Our coding is exclusively based on the applicants' institutional affiliations: our material did not include systematic information on nationality or ethnicity. We did not investigate the career paths of migrant academics in Norway (Maximova-Mentzoni et al., 2016). Furthermore, we did not collect data on the applicants' names. The bias on national or ethnic origin that names can activate is not the object of our investigation. We also coded applicants by gender. The applicants' list contains each candidate's gender, as stated by the applicant. We were constrained by the features of the online application system that offers applicants only two options as pre-sets in a drop-down menu: man (M) or woman (W). We assumed that applicants' responses reflected their primary gender identification.
The final result of our coding was a data table of 77 rows, one for each open call, and 22 columns reflecting the variables we considered. All numbers in the table are zero or integers, except for part-time job offers. We conducted descriptive statistics on the dataset. Our descriptive statistics included, for each variable, the mean and the standard deviation across calls (N = 77), and total counts as percentages. We then considered separately two sets of variables: variables A describing the composition of the applicants' pool and variables O classifying the candidates who received a job offer. Our 22 variables are listed in Table 3.
We investigated the following questions.
1. Is there a bottleneck holding back women applicants in general, suggestive of gender-specific issues in the selection process? To answer this question, we contrasted the percentages of women among applicants and among winners. 2. Do women apply as often as men, or are they clearly self-selecting? To answer this question, we compared the percentage of women applicants against that of men applicants. 3. Are women less likely to apply, when taking up the job would involve relocating?
To answer this question, we looked at the composition of the applicant pool by academic affiliation, for men and women. 4. Is there a bottleneck for women when they apply as external applicants? We answered this question by comparing the applicants' pool against the pool of winners, for both women and men.

Results
The 77 open calls we retained attracted a total of 1009 applicants; about 42% of them (to be exact, 41.53%; n = 419) were women, indicating that women did not self-select. They applied nearly as often as men did. The 77 calls led to a total of 73.7 job offers; about 46% of these (n = 34) were to women. The total of job offers is not an integer number, because we retained part-time job offers of at least 0.5 FTE. The percentage of women among applicants is roughly maintained among winners, showing that there is no bottleneck for women applicants in general. In fact, there is a small increase which suggests that, in general, women applicants between 2007 and 2017 had a slightly higher likelihood of victory compared to their men competitors. Our data also show that the prospect of relocating does not deter women from applying. The two pools of applicants (men and women) had a very similar composition. For both genders, more than half the applicants were internationals, about one-quarter had Norwegian affiliations other than NTNU, and the rest were NTNU employees, as detailed in Table 6.
The most interesting finding is the presence of an external candidate penalty and an internal candidate bonus for women. While the two pools of men and women applicants are quite similar, there are major differences in the pools of winners. As shown in Table  5, women made up roughly 64% of D winners, 67% of U winners, 36% of N winners and 33% of I winners.
Let us compare the two figures for applicants and winners using clustered stacked bar figures (Figures 1 and 2).
The figures show how most job offers to internal candidates went to women, whereas men received the lion's share of job offers to external applicants. This pattern does not fully reflect the composition of the applicants' pool. Women seem to enjoy an advantage over men when competing as internal candidates, and a disadvantage when competing as external candidates, particularly if internationals. The effect is clearly gender-specific, and it is a function of the distance. The more distant the woman is from the institution, the less likely it is that she will be taken into consideration. Tables 4, 5, 6 and 7 summarize our findings.

Discussion and conclusion
Our study is an investigation of one decade of recruitment of new permanent (associate) professors at one faculty of the largest Norwegian university. Our sample consists of over a thousand applicants. Norwegian academia lives up to its reputation in that we found no bottleneck for women applicants. In other words, there is no indication that women applicants in general were systematically dispreferred in the selection process. Moreover, about 40% of applicants were women, indicating that women did not selfselect: they applied nearly as often as men did. We do not know whether this 40% figure reflects the share of women among potential applicants on the job market. It is not possible to come up with a reliable estimate of all potential applicants during one entire decade not only nationally, but also internationally. However, we accept the assumption that has historically underlain Norwegian gender equality legislation, whereby 'at least 40% of each gender' is taken to indicate a satisfying degree of gender equality. Much of the international literature (reviewed earlier) on application patterns and selection outcomes for permanent academic positions reports that women are either reluctant to apply or systematically dispreferred (or both); that is not what we found. Contrary to the predominant perspective in the literature on gender and mobility, women do not seem deterred by the prospect of relocating: over half of the women applicants in our sample had non-Norwegian affiliations at the time of application, nearly matching their male counterparts. Gender-specific hiring patterns emerge after breaking down data by affiliation and comparing applicants and winners. There is an evident bottleneck for external women applicants (external candidate penalty). The effect is a function of the 'distance'. The more distant the woman is from the institution, the less likely it is that she will be offered the job; this finding resonates with one stream in the literature on gender, mobility and academic hiring, but it has so far been insufficiently and unsystematically investigated. International women applicants are at a major disadvantage. They have worse chances than applicants affiliated with Norwegian institutions other than NTNU, who in turn have fewer chances than NTNU-affiliated women. Women seem to be as entrepreneurial as men, in that they are willing to apply when the job involves relocating; but they are received differently. Women seem to have to work harder than men to develop institutional rapport before they are hired permanently. This calls for a reflection on the genderspecific effects of academic mobility, discussed in the international literature but hardly ever investigated systematically by breaking down a large sample of applicants by gender and academic affiliation. We remind the reader that this is not a study on racism. We could not collect data on race or nationality as our material did not include such information.
At the same time, we found that women internal applicants were at a systematic advantage. This effect cannot be simplistically attributed to 'inbreeding', because it is gender-specific. Rather, it reflects the current tension between the intensified emphasis on internationalization and external hiring on the one hand, and the equally pressing need to keep departments running with their daily tasks (particularly teaching) on the other. As discussed, at the faculty considered, about two-thirds of temporary senior teaching contracts are held by women; these findings are fully in line with the international literature on gender and job tasks distribution in academia. The internal candidate bonus for women may come from the need to secure teaching and an acceptable degree of organizational certainty. In Norway, in recent years, considerable efforts have been made to increase the status of (and the professional rewards for) teaching in academia (Ministry of Education and Research, 2017). While this would be an effective remedy against some of the current inequalities, these efforts should ideally be accompanied by a more equitable distribution of teaching tasks across genders.
In a nutshell, our study shows that underneath an apparently gender-equal picture there are gender-specific hiring patterns that are not the consequence of women's application choices: there is no bottleneck for women applicants in general, but a woman's place is in the 'home' -meaning, the institution where she already works.

Conflict of interest
The author declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research Council of Norway, grant no. 296173 -GENDIM.

Ethics
Authorization for project no. 57034 issued by the Norwegian Research Data Protection Service.