Perceived Family and Partner Support and the Work-Family Interface: A Meta-analytic Review

Introduction

Families, communities, nations, and civilizations must both provide for and nurture their constituents. This is essential for present and future generations to survive and thrive. It is not surprising that research on the interface of paid work (providing) and family life (nurturing future generations) is prolific (Hill & Holmes, 2018). The implicit goal of work-family/work-life research is to understand conditions in which the interface of paid work and personal/family life can be less conflictual and more harmonious (Hill & Carroll, 2014).

Research about the work-family interface has frequently utilized four measures of how work and personal/family life impact each other. These four measures are work-to-family conflict (the degree to which paid work interferes with personal/familial life), family-to-work conflict (the degree to which personal/familial life interferes with paid work), work-to-family enrichment (the degree to which paid work benefits and enhances personal/familial life), and family-to-work enrichment (the degree to which personal/familial life benefits and enhances paid work) (Hill & Holmes, 2018). We utilize these terms and measures throughout this paper, and note that in line with previous research, single, childless individuals who are living alone typically still consider themselves as part of a family and are thus included in these terms (Casper et al., 2016).

While these four measures are often combined in various ways (e.g., measures of work-family conflict that aim to measure both directions simultaneously, measures of work-life balance that combine aspects of all four, etc.), previous research has demonstrated the utility of separating these measures (De Simone et al., 2018; Nicklin & McNall, 2013). Further, research on this topic has often focused on only one or two of these four sides when investigating various factors that impact the work-family interface. More research is needed that examines these four distinct measures of the work-family interface together. Using meta-analytic techniques, we aim to investigate these four measures, looking specifically at the association between perceived support (both partner and family support together) and each of these four measures of the work-family interface. We aim to further understand differences between measures of partner support and family support and the work-family interface through moderator and subgroup analyses.

Work-Family Conflict

For decades, work-family conflict has been a widely studied phenomenon. The growth of this field seems to have correlated with the rise of dual-earner couples (Zhang & Liu, 2011); as an increasing number of couples strive to successfully navigate family life with two careers, many experience negative spillover between what are sometimes seen as competing domains. According to Greenhaus and Beutell (1985), this conflict often occurs for one of three reasons: 1) “time” spent in one domain takes away from time in the other domain, 2) “strain” or stress experienced in one domain negatively affects one’s ability to perform in the other domain, and 3) “behaviors” performed in one domain negatively affect one’s ability to perform in the other domain. Work-family conflict is associated with poorer work outcomes (e.g., higher job burnout, lower job satisfaction; Chen et al., 2012; Niazi et al., 2019) and poorer individual and family outcomes (e.g., lower marital satisfaction, higher hostility, higher emotional exhaustion; Carroll et al., 2013; Judge et al., 2006; Rubio et al., 2015).

Allen et al. (2015) found that although work-to-family conflict did not differ between countries, family-to-work conflict was higher in countries that were more collectivistic and had a higher economic gender gap. Previous meta-analyses (e.g., Allen et al., 2011; French et al., 2018; Kossek et al., 2011) have found that several factors predict work-family conflict, including demographic and personality variables (e.g., locus of control, neuroticism), work variables (e.g., job autonomy, organizational support), and family variables (e.g., parenting demands). The current study focuses on the family variables of partner support and family support as predictors of the work-family interface. We identified one meta-analysis that investigated the role of family and partner support as part of a larger study on the antecedents of work-to-family conflict and family-to-work conflict. This previous meta-analysis found significant negative associations between both partner and family support and work-to-family and family-to-work conflict and found no significant differences between whether the measure was family or partner support (French et al., 2018). However, we note that this meta-analysis was identified after we had obtained our initial sample of articles to investigate our research questions. In comparing the articles this previous meta-analysis used versus the articles we had identified that met our inclusion criteria, there was only a 17.5% overlap in articles. As such, we deemed that our approach was different and valuable enough to warrant continued investigation into this research question1. Further, we note that no other meta-analysis of which we are aware has examined the effects of both partner and family support on not only work-family conflict but also work-family enrichment.

Work-Family Enrichment

Research on work-family enrichment has emerged more recently than the research on work-family conflict. In the early 2000s, several researchers began calling for a more balanced approach to understanding the connections between work and family that not only considered the challenges of balancing work and family, but also looked at the potential benefits of being engaged in both domains (e.g., Frone, 2003; Parasuraman & Greenhaus 2002). Greenhaus and Powell (2006) introduced the first comprehensive conceptual model of work-family enrichment, in which they defined work-family enrichment as “the extent to which experiences in one role improves the quality of life in the other role” (p. 73). With the introduction of this framework, research on enrichment has continued to increase. The first meta-analysis focused on the outcomes of work-family enrichment was published in 2010 and found that both work-to-family and family-to-work enrichment were positively associated with job satisfaction, affective commitment to the work organization, family satisfaction, and overall physical and mental health (McNall et al., 2010).

More recently, Lapierre et al. (2018) published a meta-analysis on the antecedents of work-family enrichment. While some of the personal characteristics and factors they investigated were associated with both directions of work-family enrichment, such as work autonomy and coworker support, they found that overall, factors related to work had stronger associations with work-to-family enrichment while factors related to the family had stronger associations with family-to-work enrichment. While their analysis did look at family support among other factors and found significant associations between family support and work-to-family (r = .17) and family-to-work enrichment (r = .40), the study did not separate family support from partner/spouse support. Further, this study also included measures of social support that extended beyond the family in the analysis, thus making it difficult to distinguish the role of the family support from other social support systems2.

Perceived Familial Support

Research on the antecedents of the work-family interface has also emerged more recently, with early research focused primarily on the outcomes of work-family conflict or negative spillover. Along with the move to increase the understanding of work-family enrichment, researchers also began devoting more efforts to understanding the predictors of both work-family conflict and enrichment (Dilworth, 2004; Stevens et al., 2007). The role of social support was quickly recognized as an important factor for reducing conflict and enhancing enrichment. Indeed, Greenhaus and Powell (2006) emphasized the importance of social support in regards to the work-family interface and discussed both the physical and emotional benefits of various types of social support. Most studies have found that support is positively associated with work-family enrichment and negatively associated with work-family conflict (e.g., Lee et al., 2013; Rupert et al., 2012), with some exceptions of nonsignificant findings (e.g., Cinamon & Rich, 2002; Wayne et al., 2019) as well as rare instances in which the association was in the opposite direction of what was expected (e.g., Charles, 2018; Selvarajan et al., 2013).

Research has acknowledged several dimensions of social support that related to the work-family interface. First, support is often divided into two domains: work support (support from within the working environment) and social support (support coming from outside of the work environment). For the purposes for this study, we focus on the latter. There are three types of social support that have been frequently investigated in prior literature: general social support, familial support, and partner or spousal support. General social support can include support from any social system or group, including friends, community or religious groups, family members, and spouses or partners; some questions related to general support are broad enough to include all of these groups, while other questions have been written to explicitly exclude family members and focus on support from outside of the family system. We do not explore this measure in this paper, opting to focus only on the final two measures. Familial support measures often include both extended family members as well as immediate family members, though some measures have been worded to focus specifically on family members within the same residence. Measures of general family support often include, but do not require, a spouse or a romantic partner as part of the family system, usually depending on whether the respondent is in such a relationship. Finally, spousal or partner support measures focus in on support from only the romantic partner.

From a family systems perspective, families are made up of various interdependent subsystems (Broderick, 1993). These subsystems exist and work in a hierarchical nature, in which certain subsystems are more important and more influential to the system as a whole than other subsystems (Cox & Paley, 1997). The marital or romantic partner relationship is one subsystem that typically exerts a strong influence on the other subsystems. Within this context, measures of general family support aim to measure the entire family system, whereas measures of partner support aim to measure a single, but highly influential, subsystem. As such, while we are unable to separate or remove the notion of partner support from the family system as a whole, we can investigate it as its own subsystem. Such an investigation allows us to measure whether there is a difference in the magnitude of associations when we look at general family support (the entire family system) or when we look at the smaller, but highly important and proximal, partner support subsystem. This distinction has yet to be examined in relation to both work-family conflict and enrichment in a meta-analysis.

In exploring the role of social support in the work family interface, it is also important to note that researchers have often separated emotional support from instrumental or tangible support. Emotional support is characterized by caring and supportive words and behaviors, such as verbal encouragement, whereas instrumental support is characterized by physical contributions and assistance, such as taking care of the children to allow one’s partner time to complete their work responsibilities (Beehr, 1985; King et al., 1995). Both instrumental and emotional support are correlated with life satisfaction and job satisfaction (King et al., 1995).

Moderators

In the current study we will test whether five study-level differences moderate the associations between perceived support and the work-family interface. First, to assess publication bias, we will examine whether the associations vary based on the publication status of the studies included in our meta-analysis (Card, 2012). The second moderator we will test is women dominant samples (vs. more gender-equal or male-dominant samples). Given that participants’ experiences of the work-family interface may be influenced by gender role expectations (Miller & Bermúdez, 2004; Yavorsky et al., 2015), it is possible that the role of support may be stronger in women dominant samples wherein spousal and familial support is seen as a gift but not a given. On the other hand, in samples with more male participants, spousal and familial support may be more of an expectation (i.e., all family members believe that wives and children should support husbands’ careers) and thus be less impactful to participants’ experience of the work-family interface (Nasurdin & Hsia, 2008). Third, we will examine whether the education level of the sample moderates associations, as education has been linked to differences in work-family outcomes in previous research (e.g., Blanch & Aluja, 2009; Lunau et al., 2014; Zhang & Liu, 2011). Fourth, as having children creates additional demands for parents to navigate as they strive to balance work and family demands, we will test whether samples containing a large majority of parents differ from samples in which there are fewer parents. Finally, given differences in work-family outcomes across cultures (e.g., Adisa, 2021; Hassan et al., 2010) we will examine whether the associations differ for Western vs. Nonwestern samples.

The Current Study

Previous research testing the effects of family and partner support on the work-family interface has produced results with effect sizes of varying magnitudes. A meta-analysis can aggregate dozens of studies to elucidate these associations. While previous analyses (i.e., French et al., 2018; Lapierre et al., 2018) found that family support was positively associated with work-to-family enrichment and negatively associated with work-family conflict, a more complete meta-analysis is needed that a) utilizes more effect sizes, b) also examines work-to-family conflict and family-to-work conflict, and c) examines both conflict and enrichment together. The aim of the present study is to systematically integrate research related to perceived partner and familial support and the work-family interface. Based on the literature reviewed, we propose the following hypotheses and research questions:

• H1a: There will be significant negative associations between perceived support (both family and partner support together) and work-to-family conflict and family-to-work conflict.
• H1b: There will be significant positive associations between perceived support (both family and partner support together) and work-to-family enrichment and family-to-work enrichment.
• H2: The magnitude of effect sizes will be larger between perceived support (both family and partner support together) and the family-to-work outcomes than between perceived support and work-to-family outcomes.
• H3: Whether the measure of support is focused on partner support or family support will significantly moderate the associations between perceived support and the four measures of the work-family interface. Specifically, we expect that family support will have stronger associations with the four measures of work-family interface than partner support given the larger support system. In addition to these hypotheses, we also propose the following exploratory research question:
  • RQ1: Are there study-level differences (i.e., publication status, women dominant sample, highly educated sample, parent dominant sample, and Western or Nonwestern sample) that moderate the associations between perceived support and the work-family interface?

Method

Inclusion Criteria and Search Procedure

To be included in this review, studies were required to meet the following criteria. First, the study needed to include a measure of work-family conflict or enrichment that specified a direction of effect (i.e., work-to-family conflict (WFC), family-to-work conflict (FWC), work-to-family enrichment (WFE), or family-to-work enrichment (FWE). Studies that did not specify a direction or combined elements of work-to-family and family-to-work conflict or enrichment were excluded. Second, the study needed to include a measure of perceived familial, partner, or spousal support, where the individual of interest answered a subjective question(s) regarding how much support they perceived they received from either their family in general or their partner or spouse specifically. Measures that included family members alongside other support systems (e.g., friends and coworkers) were excluded. We sought both published articles, meaning articles published in peer-reviewed journals, as well as unpublished articles and data, meaning results that were not in peer-reviewed journals, including dissertations and conference proceedings. The inclusion of unpublished findings is recommended in meta-analyses to test and account for publication bias (Card, 2012).

To obtain these studies, a database search was conducted in March 2021 using the following databases: PsycINFO, Academic Search Premier, OpenDissertations, and Psychology and Behavioral Sciences Collection3. The same keywords were put into Google Scholar where the first 120 results were reviewed (the point at which no additional relevant results seemed to appear). Several additional studies were also obtained through looking at the references of highly relevant articles (backwards searching) and by looking at citations of the most relevant articles (forward searching). Overall, this resulted in 548 articles. For each of these articles, the title and abstract were reviewed to assess whether they met the inclusion criteria. For those that appeared relevant, a full text of the article was obtained. Those that included appropriate measures of work-family conflict or enrichment and perceived family or partner support were included and those that did not were excluded (see Figure 1 for a breakdown of this process).

Figure 1

Search Procedures. Figure adapted from Page et al., 2020.

Coding

The following information was coded from each report: percentage of women in the sample, the mean age of the sample, the average number of children in the family, weekly hours worked, the percentage of the sample that were parents versus not parents, the percentage of the sample that was married or cohabiting versus single at the time of the study, the percentage of the sample that was in a dual-earner relationship versus not in a dual-earner relationship, the percentage of the sample that had a college degree versus less than a college degree, the location of the study, the measures of support used, and the measures of work-family interface that were used. Given the large amounts of missingness on number of children, relationship status, and dual-earner relationship status, these sample characteristics are not discussed further. Further, given variations in reporting styles for other variables, we dichotomized the percentage of parents into parent dominant (1, more than 70%) vs. not parent dominant sample (0), percentage of women into women dominant (1, more than 70%) vs. not women dominant sample (0), and whether the sample was from a Western country (1) or a Nonwestern country (0).

All coding was completed by the first author in a limited period of time to reduce coder drift. Approximately 15% of the studies were also coded by the second author in order to calculate interrater reliability. Across all of the categories coded for, agreement between the two coders ranged from 90-100%, with an average agreement of 97.7% across all of the categories. All differences were discussed by the authors until mutual agreement was reached.

In coding the effect sizes, there were several instances in which the effect size included in the current study reflects a weighted average. Most commonly, there were several studies that reported separate effect sizes for associations between instrumental family/partner support and emotional family/partner support and the various measures of the work-family interface. In these situations, the two effect sizes were averaged to obtain a single effect size. There were several instances where effect sizes were reported separately for men and women; in these instances, given the differing sample sizes, a weighted average was calculated. Further, there were two instances where the study reported effect sizes for both husbands and wives in dual-earner relationships. The husbands and wives scores were averaged together to account for the interdependence between them.

Key Variables

Across the studies, there were a variety of measures used to conceptualize support. The most common measure of support, comprising approximately a quarter of the studies, was King et al.’s (1995) Family Support Inventory for Workers, or various adaptations of it, which measures perceived instrumental and emotional support. Other common measures of support included measures from Caplan et al. (1980), Procidano and Heller (1983), and O’Driscoll et al. (2004). Overall, there was more uniformity in the measures of work-to-family conflict and family-to-work conflict, with over a third of the studies using Netemeyer et al.’s (1996) work-family conflict and family-work conflict scales and approximately a quarter of the studies using Carlson et al.’s (2000) multidimensional measure of work-family conflict or adaptations of these scales. The most common measure of work-to-family and family-to-work enrichment was Carlson et al.’s (2006) work-family enrichment scale, or adaptations of it.

Analytic Plan

Meta-analytic analyses were completed in STATA version 16.1 using the META package. First, using random effects maximum likelihood (REML) models, forest plot summaries were obtained for each of the four outcome variables using the weighted standardized r (Zr; see Tables 1-4). The Zr was used as r is bounded from -1 and 1 and therefore might have a distribution that is inappropriate for the analyses (Card, 2012). For ease in interpreting results, Zrs were converted back to rs in text of this article and are reported in Table 5. For each of the four outcome variables, there was significant heterogeneity, suggesting that a random-effects model was appropriate (Card, 2012). Further, the random-effects model allows us to view the effect sizes included in this study as a sample of the population of effects, and thus generalize our findings beyond the studies included in this analysis. Meta regression was used to probe whether various sample-level variables explained heterogeneity in the effect sizes by assessing whether there were statistically significant differences in effect sizes between groups. Subgroup analyses were used to further understand how the groups investigated (i.e., partner vs. family support, publication status, women dominant sample, highly educated sample, parent dominant sample, and Western or Nonwestern sample) differed from each other.

Table 1

Forest Plot of Work-to-Family Conflict Effect Sizes.

Note. Effect sizes pictured are the standardized Pearson’s r coefficients (Zr).

Table 2

Forest Plot of Family-to-Work Conflict Effect Sizes.

Note. Effect sizes pictured are the standardized Pearson’s r coefficients (Zr).

Table 3

Forest Plot of Family-to-Work Conflict Effect Sizes.

Note. Effect sizes pictured are the standardized Pearson’s r coefficients (Zr).

Table 4

Forest Plot of Family-to-Work Conflict Effect Sizes.

Note. Effect sizes pictured are the standardized Pearson’s r coefficients (Zr).

We first assessed publication bias using Egger’s regression to detect small sample bias (Card, 2012; Lin et al., 2018). In assessing publication bias it is also important to note that for many of the studies included in this analysis, perceived support was not the primary focus of the study, but rather was in some cases a control variable or was tangential to the primary research question. As such, we coded for whether support was included in the title of the article (1 = support included, 0 = support not included) and used meta-regression to assess whether referencing support in the title was a significant moderator of the effect sizes. Our expectation was that if perceived support was the main focus of the study, we would see larger effect sizes among these studies if there was indeed publication bias.

Results

Descriptive Information

A total of N = 36,226 participants were included from the 82 studies, with an average sample size of N = 442 and a range of 57 to 2,759 participants (SD = 504). Across the studies, the mean age of participants was 38.7 years (SD = 5.60) and the samples were on average 60.7% women. Participants came from a total of 24 countries; 68.8% of the samples were from Western countries.

Central Tendencies and Heterogeneity

Across the four outcome variables, using random effects modeling, we found effect sizes significantly different from zero of varying magnitudes. Tables 1-4 display forest plots showing weighted mean effect sizes and heterogeneity statistics for the association between support and each of the four outcomes, respectively. In interpreting the magnitude of these effect sizes, as suggested by Card (2012), we utilize Cohen’s (1969) benchmarks of r = ± .10 as representing small effect sizes, r = ± .30 as representing medium effect sizes, and r = ±.50 as representing large effect sizes.

In support of H1a, familial/partner support had a small negative association with work-to-family conflict (r = -.099, SE = .064, CI4 [-.125, -.074], p < .001; tests of heterogeneity - I² = 75.9, Q(68) = 298,69, p < .001; see Table 1) and a medium-small negative association with family-to-work conflict, r= -.178, SE = .065, CI[-.2.11, -.150], p < .001; tests of heterogeneity - I² = 81.43, Q(33) = 261.66, p < .001; see Table 2. Supporting H1b, there was a small positive association between familial/partner support and work-to-family enrichment, r = .173, SE = .059, CI = [.123, .222], p < .001; tests of heterogeneity - I²= 83.59, Q(24) = 137.58, p < .001; see Table 3, and a medium positive association with family-to-work enrichment r = .378, SE = .058, CI = [.316, .440], p < .001; tests of heterogeneity - I² = 91.48, Q(26) = 389.60, p < .001; see Table 4. Consistent with what we predicted in H2, associations were stronger for family-to-work outcomes than work-to-family outcomes. Across all four outcome variables there was significant heterogeneity, indicating the need to investigate moderating factors.

Moderation Analyses

Meta regression was used to assess moderation. The following five dichotomous sample level variables were regressed on to the four outcome variables: whether the support measure was family or partner specific, whether the majority (over 70%) of the sample were parents, whether the majority (over 70%) of the sample were women, whether the sample was from a Western or a Nonwestern country, and whether the study was published or unpublished. Given the large amount of missing data on the education measure, which led to listwise deletion of effect sizes for which this code was unavailable, it was excluded from the overall analysis and was assessed separately. In partial support of H3, results showed that whether the support measure was family focused or partner focused was a significant moderator of work-to-family conflict (B = -.088, SE = .027, p = .001) and of family-to-work conflict (B = -.088, SE = .033, p = .008), where the association between partner support was weaker than the association between family support and work-to-family conflict. However, whether the measures was partner or family focused was not a significant moderator for work-to-family or family-to-work enrichment.

In investigating RQ1 we found that none of the other moderators were significant for any of the four outcomes. Following these analyses, education (measured as whether 70% of the sample had a college degree or not) was added to the regression given the high amount of missingness on this variable. It was not a significant moderator for any of the four measures the of work-family interface (WFC: B = -.011, SE = .039, p = .788; FWC: B = -.035, SE = .037, p = .341; WFE: B =.119, SE = .074, p = .109; FWE: B =.066, SE = .107, p = .537).

Assessing Publication Bias

We investigated publication bias using Egger’s test for small study effects (Card, 2012; Lin et al., 2018). The test was not significant for any of the four measures, suggesting that these results were not influenced by publication status (WFC: B = 0.28, SE = 0.588, p = .629; FWC B = 0.40, SE = 0.697, p = .566; WFE: B = 0.52, SE = 1.414, p = .711; FWE: B = -1.79, SE = 1.701, p = .294) .As many of our studies included family/partner support as a peripheral or tangential focus to the primary focus of the study, we further assessed publication bias by coding whether the title explicitly referenced support and then used meta regression to see if this moderated the effect sizes. Referencing support in the title was a significant moderator for work-to-family enrichment (B = .169, SE = .053, p = .001), with larger effect sizes coming from studies that mentioned support. However, as there were only four studies that referenced support in the title for this outcome variable and other tests of publication bias were not significant, we are cautious in interpreting this, but acknowledge that it may suggest that publication bias is influencing the results. Across the other three outcome variables, referencing support in the title of the article was not a significant moderator (WFC: B = .002, SE = .026, p = .929; FWC: B = .048, SE = .031, p = .120; FWE: B = -.001, SE = .078, p = .988).

Table 5

Subgroup Analyses Effect Sizes and Heterogeneity

Note. *Signifies that the effect size is significantly different from zero. WFC = work-to-family conflict; FWC = family-to-work conflict; WFE = work-to-family enrichment; FWE = family-to-work enrichment.

Discussion

In this meta-analysis, we examined the effects of perceived partner support and perceived family support on work-to-family conflict, family-to-work conflict, work-to-family enrichment, and family-to-work enrichment. To calculate these effects, we utilized 183 effect sizes from 78 articles (82 unique samples), with reports from a total of 36,226 participants. Although most of the studies we utilized focused on conflict (69 studies included WFC measures and 62 included FWC measures), we were still able to find a fair number of studies focusing on enrichment, the majority of which were published in the last ten years (25 studies included WFE measures and 27 included FWE measures). This seems to align with a general trend in the work-and-family field: many researchers are shifting their focus to also include the positive aspects of the work-family interface (e.g., work-to-family enrichment and family-to-work enrichment; Frone, 2003; McNall et al., 2010; Parasuraman & Greenhaus, 2002). Relatedly, while we interpret these findings together, it is important to note most of the work-family enrichment and work-family conflict effect sizes came from different studies; indeed, only 15 of the studies that met our inclusion criteria included effect sizes between perceived support and all four of the work-family outcome measures. Several noteworthy findings emerged from our analysis.

First, overall familial support (collapsing partner and family support together) was negatively associated with both work-to-family conflict and family-to-work conflict, and positively with both work-to-family enrichment and family-to-work enrichment. Together, the 82 studies show that perceived support from family benefits employees at home and at work, while lack of perceived support from family harms employees’ home life and work life. It is also noteworthy that the effect of support on work-to-family processes (i.e., small negative effect on work-to-family conflict, medium-small positive effect on work-to-family enrichment) appears to be smaller than the effect of support on family-to-work processes (i.e., medium-small negative effect on family-to-work conflict, medium positive effect on family-to-work enrichment). Thus, it is possible that support is more impactful for employees’ success at work than their success at home. This is consistent with Lapierre et al.’s (2018) findings which showed that family factors typically had stronger associations with family-to-work enrichment than with work-to-family enrichment.

Parsing out differences between measures of perceived partner support and perceived family support, we found that the effect size for partner support was significantly smaller than the effect size for family support work-to-family conflict and family-to-work conflict. Indeed, looking at the subgroup analyses, the effect of “partner” support on work-to-family conflict was nonsignificant (see Table 5). On the other hand, “family” support remained a significant predictor of all four outcomes. Thus, the 82 studies show that while “family” support affects both conflict and enrichment processes, “partner” support may be less important in conflict processes. This finding is in contrast to French et al.’s (2018) meta-analysis, even after incorporating many of the same studies they included, showing the value of continued meta-analyses that utilize diverse search techniques. We note, however, that for participants with a partner, family support would “include” partner support but would more fully capture support from the “entire” family system. Further, we highlight the importance of partner support in enrichment processes, in which there were no significant differences between family and partner support measures and enrichment outcomes, suggesting that while this is a smaller support system, the partner subsystem appears to be particularly salient in fostering work-family enrichment processes.

Finally, in testing other factors as moderators of the associations between support and the four outcomes, overall, we found no significant differences. This was surprising. The lack of significant moderating variables (whether the majority of the sample were parents, whether the majority of the sample were women, whether the sample was from a Western or a Nonwestern country, and whether the study was published or unpublished and whether the sample was highly educated) suggest that the effects of perceived familial support are fairly consistent across groups. However, future research should explore other potential moderators of these associations.

Implications

Practitioners (e.g., psychologists, marriage and family therapists, social workers, etc.) and workplace policy makers can apply these findings to help employees succeed in both their home life and work life. To reduce work-family conflict, practitioners and policy makers should focus on “family” support for the best results. To increase work-family enrichment processes, focusing on improving either general family support or focusing on partner support specifically will likely produce similar results. Additionally, perceived familial (family and partner) support seems to be especially impactful for improving work outcomes. Given that support was found to predict higher family-to-work enrichment and lower family-to-work conflict, a major takeaway of this study is that employers should pay heed to employees’ support network. Employers may want to invest in strategies and policies that not only increase employees’ work dedication but also that get the employees’ partners and families on board. Employees’ home lives can spill over into work life either positively (i.e., family-to-work enrichment) or negatively (i.e., family-to-work conflict). Thus, employers may also consider investing in policies which promote a healthy, successful home life (e.g., parental leave, paid time off, flexible hours, option to work from home, etc.).

The results suggest that as practitioners help employed clients find success at home and at work, increasing clients’ perceived family and partner support would be beneficial in reducing negative spillover and optimizing positive spillover. Practitioners can 1) help clients identify and better utilize existing support, 2) help clients increase the support they receive, and 3) help clients change their perceptions of the support they receive. This third point may be especially impactful and achievable, as one of the primary goals of many therapists is to help clients recognize their thought patterns and behavior patterns and how those are interrelated (e.g., in cognitive behavioral therapy). Perceptions can help shape our reality, and as such, perceptions can be just as important as reality in shaping outcomes (LeBaron et al., 2017; LeBaron et al., 2020). By helping clients develop positive perceptions of their support network (e.g., that they can turn to family members for support, that their partner or other family members are on their side and rooting for their success, that their family will give them the support they need, etc.), practitioners may help decrease clients’ work-to-family conflict and family-to-work conflict and increase clients’ work-to-family enrichment and family-to-work enrichment.

Strengths, Limitations, and Future Directions

We recognize the random-effects modeling as a strength of this study which allows us to extend the findings of this study to articles beyond those include in this review. Further, our investigation of work-to-family conflict, family-to-work conflict, work-to-family enrichment, and family-to-work enrichment in the same study is an important strength of this article. Despite these strengths, there are also several important limitations. First, we note that despite the various moderating factors we investigated, there was still significant heterogeneity across the four outcome variables, suggesting that there are additional moderating factors that we were not able to account for. Future research should continue to investigate additional factors that may moderate the associations between family and partner support and work-family outcomes.

The measures of support we looked at were the individuals’ “perceptions” of the support they receive. More research is needed that also looks at more objective measures of support, including obtaining information on how family members believe they give support to each other. Only two of the 78 articles we reviewed included data from more than one family member. Future research should strive to include multiple perspectives and to integrate these perspectives in multi-group actor partner interdependence models to better understand the associations between family and partner support and work-family outcomes. Finally, as the large majority of the studies we reviewed did not separate instrumental support from emotional support, we chose to combine these measures of instrumental and emotional support when they were reported separately. Thus, while it was beyond the scope of the present study to analyze these measures of support separately, future research should work to meta-analytically understand differences between emotional and instrumental support and the work-family interface.

Conclusion

This study makes a valuable contribution to the contemporary milieu of research on the work-family interface. Using meta-analytic techniques, this study examined 183 effect sizes from 82 unique samples to understand associations between partner and family support and the four most salient measures of the work-family interface. Of special note, results showed that overall family support appeared to be more important in protecting against conflict between work and family than partner support while there were no significant differences between partner and family support for work and family enrichment processes. Of the four outcomes, family and partner support appeared to be particularly important in increasing family-to-work enrichment. This study suggests that focusing in on how to garner greater partner and familial support for individuals in their paid work is key to reducing work-family conflict and increasing work-family enrichment. Focusing on this may better enable individuals, families, communities, and nations provide for and nurture their members in harmony in order to survive and thrive.