Family Supportive Supervisor Behaviors (FSSB), defined as behaviors exhibited by supervisors that are supportive of employees’ family roles (Hammer et al., 2009), are a critical link in the work-family support—employee outcome chain. In their FSSB measurement development study, Hammer et al. (2009) found FSSB was significantly negatively related to work-to-family conflict, turnover intentions, and significantly positively related to work-to-family positive spillover and family-to-work positive spillover, as well as to job satisfaction, over and above the effects of general supervisor support. And more recently, Odle-Dusseau, Britt, and Greene-Shortridge (2012) found significant relationships over time between employee perceptions of FSSB and reduced turnover intentions, increased job satisfaction, and increased supervisor ratings of employee job performance. Interestingly, Odle-Dusseau et al. (2012) found that while work-family enrichment was a significant mediator of the effects of FSSB on the outcomes over time, work-family conflict failed to mediate any of the effects over time.

In a randomized controlled study, Hammer, Kossek, Anger, Bodner, and Zimmerman (2011) developed and evaluated an FSSB training intervention in which supervisors received both face-to-face and computer-based training. For employees with high levels of family-to-work conflict, the FSSB training led to improved reports of physical health, job satisfaction, and decreased turnover intentions among the employees of those supervisors who were in the training condition. The effects of the training were mediated by employee reports of FSSB, confirming that the training of supervisors lead to increased employee perceptions of FSSB which, in turn, were related to improved job satisfaction and decreased turnover intentions. Further, the main effects of the FSSB training were significant over time on health outcomes although FSSB was not a mediator of the effects of the training on reports of physical health. Finally, meta-analytic literature reviews show that work-family specific supervisor support has consistently been shown to be more strongly linked to work-family conflict than general support, and that supervisor support is the pathway through which employees develop organizational support perceptions (Kossek et al., 2011). Taken together, these findings suggest that the FSSB construct has both theoretical, as well as practical implications for work-life researchers and practitioners. The only validated published measure of FSSB is based on a 14-item tool (Hammer et al., 2009), which has limited use for researchers and practitioners who need abbreviated assessments. As a way to pre-empt researchers and practitioners from using inconsistent and potentially unreliable short forms of the FSSB, the focus of the present study is to develop and validate an FSSB-SF.

Conceptually, the critical importance of FSSB is based in social support theory, suggesting that psychological and instrumental support from significant people in an individual’s role space is critical to providing resources. Such resources enable individuals to better cope with stressors, leading to improved health and well-being outcomes. These relationships have been shown repeatedly with respect to general social support (e.g., Cohen & Willis, 1985), as well as with respect to other various sources of social support such as spouses, friends, coworkers, and even supervisors. The role of supervisor support has been documented in the workplace/organizational literature and more specifically, the role of work-family specific supervisor support has been demonstrated above and beyond general levels of supervisor support in reducing work-family conflict and improving well-being (e.g., Hammer et al., 2009; Kossek et al., 2011). In a recent work-family conceptual model presented by King and colleagues (2013), perceived supervisor support for family plays a major role, along with perceived schedule control, in impacting health and well-being of workers. Additionally, Straub (2012) recently delineated a theoretically based multi-level conceptual framework of the antecedents and outcomes of FSSB to guide future research. Thus, the development of a measure to facilitate further conceptual work in the area of supervisor support for work and family is important.

Hammer et al. (2009) developed the initial FSSB 14-item instrument in response to concerns that prior measures (e.g., Thompson, Beauvais, & Lyness, 1999) confounded the constructs of supervisor support for work and family with organizational work-family culture. Hammer and colleagues argued that providing prescriptive information, or tools, to managers concerning the actual behaviors that were important to enact when being family supportive would be helpful to organizational practitioners. The original FSSB measure was developed to assess four dimensions originally conceptualized by Hammer et al. (2007) (i.e., emotional support, instrumental support, role modeling behaviors, and creative work-family management). We believe that while assessing separate dimensions of FSSB is important, identifying an overall global construct that can be measured with four, instead of 14 items, will be more practical and less costly for researchers to include in their survey instruments. It will also provide a standardized short measure for the field. Indeed, results from the 2009 validation study showed that the four dimensions loaded on one main factor, suggesting this construct operates as a gestalt of the combination of these behaviors into a general perception of family specific supervisor support. Although this prior research was important and contributes to both research and practice, an independent validation of a short form of the FSSB instrument is also needed.

Validating a short form of the FSSB measure would also make it more likely that researchers would include the FSSB measure in future studies (either as a primary or secondary/exploratory variable), leading to increased organizational implications for the training of managers on family supportive behaviors. It is also argued that employees in organizations have been over-surveyed and in turn are less likely to respond to lengthy surveys (e.g., Rogelberg & Luong, 1998). Thus, it is important to develop parsimonious measures of work-family specific support in order to ensure managerial work-family support is mainstreamed into organizational studies.

As part of the FSSB training intervention developed and evaluated by Hammer et al. (2011), the FSSB is used as an assessment tool that is then fed back to supervisors to help motivate transfer of training. The motivation comes from learning that supervisors consistently tend to rate themselves as more supportive than do employees. This information is critical feedback during the FSSB training for the supervisors. Thus, an effort to minimize the number of survey items by developing shorter, psychometrically sound measures could benefit the researcher as well as the practitioner (Stanton, Sinar, Balzer, & Smith, 2002).

Perhaps the most compelling evidence solidifying why a short form of FSSB is important comes from a recent meta-analysis conducted by Kossek et al. (2011) that demonstrates the significant effects of family-specific support in predicting work-family conflict outcomes over more general forms of social support. These effects were evident for both supervisor and organizational level support. That is, family-specific measures of supervisor and organizational support were stronger predictors of work-family conflict outcomes than were general measures of supervisor and organizational support.

It should be mentioned that, to our knowledge, only two other measures of supervisor support for family exist in the literature (i.e., Clark, 2001; Thompson et al., 1999), however both are limited in their application, and neither was psychometrically validated. First, the Clark (2001) measure is only three items and not multidimensional, only capturing the emotional support component of supervisor support for family. Second, the Thompson et al. (1999) measure has been widely used as a more general measure of work-family culture, but has items that confound perceptions of organizational and managerial support for family. Additionally, this measure is 11 items and assesses only emotional and instrumental support, but not the additional two dimensions of support, role modeling and creative work-family management, found by Hammer et al. (2009). See Table 1 for a list of each of the measures and the items for comparison purposes.

Additionally, the development of the FSSB-SF will enable better integration of the work-family and leadership research areas (Major & Cleveland, 2007). For example, building on research on the relationship between leader member exchange (i.e., LMX) and work-family conflict (e.g., Major, Fletcher, Davis, & Germano, 2008), a general measure of how individuals and groups of subordinates rate the family supportiveness of leaders provides valuable diversity climate feedback. Studies show that a positive work-family climate (Kossek, Colquitt, & Noe, 2001) is critical for employee well-being and performance as a group level rating, which can be easily monitored over time.

Below we report on two studies conducted to support the development and use of a new FSSB-SF. Study 1 draws on the original data from the Hammer et al. (2009) study to determine if the results using the 14-item measure of FSSB replicate with the four-item FSSB-SF measure. Study 2 extends the validation of the FSSB-SF to a new sample of professional workers and new outcome variables important to occupational health psychology, in addition to the originally studied variables. Before doing so, we discuss some challenges of short form validation work.

Smith, McCarthy, and Anderson (2000) provide a review of the challenges of short-form development, divided into two general and nine specific “sins,” as well as remedies. These sins are discussed in terms of their foundations in psychometric theory and their prevalence is demonstrated using a sample of 12 published short-form development and validation articles in top psychology journals. Below we draw on Smith and colleagues’ research to examine the strengths and limitations of our short-form validation of FSSB. We introduce relevant sins and recommendations in this article as substantively appropriate and summarize our results relative to this framework in the Discussion Section of this paper.

In Smith et al. (2000), a prominent theme reflecting several “sins” relates to a potential and unintended shift in the content or construct domain between the long- and short-form measures. Hence, the manner in which items are chosen from the long-form measure for inclusion in the short-form measure is of paramount importance. Indeed, blindly selecting only items with the largest item-total correlations or factor loadings may result in an unintended narrowing of the content or construct domain (Little, Lindenberger, & Nesselroade, 1999; Smith et al., 2000). Thus, Smith et al. (2000) recommend the use of both quantitative information and qualitative expert review to mitigate the likelihood of such shifts. We adopted this approach in our selection of items for the FSSB-SF.

Our purpose was to create a short-form measure of the overall FSSB construct rather than its sub-domains because the primary validity evidence presented in Hammer et al. (2009) focused on the overall FSSB construct. One important decision involved the importance of the sub-domains when selecting items. Smith et al. (2000) describe, and are critical of, the development of a short-form NEO-Five-Factor Inventory (Costa & McCrae, 1992) in which the study authors opted to initially ignore the facet information when constructing the short-form measure for each personality dimension. Viewing the FSSB sub-domains as facets, we chose to use these sub-dimensions explicitly in our selection of items to maintain the breadth of the conceptual FSSB domain. Thus, we thought it important to retain one item from each of the theoretically motivated sub-dimensions of the original FSSB (Hammer et al., 2009).

To select the items for the short form, we relied on the content of the items (i.e., based on the item wording), as well as on the first-order factor loadings and error variances (i.e., within each sub-domain), all of which was available in Table 2 of Hammer et al. (2009). We first computed the item information statistic (i.e., squared unstandardized factor loading divided by the error variance; McDonald, 1999) for each item. Within each FSSB sub-dimension we identified the two items with the largest item information statistics. Using the item content information and conceptual definitions of the FSSB sub-domains, we chose the item as the one that was most reflective of its conceptual sub-domain. The selected items were Items 3, 7, 9, and 11 reported in Table 2 of Hammer et al. (2009).

If the FSSB-SF is a sufficiently reliable and valid measure of FSSB, we should be able to replicate the psychometric and validity results on the sample used in the original FSSB development and evaluation article (Hammer et al., 2009). Smith et al. (2000) are rightly critical of basing short-form validation efforts exclusively on samples used to validate the long-form of a measure. We do, however, see this replication of results with these data as important. Indeed, failure to replicate these results on the same sample would certainly be damaging to any claims to the validity of the FSSB-SF. This concern motivated our approach in Study 1. We discuss the limitations of this approach at the end of this section; some of these limitations motivate Study 2 presented later. We therefore see replication of these prior results in Study 1 as being informative but not necessarily conclusive regarding the validity of the FSSB-SF.

Further validation of the FSSB-SF was conducted using a new sample of professional information technology employees from a large U.S. firm, which adds to the generalizability of the measure, as the original measure was validated using a low-wage, hourly workforce. These data were from the baseline assessment of a larger study by the Work, Family, and Health Network on the impact of work conditions on work, family, and health outcomes of workers. We included the same outcome variables that were part of the Hammer et al. (2009) validation study (i.e., work-to-family conflict, family-to-work conflict, work-to-family positive spillover, job satisfaction, and turnover intentions), with the exception of family-to-work positive spillover, which was omitted from the newer study but was part of the first study.

Consistent with the multilevel conceptual framework for FSSB research proposed by Straub (2012), we included four additional theoretically-driven outcomes which enabled us to assess the FSSB-SF’s ability to also predict control over work hours, obligation to work when sick, perceived stress, and reports of family time adequacy. Especially for professionals, a key behavior critical to being family supportive is giving employees more control over work hours (Kossek, Latusch, & Eaton, 2006). In addition, we expect that when a supervisor is supportive of family, he/she is more likely to place less pressure on employees to work when they are sick given they understand the need for self-care in addition to family care, resulting in lowered reports of employees feeling obligated to work when sick. This expectation is also consistent with research by Moen, Kelly, Tranby, and Huang (2011) who found when professionals felt more in control of their work hours they were less likely to work when they themselves were sick. Likewise, we expect that the increased supervisor support for family will be negatively related to perceived stress and positively related to perceptions that one has an adequate amount of family time, consistent with social support theory (Cohen & Willis, 1985) and the conservation of resources theory (COR; Hobfoll, 1989). This suggests that social support at work provides employees with more resources to participate in the family role outside of work. This expectation is consistent with previous research showing that increased organizational cultural support for family is positively related to employee reports of family time adequacy (Kelly, Moen, & Tranby, 2011). Therefore, we hypothesize the following:

To handle occasional missing values in the items underlying each scale, we created mean scale scores only if 75% or more of the scale item responses were observed; otherwise, the scale scores were set to missing. Using this method, missing scale score data were sparse for the variables under study, ranging from no missing data to 20 cases (2.4% of total) with missing values for the Family Time Adequacy scale. Analyses were conducted using Mplus 4.2 (Muthen & Muthen, 2007) using the multi-level add-on package and using maximum likelihood estimation permitting estimation accounting for missing data. In the analyses that follow, α = .05 was used to define statistical significance. Table 5 provides within-supervisor employee descriptive statistics for the study variables.

Drawing on the work of Smith et al. (2000), we believe our approach has addressed many of the “sins” of short-form validation. We used both qualitative and quantitative analysis to select items in developing the short form to prevent an unintended “narrowing” of the construct domain between long and short forms. We used a new sample to cross-validate the short form results in Study 2. Finally, we believe that use of the FSSB-SF offers a substantial decrease (i.e., 10 items or 71%) in survey space and administration time. The net result is a four-item scale that behaves much like the original 14-item scale when the focus is on the overall FSSB construct. It is clear from our development and analysis of the FSSB-SF that the new measure demonstrates significant reliability and validity and offers both researchers and practitioners a more succinct assessment of this critical concept for future use.

The scientific contributions of our paper’s findings add to the nomological network and generalizability of FSSB by demonstrating its significant relationship not only with work-family conflict, work-family positive spillover, job satisfaction, and turnover intentions, as was done by Hammer et al. (2009), but also by extending the outcomes to control over work hours, obligation to work when sick, perceived stress, and reports of family time adequacy. More specifically, employees who rate their supervisors high on FSSB-SF, indicate that they have more control over their work hours, less obligation to work when they are sick, lower perceived stress, and higher reports of family time adequacy. In addition, the findings of this study extend the generalizability of the FSSB assessment to a professional sample of high-technology workers. These results are important particularly for future research as many studies either, examine work-family issues in a very specific sample or job type, or a massive general population survey. We believe our approach of building on the study of work-family supervisor support in specific job contexts and samples is a great method for researchers to see if constructs are replicable across work contexts and job occupations. We showed that FSSB could be captured in a shortened conceptual form across low-income retail worker and in higher income information technology worker samples.

In addition to our own validation findings, Odle-Dusseau et al. (2012) demonstrated that a measure of FSSB was a predictor of critical well-being outcomes, and even more importantly, it was a predictor of supervisory ratings of employee performance over time. These findings suggest that employee reports of FSSB are not only related to employee reports of job attitudes and well-being over time, but the potential beneficial effects of FSSB extend to non-same source performance ratings. Very few studies have linked work-family programs or supports, informal or formal, to employee performance outcomes (for an exception see, Kossek et al., 2001). Furthermore, our findings, and those of most work-family studies to date, have limited ability to draw causal conclusions. There is a need for longitudinal studies and randomized experiments to explore the causal relations among the FSSB-SF and health, well-being, and performance outcomes.

This study, and specifically the development and validation of the FSSB-SF, opens up great potential for future research using multi-level modeling. Straub (2012) offers a multilevel framework linking individual and contextual level factors to FSSB, which in turn is linked to both individual and team level outcomes. Future research should include the FSSB-SF and include ratings from supervisors and employees. Studies can use multi-level modeling of the supervisor level and characteristics to predict individual employee attitudes and behaviors. Group level context studies can also be conducted to access variation in the family supportiveness of work groups, as most work-family policies are enacted at the work group level. Cross-organizational studies can be conducted to assess variation in family specific supervisor supportiveness as an important construct in general organizational behavior, psychological, and management research.

Not only are the theoretical and research implications significant for the continued examination of the FSSB construct that is only enhanced by the availability of a short-form version of the measure, but the practical implications are noteworthy. As can be seen in the Hammer et al. (2011) randomized controlled study, training was developed based on enhancing supervisor FSSB, and this training led to increased employee perceptions of FSSB and in turn increased employee job satisfaction, physical health, and decreased turnover intentions for those employees with high levels of family-to-work conflict. Additionally, we would expect that future FSSB training has the potential of impacting employee performance, as suggested by Odle-Dusseau et al. (2012), as well as the additional family and well-being outcomes examined in the present validation study (i.e., work-family conflict, work-family positive spillover, control over work hours, obligation to work when sick, perceived stress, and reports of family time adequacy). Given that the suggested training by Hammer et al. (2011) involves an assessment of supervisor and employee FSSB, the FSSB-SF provides a more practical assessment instrument for organizational practitioners in the future.