Employee Rights and Employment Policy Journal

THE FAMILY AND MEDICAL LEAVE ACT: ASSESSING THE COSTS AND BENEFITS OF USE

BY:
HOLLY B. TOMPSON* AND JON M. WERNER**

THE FAMILY AND MEDICAL LEAVE ACT: ASSESSING THE COSTS AND BENEFITS OF USE

[P.125]In this study, 430 human resource (HR) managers were surveyed concerning their perceptions of costs, benefits, and compliance with the Family and Medical Leave Act of 1993. In addition, a model was developed and tested which predicts usage of the FMLA. From 1993- 1995, a mean usage rate of 3.54% per organization was obtained, and our model of usage received general support. Top management sup-port, in particular, emerged as a strong predictor of FMLA use. HR managers perceived employee retention to be among the greatest ben-efits of the FMLA and lost productivity to be the most significant concern.

The Family and Medical Leave Act (FMLA) or the (Act) is the most comprehensive U. S. federal legislation to date dealing with work and family concerns. 1 Since the Act took effect on August 5, 1993, organizations with 50 or more employees 2 have been required to al-low up to 12 weeks of unpaid leave to employees in a number of situa-tions, such as to care for a new 3 or adopted child 4 or an ill family [P.126] member, 5 or to recover from a serious personal illness. 6 Employee health benefits must be maintained during such leave, 7 and employees must be reinstated either to their same jobs, or to ones equivalent in pay and responsibilities. 8 Both private and public-sector employers are covered by the Act. Employees who have worked at least 1,250 hours in the previous 12 months are eligible for such leave. 9 Even though only approximately 11% of private U. S. business establish-ments are covered by the Act, these organizations account for almost 60% of U. S. private-sector employment. 10 Additionally, between 18 and 19 million public sector employees are also covered by the Act. 11 In total, it is estimated that roughly 50 million U. S. workers are cov-ered by the FMLA. 12

Given the number of workers covered by the FMLA, it isn't sur-prising that its passage was hotly debated. Proponents of the Act highlighted advantages to both employees (e. g., time to bond with a newborn) and employers (e. g., improved recruitment vis a vis smaller firms) which they believed would be associated with the enactment of the FMLA. Opponents of the Act were quite vocal as well. For ex-ample, prior to enactment, critics of the FMLA argued that the costs of recruiting and training replacement workers for employees on leave would be substantial. 13 The cost of maintaining health care coverage for employees on leave was a concern, as was the potential for em-ployees to take FMLA leave and then not return to work after-wards. 14 At present, the extent to which these "revolutionary" [P.127] benefits and "daunting" costs have actually been realized by employ-ees and organizations is not well documented. However, managers should now be in a position to assess which of the proposed benefits affected productivity and morale in their organizations and which were mostly hype with little substance. Therefore, the primary pur-pose of our study was to assess the impact of the FMLA in terms of its costs and benefits as perceived by current human resource managers.

Recent research sponsored by the Society for Human Resource Management (SHRM) has attempted to document specific costs asso-ciated with the FMLA. Results of their surveys suggest that the big-gest problem from the perspective of employers may be that of intermittent leave. 15 The FMLA allows employees to take leave in increments of as little as an hour at a time for any serious medical condition. 16 Human resource professionals expressed concern about the administrative difficulties associated with intermittent leave, and also with the apparent vagueness of the FMLA's definition of "seri-ous." 17 According to the Act, a serious illness is "an illness, injury, impairment or physical or mental condition that involves . . . inpatient care in a hospital, hospice, or residential medical facility or . . . contin-uing treatment by a health care provider." 18 While employees are re-quired to request leave 30 days in advance, 19 except in emergencies, employees are reported to be asking that absences still be covered under the Act after the fact of leave taking, and without prior agree-ment. This suggests that many employers are experiencing difficulties accommodating the provisions of the FMLA with their existing bene-fits packages and policies.

In contrast to these findings, however, the SHRM report also in-dicates that most businesses have incurred relatively few costs to com-ply with the Act, and that only one in five respondents to the survey would repeal the FMLA. 20 This is similar to the findings provided in the Final Report of the Department of Labor's Commission on Family and Medical Leave (Commission on Leave), which will be discussed [P.128] below. 21 Further, to counter arguments concerning the "vagueness" of the legislation, Howard Ostmann, FMLA team leader with the De-partment of Labor's Wage and Hour Division, stated that the final rules (issued in January of 1995) clarify that it is the employee's re-sponsibility to keep the employer informed of reasons for and details surrounding a covered absence. 22

The extent to which the costs and benefits of the FMLA are real-ized is, to a large degree, determined by how many employees use the FMLA. To date, however, little empirical research has been con-ducted concerning which factors are associated with usage rates. In light of this deficiency, a second purpose of our study is to look for potential antecedents of employee usage of the FMLA. Given the paucity of data concerning the effects of family leave policies in gen-eral, and the FMLA in particular, such research is critical for FMLA-covered organizations and their employees, as well as for the ongoing public policy debates concerning mandated family leave. 23

Recent academic research on work-family initiatives has argued that organizations adopt work-family innovations (such as increased child care or family leave options) because of pressures from their institutional environments, as well as because of the strategic choices and "dominant logics" held by top management. 24 Institutional pres-sures to adopt work-family programs can be classified as mimetic (i. e., modeling the initiatives of successful competitors), normative (i. e., arising out of HR professionals desire to improve both productivity [P.129] and morale), or coercive (i. e., responding to external governmental or societal pressure). 25 The FMLA can be seen as a coercive pressure on organizations, in that covered organizations are mandated to provide unpaid leave to qualified individuals. Kossek et al. argue that coercive pressures may be less influential than mimetic or normative influ-ences, as coercive pressures are most likely to be viewed negatively and resisted. 26 Most of the recent research on work-family initiatives has utilized data that was collected prior to enactment of the FMLA, and has thus captured voluntary innovations undertaken by organiza-tions. 27 Most of this research has surveyed HR professionals since, as argued by Milliken et. al., these individuals play a critical role in "no-ticing, interpreting, and responding to environmental change[ s]," and are therefore crucial to implementing work-family policies concerning child care, leave, or flexible work options. 28 We extend this line of research to a mandated (nonvoluntary) work-family initiative, namely unpaid leave. Further, Morgan and Milliken found the lowest number of work-family initiatives reported in the southwestern region of the United States. 29 We conducted our surveys specifically in this geo-graphic region, to see what effect mandated leave has had in a region that previously reported below average amounts of such policies. Fi-nally, Kossek et al. found that HR professionals own attitudes toward child care, as well as their perceptions of top executives' attitudes pre-dicted the number of work-family initiatives reported. 30 When leave has been mandated, it might be argued that executive attitudes would be relatively less important (i. e., a mind set of "we do it because the law requires it"). Our study tested whether perceived top manage-ment support would predict amounts of usage for mandated leave.

Overall, then, our study: a) assesses the extent to which employ-ers are complying with the provisions of the FMLA; b) provides fur-ther information about FMLA usage (broken down by industry, percent female, and organization size); 31 c) probes HR professionals' [P.130] perceptions concerning costs and benefits of the FMLA; and d) devel-ops and tests a model of FMLA usage. Each of these areas is dis-cussed next.

EMPLOYER COMPLIANCE AND EMPLOYEE USAGE OF FMLA-COVERED LEAVE

Data have appeared recently concerning employer compliance with the FMLA. A November 1993 survey of 302 employers (mostly from California, which already had a state leave law in place), found that 40% of employers reported being out of compliance with at least one of the provisions of the FMLA, e. g., leave to care for a newborn or newly-adopted child, or an ill child, spouse, or parent. 32 Firms with more than 200 employees were more likely to be in compliance than were smaller firms. Firms where over half the workforce was female were somewhat more likely to be in compliance than firms with less than 50% female workers (compliance rates of 66% versus 58%, respectively). 33

FMLA-related complaints filed with the U. S. Department of La-bor (DOL) increased substantially from 1993 to 1995, 34 and have slowed down some since 1995. According to recent figures available from the Department of Labor, 98 complaints were filed in Fiscal Year 1993 (FY 1993 covered only August and September, the first two months the Act was in effect); 1,881 complaints were received in Fiscal Year 1994, 2,619 in Fiscal Year 1995, 2,534 in Fiscal Year 1996, and 2000 from October, 1996 -June, 1997 (the first nine months of FY 1997). 35 These figures have been attributed to growing awareness of the law by both employers and employees, though management law-yer Gary Mathiason has argued that the overall number of complaints is "not that high" compared to what might have been expected. 36

As with employer compliance, data are emerging concerning the extent to which employees are making use of the Act. Two recent national telephone surveys were undertaken at the request of the Commission on Family and Medical Leave, chaired by Senator Dodd. 37 Both were conducted between June and August of 1995. 38 [P.131] The first, by researchers at the University of Michigan (McGonagle Survey), was a random sample of 2,254 employees. 39 Of these, 378 or 16.8% of all employees reported taking leave within the 18 months prior to their interviews. 40 From this group, 230 (61% of all leave-takers) reported working for an FMLA-covered employer. Leave takers were more likely to be female (58%), between the ages of 35-49 (40.8%), and to have taken leave to address their own serious health conditions (58.8%). 41 An additional group of 76 respondents (3.4% of the total sample) reported needing leave, but not taking it during the covered time period. 42 The majority of survey respondents (1,800, or 79.8% of the sample) reported being employed and neither taking nor needing leave during the covered time period. 43

The second survey was of 1,206 private-sector employers, and was conducted by research firm Westat, Inc. (Cantor Survey). 44 Employ-ers reported that 3.6% of their employees had taken FMLA leave since January 1994. Firms with more than 250 employees reported higher usage (5.3%) than firms with 250 or fewer employees (2.4%). Manufacturing firms reported higher usage (4.4%) than firms in the service or retail industries (3.7% and 2.0%, respectively). 45 Most or-ganizations reported "no noticeable effects" of the FMLA on em-ployee turnover, absenteeism or productivity, although larger firms were more likely to report noticeable effects than smaller FMLA-cov-ered firms. 46 [P.132]

HR PROFESSIONALS' PERCEPTIONS OF FMLA RELATED COSTS AND BENEFITS

Proponents of the Family and Medical Leave Act have argued that employees should not have to choose between family responsibil-ities and their jobs. 47 As a result of various social and demographic changes in the United States, more than 60% of women now work outside the home, more than double the rate in 1950. 48 In the past, women with children under age six were underrepresented in the workforce, compared to other women. Today, "nearly 60% of mothers with children under the age of six are in the workforce." 49 Research evidence is mounting which suggests that the "quality time" children need with parents for healthy development is attained only through a substantial quantity of time invested. 50 While it is primarily those who can afford time off without pay who benefit from the FMLA, supporters of the FMLA argue that it is a step toward provid-ing "quality" interaction from the beginning. Further, as the number of older Americans continues to rise, both male and female employees increasingly need to assist aging parents and other ill relatives. 51

Prior to the passage of this Act, some employers reported that their family leave policies improved their ability to recruit desired em-ployees, reduced worker stress and turnover, and increased their pro-ductivity. 52 In general, however, both large and small employers were opposed to the FMLA, as well as to previously proposed versions of the Act. For example, "a 1988 survey of human resource professionals by the American Society for Personnel Administration [now the Soci-ety for Human Resource Management], found that 83% disagreed with federally-mandated paid parental leave and 69% also disagreed with mandated unpaid leave." 53 Concerns have been raised about a number of issues, including: a) productivity losses and operational [P.133] instability, 54 recordkeeping difficulties, 55 recruiting and training tem-porary workers, 56 and litigation due to the vagueness of the Act. 57 While employers have expressed concerns about costs associated with implementing the Act, these concerns become especially relevant given the difficulty of predicting how many employees will utilize the Act at a given time.

To summarize, a major focus of our study was to probe HR pro-fessionals' perceptions of the costs and benefits of the Act two years after enactment. Among the particular issues we addressed were the following: "How do HR professionals view the Act after having had some time to work with it?", "Do the benefits of the Act outweigh the costs, or vice versa?", "Do HR managers believe that employees in their organization feel they would be punished if they took leave under the Act?" and "Have their employees misused provisions of the FMLA?" These questions are exploratory in nature and, therefore, formal hypotheses may not be warranted. However, given the earlier negative reactions by SHRM members as well as the regional differ-ences detected by Morgan and Milliken, 58 we expected our respon-dents to also perceive more costs than benefits.

PREDICTORS OF FMLA LEAVE USAGE

Before passage of the FMLA, there was considerable debate con-cerning the extent to which employees would make use of the Act's provisions. Many employers were concerned that significant portions of their workforce could take leave at the same time. Therefore, ac-tual usage rates are of great concern to FMLA-covered organizations. As discussed in the Cantor Survey, there has been very little prior research conducted on the topic of leave-taking in general. 59 Little empirical or theoretical work is available to guide research on the predictors of FMLA usage. Thus, our proposed model is clearly ten-tative and exploratory. However, we put it forward in the hopes of stimulating further research in this area. Two categories of variables [P.134] -- organizational-level and individual-level -- are included as primary predictors of FMLA leave usage.

The first category, organizational-level variables, includes top management support for the FMLA, as well as the extent to which paid leave policies were in place in respondents' organizations prior to passage of the Act. Top management support for the FMLA should have a positive effect on usage, in that employees should feel freer to make use of the Act when top management appears accepting of its intent and specific provisions. 60 Conversely, if employees perceive that top management does not support the FMLA and that negative consequences would result, usage rates are expected to decline. Given that the FMLA only guarantees unpaid leave for employees who qualify, we propose that having a paid leave policy in place prior to the enactment of the FMLA will be associated with lower usage rates (i. e., employees who do not have the option of taking a paid leave of absence or maternity leave will be most likely to utilize the FMLA). More formally, this is stated as follows:

Hypothesis 1 (a & b): Usage rates for the FMLA will be greater in organizations where: a) HR professionals perceive greater support for the FMLA from top management and, b) when no paid leave policy existed prior to passage of the Act.

Three individual-level variables -- the extent to which employees have difficulty interpreting the act, employee gender, and employees' beliefs that negative consequences would result if they used the Act -- are proposed to be associated with usage rates. First, the more difficulty employees have understanding the general provisions of the Act and how it applies to their particular situation, the less likely they are to actually utilize it. Conversely, when employees feel confident concerning their understanding of the Act and, therefore, are aware of when their situation qualifies them for the FMLA, they will be more likely to pursue FMLA leave. In other research, employees' ability to confidently interpret the provisions of an Employee Assistance Pro-gram (EAP) has been found to be associated with increased use. 61 Similarly, when employees receive education that helps them interpret EAPs and how the different options apply to their needs, their pro-pensity to use them significantly increases. 62 [P.135]

Despite the increasing participation of women in the workforce, women continue to be responsible for the majority of household and family responsibilities. 63 It follows, then, that women will on average be more likely than men to assume the role of primary caregiver after the birth of a child, or when a family member is ill. Because of such circumstances, it is predicted that women will make greater use of the FMLA. While gender is not likely to be related to leave taking for recuperation from personal illness, women are expected to take ad-vantage of the FMLA more often than men. The McGonagle Survey found that leave takers were more likely to be female, although by a relatively narrow margin, i. e., in their study, 58% of all employees tak-ing leave were female. 64 In a survey completed by Jerry D. Goodstein in 1994, the percentage of female employees was significantly and pos-itively related to the amount of child care and flexible work options, although this same variable did not predict use of flexible work op-tions in a later survey by Goodstein or by Morgan and Milliken. 65

The third individual-level variable, employee perceptions of pun-ishment, is expected to be negatively related to employee FMLA leave usage. That is, when employees believe that FMLA use will re-sult in punishment, whether in the form of a supervisor's disapproval or possible career derailment, this will deter use. While inappropriate uses of punishment have come under scrutiny in recent years, many managers still view punishment (or the threat of punishment) as an effective tool, at least in the short-term, in shaping employees' behav-ior. 66 Therefore, we hypothesized:

Hypothesis 2 (a -c): FMLA usage will be greater: a) when employ-ees are perceived to have little difficulty interpreting the Act; b) for female employees, and c) when employees do not believe that they will be penalized for using the Act.

Recent empirical studies have found strong and consistent indica-tions that the degree to which employees perceive support or lack of support for a program or policy is a reliable proxy for employee be-havior. 67 In this case, Hypothesis 1 proposed top management sup-port for the FMLA as a predictor of employee usage. If this [P.136] relationship is supported, then it is also important to identify the fac-tors that influence management support of the FMLA. 68 While far from a complete list of antecedents, we propose that top management support will be predicted by two of the HR professionals' perceptions, namely the extent to which they perceive significant benefits and min-imal costs associated with the FMLA. Since HR professionals are generally charged with interpreting and implementing laws and regu-lations such as the FMLA, it is expected that their perceptions of FMLA benefits and costs will influence the extent to which top man-agers in the organization are supportive or unsupportive of the Act. Thus, we hypothesized the following:

Hypothesis 3 (a -b): Top management support for the FMLA will be greater when HR professionals perceive: a) greater overall ben-efits, and b) lower overall costs associated with the FMLA.

METHOD

DESIGN AND SAMPLE

Human resource (HR) professionals were surveyed concerning their organizations' experiences with the FMLA, as well as their per-ceptions concerning a number of issues related to the Act. 3,000 mail-ing labels were purchased through the Society for Human Resource Management, and surveys were sent to HR professionals in six south-western states (Arkansas, Mississippi, Louisiana, Texas, New Mexico, and Oklahoma). 69 Our final sample consisted of responses from 430 HR professionals, for a usable response rate of 17%. 70 While low, Alreck and Settle note that "response rates are often only about 5 or 10 percent" for mail surveys. 71 Cover letters were addressed to the [P.137] Human Resource Director at each organization. Respondents were assured of the confidentiality of their individual responses, and only included their names and addresses if they desired a copy of the find-ings once our study was completed. Questionnaires were mailed out in August 1995. This time frame coincided with the two-year anniver-sary of the date when the Act took effect. This allowed us to compare our findings to other surveys conducted recently. 72

QUESTIONNAIRE

The survey requested basic information from respondents such as industry, total number of employees, percentage of female employees, and the number of employees who had used the FMLA in the past two years. 73 Respondents were asked if a family leave policy (e. g., maternity leave) had been available to employees before the FMLA. If so, the average length of leave taken and whether this leave was paid or not was determined.

Questions were included on the survey concerning what respon-dents viewed as the greatest benefits and costs associated with the FMLA. 74 Respondents were then asked to rate the extent to which they thought overall benefits of the Act outweighed costs, or vice versa. Additionally, we questioned HR professionals concerning whether employees in their organization felt they would be punished if they took leave under the Act, and whether employees had misused provisions of the FMLA. [P.138]

Next, respondents were asked the extent to which they agreed or disagreed with several statements concerning specific aspects of the FMLA. 75 Finally, respondents were asked four questions concerning the extent to which their organizations were in compliance with the FMLA, i. e., whether the following steps had been taken in their or-ganization: posting of FMLA information, modification of health in-surance, and policy administration. Response options were as follows: yes = 2, in progress = 1, or no = 0.

DATA ANALYSIS

Descriptive statistics (means, standard deviations, and correla-tions) were used to understand the characteristics of the sample and to gain insight into FMLA usage rates. 76 Multiple regression analyses were used to test Hypotheses 1-3.

RESULTS

Information concerning our sample can be found in Tables 1 and 2. For the first two years the Act had been in effect, the average ratio of leave-takers to number of employees was 3.54% in the total sam-ple. 57.3% of the respondents reported that their organizations had family leave policies in place prior to passage of the Act. Roughly one-third of these organizations (i. e., 20.4% of the sample) offered paid leave. The average length of leave for those providing paid leave was in the category of 6-12 weeks. Responses ranged from 0 to 250 employees taking leave, and the distribution was positively skewed, i. e., 27% of our sample reported no FMLA leave-takers, and an addi-tional 23% reported 1-3 leave-takers.

Reported compliance with the Act was relatively high. For exam-ple, 96% of respondents reported posting FMLA information for em-ployees (2% in progress); 88% had amended their policies and/ or handbooks to comply with the Act (9% in progress); 77% had modi- [P.139] fied their health insurance to cover lapses (5% in progress); and 65% reported providing information to employees about filing FMLA complaints with the U. S. Department of Labor (5% in progress).

TABLE 1
RESPONDENT CHARACTERISTICS AND USAGE RATES BY ORGANIZATION SIZE AND INDUSTRY

Regarding organization size, our sample was roughly equally di-vided between organizations with 50-250 employees, and those with more than 250 employees. The mean number of employees across all organizations in our sample was 584.9 (standard deviation= 986.1). The percent of the workforce that was female was relatively consistent regardless of organization size (mean = 49.8%). Based on the findings of the Cantor Survey, we tested whether organizations with more than 250 employees would indicate greater FMLA usage. This was not the case in the present data, in that a t-test of usage rates for organizations with more or less than 250 employees was not statistically significant (organizations with 250 or fewer employees had a mean usage of 3.45, versus a mean of 3.43 for organizations with 251 or more employees; [P.140]

TABLE 2
DESCRIPTIVE STATISTICS AND CORRELATION COEFFICIENTS (a)

(a) - N=430; the sample sizes for individual correlations ranged from 400-430.

* - p < .05;
** - p < .01;
*** - p < .001

[P.141]

t= .00, ns). 77 Unexpectedly, usage rates were highest (4.6%) for the middle-size category (251-500 employees), followed by the smallest category (4.1% for organizations with 50-125 employees).

In terms of industry affiliation, the largest number of organiza-tions reported their industry as manufacturing, public sector, and banking/ finance. FMLA usage was highest in four industries: non-profit, banking/ finance, service, and mining/ geology, and ranged from 5.9% in nonprofit organizations to 1.5% in professional organizations. Unlike the Cantor Survey, manufacturing organizations reported lower usage (2.8%) than either retail or service organizations (3.3% and 4.9%, respectively). 78 Sizeable differences were observed in per-cent female and FMLA usage by industry. Percent female was above 60% in four industries: nonprofit, banking/ finance, professional, and public sector, and ranged from 72.3% in nonprofits to 6.1% in mining/ geology. However, as seen in Table 1, there was no consistent pattern of relationships between percent female by industry and usage rates. For example, communications firms were above average in their per-centage of female employees, but were next to last in usage, whereas construction and mining had very few female employees, yet were above average in their FMLA usage rates.

Concerning the rankings of various perceived benefits and costs of the FMLA, there was not a clear first choice among respondents in terms of perceived benefits. 79 Greatest benefits were perceived con-cerning employee retention (mean = 2.97) and satisfaction (mean = 3.09), followed by decreasing stress (mean = 3.58), creating a more positive atmosphere (mean = 3.63), and increasing loyalty and com-mitment (mean = 3.81). Respondents ranked encouraging employee reciprocity (mean = 4.56) and use as a recruiting tool (mean = 6.51) lowest. As far as overall costs of the FMLA, respondents' top concern was decreased productivity (mean = 2.09), followed by administrative concerns (mean = 2.78), operational instability (mean = 3.05), and manager training (mean = 3.22). Respondents gave their lowest rank-ings to the difficulty of getting qualified temporary help (mean = 3.85). [P.142]

Results of our regression analyses can be seen in Table 3. 80 Sup-port was obtained for one of the two predictions made in Hypothesis 1. As expected, top management support was positively related to us-age. Contrary to Hypothesis 1( b), existing paid leave was not signifi-cantly related to usage rates. Hypothesis 2 was supported for two of the three predictors. Specifically, greater difficulty interpreting the FMLA was associated with decreased FMLA use. Similar to prior research, we found a moderate relationship between females as a per-centage of the workforce and FMLA usage. 81 As shown in Table 2, percent female correlated r= .16 with FMLA usage (p < .01) 82 , and in the regression equation, percent female was one of three variables that significantly predicted FMLA usage. Contrary to Hypothesis 2( c), employee penalty perceptions were not significantly related to FMLA usage. Altogether, Model R 2 for this equation was .24 (Ad-justed R 2 = .23; F= 3.41, p < .05). 83 Finally, Hypothesis 3 was sup-ported, in that perceived overall benefits were positively related, and perceived overall costs negatively related to top management support. The Model R 2 for this equation was .08 (Adjusted R 2 = .07; F= 9.08, p < .01). [P.143]

TABLE 3
RESULTS OF REGRESSION ANALYSES

* - p < .05;
** - p < .01;
*** - p < .001

DISCUSSION

The HR professionals in our sample saw the greatest benefits from the FMLA in increasing employee retention and satisfaction. This is similar to the findings by the Cantor Survey researchers that companies where employees had taken leave during the covered time period said the FMLA had a greater impact on reducing turnover than did other companies. 84 It appears that the FMLA has permitted many employees to stay in jobs where they might otherwise be forced out, or they might choose to work for a competitor. However, there seemed to be little evidence in our data that employers were seeking to use the leave law proactively, i. e., as a competitive advantage vis-a-vis smaller firms not covered by the Act. Despite considerable discus-sion in the popular press about organizations that are actively market-ing themselves as "family friendly" to potential employees, 85 [P.144] respondents in our study were least inclined to view family leave as a recruiting tool.

Similar to the concerns cited in a report by Michael Davis, re-spondents were most concerned with the impact of the FMLA on pro-ductivity as well as administrative concerns related to the Act. 86 Objective productivity measures were not available to us, but instead this and other variables were measured from the perspective of human resource professionals. Such perceptual measures are impor-tant in their own right, as it is difficult to isolate the precise effects of the FMLA (or any other legislation) on productivity measures. Fur-ther, many employment decisions are based upon top managers' and HR professionals' perceptions of reality. Our respondents viewed losses in productivity to be the greatest cost of the FMLA, even though this point continues to be debated in both the research and political arenas. 87 Further, administrative problems continue to trouble HR professionals, including training and updating current managers concerning the FMLA, issues pertaining to intermittent leave, and clarifying the definition of a "serious health condition." 88 However, contrary to the views expressed by opponents of the Act prior to its passage, 89 in our study, respondents were least concerned with problems getting temporary help while employees were on leave.

Given the importance of top managers to organizational success, and the increasing linkage of organizational strategy and human re-source management, 90 the relationship observed between our percep-tual measure of top management support and a more objective measure of FMLA usage was an important finding from our study. This suggests that management support (or lack of support) for em-ployee programs impacts employee use of such programs. 91 Further, while Kossek et al. found that perceived executive support was predic-tive of the amount of voluntary work-family initiatives undertaken, 92 our results indicate that a similar variable was also significantly re-lated to government-mandated family and medical leave. We also found that top management support was predicted by the extent to [P.145] which HR professionals perceived greater overall benefits and fewer overall costs associated with the FMLA, adding support to the argu-ments of Milliken et al. 93 and others concerning HR professionals' strategic importance for organizational success.

Contrary to our expectations, the existence of a paid leave policy as well as employee perceptions that they would be penalized for making use of the Act were not related to FMLA usage. Concerning paid leave policies, we unfortunately did not ask respondents to state specifically what was covered under their leave policies. There is a wide variation in what various organizations cover with paid leave. 94 One thing we had not anticipated when we began this research study was that roughly 60% of all leave-taking at FMLA-covered organiza-tions would be for an individual's own serious health conditions. 95 Thus, if we assume that the majority of leave-taking in our sample was for employees' own health conditions as well, then it is likely that dif-ferences between organizations in paying for maternity/ paternity leave, or child or elder care were overwhelmed in our data set by dif-ferences between organizations in the manner in which they handle employees' individual health needs. As far as why employee penalty perceptions did not predict FMLA usage, this would clearly be an area where it would be advisable to ask employees directly. We spec-ulate that a direct measure of employee penalty perceptions might capture the impact of this variable more than did our question ad-dressed to HR professionals. 96 We would note that, in Table 2, there was a moderately strong negative relationship between top manage-ment support and employee penalty perceptions (r= -. 34). However, in Table 2, and more important, in the regression analysis shown in Table 3, top management support had the stronger relationship with FMLA usage. As noted above, in our study top management support was the variable that best predicted usage of the FMLA. Future sur-vey research should determine whether employee penalty perceptions also predict usage, as we had expected. [P.146]

With such a large number of employees eligible for FMLA leave, the usage rates reported in our study were relatively low. Our mean usage rate (3.54%) is almost identical to the 3.6% rate obtained by the Cantor Survey. 97 Altogether, approximately half our sample had three or fewer leave takers from August 1993 to August 1995. Some organizational leave policies were reported that were more generous than the FMLA provisions. However, this was only true in about 20% of the organizations in our study.

This raises the obvious question of why usage was so low. Usage would be expected to be slightly lower than those obtained by the Cantor Survey, 98 given the regional findings reported by Morgan and Milliken, 99 and the general pro-business climate prevalent in the Southwestern states where our surveys were distributed. For example, all of the states except New Mexico and Oklahoma have "right-to-work" laws which allow employees to work in unionized organizations without being required to join the union, and which are generally thought to weaken the power of labor unions. 100 In 1995, unionization rates in these states ranged from 5.2% in Mississippi to 9.4% in New Mexico. 101 In contrast, the overall unionization rate for the entire United States in 1995 was 14.9%. 102

It is difficult to make generalizations about our results concerning organization size and industry. It is unclear why the highest usage rates were reported by organizations in the 251-500 employee cate-gory, with lower rates reported in organizations with greater than 500 employees. However, comparisons to the Cantor Survey data suggest that organizations with over 500 employees have greater difficulties administering the Act than organizations with 50-250 or 251-500. 103 [P.147] Further, usage differed widely among industry groups, ranging from 1.5% to 5.9%. We note that our finding of a usage rate of 3.0% for public-sector organizations is somewhat lower than the mean for all organizations. Ours is the first survey we are aware of to report usage rates for public-sector employees (the Cantor Survey surveyed only private-sector establishments). 104 One reason why usage rates would be lower for public-sector employees is the availability of compensa-tory time, i. e., if management agrees that an employee may cover leave with any compensatory time that s/ he has built up, this can be used in lieu of FMLA leave. 105 Overall, differences in usage rates by industry were not related to the percent of the workforce that was female. Further research concerning these and other demographic variables is needed.

Family leave laws have been both hailed and criticized concerning their value for working women with children. 106 Similar to previous research, we found a small positive correlation between percent of the workforce that was female and FMLA usage. We think this modest relationship is understandable, in that more women would be ex-pected to take leave for the birth of a child. The other two major provisions of the Act (personal illness, 107 and illness of a family mem-ber 108 ) are not necessarily correlated with gender. The FMLA was written to apply to both women and men; early indications are that both genders are making use of the Act.

In terms of practical implications, we have provided practitioners with a host of descriptive and predictive information. It is important for managers to carefully present their attitudes concerning FMLA leave taking. Because perceptions of top management attitudes ap-pear to impact employee intentions to take leave, managers should avoid letting negative perceptions of the Act color their conversations with eligible employees. This is true not only because of potential legal ramifications, but also because the quantity and quality of work can suffer if employees do not take leave when it is necessary. [P.148]

LIMITATIONS OF THE RESEARCH / SUGGESTIONS FOR FUTURE RESEARCH

We have emphasized the perceptions of HR professionals. We feel this is appropriate, given the strategic role of such individuals in disseminating information and dealing with new legislation. However, it is also necessary to further survey employees who have taken leave, as well as those who were eligible to take leave, but chose not to. Top management should also be surveyed directly, since their support or lack thereof impacts employees' usage decisions.

This study is limited by the cross-sectional, self-report nature of the research design. Caution is warranted concerning the causal orderings implied by our model of usage. It would be advantageous to use a longitudinal design in future research. In addition, our predictors of usage, although exploratory, are based on single-item measures. Further research is needed to strengthen the measurement of the constructs represented in our model. Finally, our response rate was low. Since most surveys were returned to us anonymously, we were not able to check for differences between respondents and nonrespondents.

This study of FMLA perceptions and usage has demonstrated that HR professionals see both costs and benefits of the Act. Con-cerns continue to be expressed about the burdens of the Act, as well as the tendencies of some employees to view the Act as "an excuse for a day off". 109 Nevertheless, usage of the Act remains relatively low. Some of this can be attributed to lack of awareness of the law by both managers and employees. Yet, it is also important to study percep-tions of the Act (of top managers, HR professionals, and employees). Our study is an initial effort in this regard.

As research continues on work-family conflict, 110 we believe that legal issues such as the FMLA need to be considered as well. Imple-mentation of the FMLA continues to evolve, primarily via interpreta-tions made by the DOL and the courts. 111 To the surprise of many, as of June 1997, a total of twenty legal actions pertaining to the FMLA had been filed by the DOL. Three cases were resolved through litiga-tion, [P.149] three were settled without litigation, three were dismissed, one case participated as amicus, and eight cases were still pending. Fur-ther, 11,795 complaints had been received by the DOL and 9,132 com-pliance reviews completed as of June 30, 1997 (2,663 complaints were still pending). Of these, no violations were found in 42% of the cases, e. g., the situation was not covered by or did not violate the FMLA. However, in the remaining cases, 91% of the violations found had been successfully resolved (i. e., 4,840 successful resolutions out of 5,316 violations found). 112

There appears to be some discrepancy here between the high rate of compliance with the FMLA, and compliance rates for other em-ployment legislation such as Title VII of the 1964 Civil Rights Act. 113 The FMLA has proven to be quite popular with the public. Addition-ally, several bills currently before the U. S. Congress would expand the FMLA, e. g., to cover more businesses, to expand the conditions under which leave may be taken, and to spell out the situations where paid leave is to be used in lieu of unpaid leave. 114

In contrast to this, an FMLA corrections coalition has been formed, led by the Society for Human Resource Management. The coalition has proposed amending the current FMLA to, among other things: narrow the definition of a serious health condition; allow em-ployers to require intermittent leave be taken in half-day increments; eliminate leave benefits for intermittent and reduced work schedules; and revise employee notification requirements. 115 Whatever the out-come of current initiatives before Congress, it is certain that family and medical leave issues will remain controversial and will continue to be widely debated in U. S. national politics.

Prior to the passage of the FMLA, critics warned that the costs of the Act would be substantial. By and large, these fears appear not to have been born out by subsequent experience. 116 The FMLA has pro-vided [P.150] an important safety net for many millions of U. S. employees. However, usage of the Act has been relatively low, and given the high rate of resolution of formal FMLA complaints with the Department of Labor, it would appear that much of the noncompliance has been caused more by uncertainty or ignorance of the law, rather than will-ful violation of it. 117

In the survey of SHRM members taken in January 1997 that was mentioned earlier, many concerns were expressed about the FMLA. However, only 19% of respondents were in favor of repealing the Act. 118 In our survey, respondents saw both costs and benefits associ-ated with the Act, i. e., the means for the items concerning "overall costs" and "overall benefits" of the FMLA were 3.44 and 3.21, respec-tively. Both means were above the midpoints of the scale, and were not significantly different from one another. Taken together with all the anecdotal and survey data collected by the Commission on Leave, this suggests that there are both costs and benefits associated with the FMLA. However, it appears to be very difficult to link direct costs to the Act. We would speculate that the larger category of costs associ-ated with the FMLA are indirect (i. e., increased administrative re-quirements, possible impacts on employee morale). These costs must be weighed against the very real benefits experienced by employees who, in many circumstances, are guaranteed to get their job (or an equivalent one) back after taking covered leave. 119

Employers, HR professionals, and the legal profession all have vital interests in the impact of the FMLA on organizations and their employees. Legal actions are not the only (or necessarily the best) means of resolving work-family disputes. Nevertheless, legal issues should not be neglected by researchers and practitioners interested in work-nonwork conflict. The FMLA provides one avenue available for employees and managers seeking to balance the interests and de-mands of multiple life roles. 120 It is our hope that employers and em-ployees, [P.151] in conjunction with the research and legal communities, will continue to study and develop this avenue, in addition to other more proactive means of attaining work-life balance. Our study, while rela-tively modest in its scope, provides valuable additional information on the impact of the FMLA on organizations and their employees.

FOOTNOTES

* Holly Tompson is a Lecturer in the Department of Strategic Management & Leadership at the University of Wikato, Hamilton, New Zealand. She received her Ph. D. in Organizational Behavior from the University of South Carolina.

** Jon Werner is an Associate Professor in the Department of Management, University of Wisconsin-Whitewater. He received his Ph. D. in Organizational Behavior from Michigan State University.

Please direct correspondence to the second author.

We gratefully acknowledge the support of the University of Central Arkansas Research Council, and the University of South Carolina Riegel & Emery Human Resource Center, for funding this research. This research was begun when Professors Tompson and Werner were at the University of Central Arkansas and University of South Carolina, respectively. Funding for this research from the University of Central Arkansas Research Council and the University of South Carolina Riegal & Emery Human Resource is gratefully acknowledged.

1. 29 U. S. C. º º 2601-54 (1997)

2. Id. º 2611( 4)( A)( i)

3. Id. º 2612( a)( 1)( A)

4. Id. º 2612( a)( 1)( B)

5. Id. º 2612( a)( 1)( C)

6. Id. º 2612( a)( 1)( D). Under the Act, employers may choose among several options for determining the 12-month period during which leave may be taken. This includes use of a calen-dar year, a fixed 12-month period (such as a fiscal year), or "a 12-month period prior to or after the commencement of leave." COMMISSION ON FAMILY AND MEDICAL LEAVE, U. S. DEP'T OF LABOR, A WORKABLE BALANCE: REPORT TO CONGRESS ON FAMILY AND MEDICAL LEAVE POLICIES (1996) [hereinafter COMMISSION ON LEAVE].

7. Id. º 2614( C)

8. Id. º 2614( a)( 1)( A)-( B)

9. Id. º 2611( 2)( A)( i)-( ii)

10. See DAVID CANTOR ET AL., WESTAT INC., THE IMPACT OF THE FAMILY AND MEDICAL LEAVE ACT: A SURVEY OF EMPLOYERS, Submitted to the U. S. Dep't of Labor (Oct. 2, 1995) [hereinafter CANTOR SURVEY].

11. See COMMISSION ON LEAVE, supra note 6.

12. Deborah Billings, DOL Statistics Show Significant Jump in Complaints Lodged Under Leave Law, 13 BNA'S EMPLOYEE RELATIONS WEEKLY 1263 (1995).

13. CANTOR SURVEY, supra note 10, at 1-1.

14. In section 2614( c)( 2) of the FMLA, it is stated that employers may recover premiums that were paid to maintain the health coverage of employees who fail to return from FMLA leave. In an earlier survey of employers, it was reported that in one-third of all work sites, at least one employee failed to return to work after taking FMLA leave. However, these same employers reported very few attempts to recover these premiums, i. e., payment recovery was attempted in only 7.3% of the situations where leave-takers failed to return to work. See CAN-TOR SURVEY, supra note 10, at 4-14 tbls. 4-12, 13, 14.

15. Alicia Ault Barnett, Fixing Dysfunctional Family Leave, Bus. & Health, March 1997, at 22.

16. See id.

17. See id. at 23-24 (quoting 29 U. S. C º 2611( 11)).

18. 29 U. S. C. º 2611( 11)( A)-( B) (1994).

19. Id. º 2612( e)( 1).

20. Barnett, supra note 15, at 24.

21. As part of the FMLA, Congress established the Commission on Leave. The Commis-sion, under the Chairmanship of Senator Christopher Dodd (D-CT), was a diverse, bipartisan group representing business, labor, women's organizations, and both political parties. They were charged with researching a number of issues, including the benefits and costs of the FMLA for both employers and employees. They held three public hearings on the FMLA at different loca-tions around the U. S. Additionally, they commissioned two large research studies on the FMLA. One was a survey of employers by Westat, Inc., see CANTOR SURVEY, supra note 10, and the other was a survey of employees by the Survey Research Center at the University of Michi-gan. KATHERINE A. MCGONAGLE ET AL., INSTITUTE FOR SOCIAL RESEARCH, SURVEY RE-SEARCH CENTER, THE UNIV. OF MICH., COMMISSION ON LEAVE SURVEY OF EMPLOYEES ON THE IMPACT OF THE FAMILY AND MEDICAL LEAVE ACT, Submitted to the Commission on Family and Medical Leave (Oct. 13, 1995) [hereinafter MCGONAGLE SURVEY]. The results of both research studies are emphasized heavily in the Final Report released April 30, 1996. See COM-MISSION ON LEAVE, supra note 6.

22. See Jennifer Click, Take a Cooperative Approach to Intermittent Leave, HRMAGAZINE, Dec. 1996, at 48, 49.

23. See COMMISSION ON LEAVE, supra note 6.

24. See generally Jerry Goodstein, Institutional Pressures and Strategic Responsiveness: In-volvement in Work-Family Issues, 37 ACAD. MGMT. J. 350, 372-75 (1994) [hereinafter Goodstein, Institutional Pressures]; Ellen Ernst Kossek et al., The Dominant Logic of Employer-Sponsored Work and Family Initiatives: Human Resource Managers' Institutional Role, 47 HUM. REL. 1121, 1124-25 (1994).

25. See Kossek et al., supra note 24; See generally Paul J. DiMaggio & Walter W. Powell, The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields, 48 AM. SOC. REV. 147 (1983).

26. Kossek et al., supra note 24, at 1127

27. See generally Kossek et al., supra note 24; Hal Morgan & Frances J. Milliken, Keys to Action: Understanding Differences in Organizations' Responsiveness to Work-and-Family Issues, 31 HUM. RES. MGMT. 227 (1993).

28. Frances J. Milliken et al., Understanding Organizational Adaptation to Change: The Case of Work-Family Issues, 13 HUM. RES. PLANNING 91, 93 (1990).

29. Morgan & Milliken, supra note 27, at 241-42.

30. Kossek et al., supra note 24, at 1141-43.

31. See generally Morgan & Milliken, supra note 27.

32. See Andrew E. Scharlach et al., The Family and Medical Leave Act of 1993: How Fully is Business Complying?, 37 CAL. MGMT. REV. 66, 73 (1995).

33. Id. at 71 tbl. 3.

34. See Billings, supra note 12, at 1263.

35. DEPARTMENT OF LABOR, BUREAU OF LABOR STATISTICS, THE FAMILY AND MEDICAL LEAVE ACT: 47 MONTHS OF ENFORCEMENT AND OUTREACH ACTIVITY (1997).

36. Supra note 34, at 1263.

37. COMMISSION ON LEAVE, supra note 6, at 15-32.

38. Id. at 21-26.

39. See MCGONAGLE SURVEY, supra note 21.

40. COMMISSION ON LEAVE, supra note 6, at 17.

41. Id. at 17-19.

42. MCGONAGLE SURVEY, supra note 21, at 17.

43. Id. The telephone surveys were relatively brief, ranging from 3 minutes for respondents who reported being employed and needing no leave, to 5 minutes for those who were employed and reported needing leave but not taking it, to 10 minutes for those who reported taking leave in the covered time period. Altogether, approximately 10% of the respondents reported taking leave and working for an FMLA-covered organization. However, this figure probably overesti-mates FMLA usage, as not all of this leave was likely to have been taken under the FMLA. The McGonagle Survey recommended that administrative data be collected to determine actual FMLA usage rates. See MCGONAGLE SURVEY, supra note 21, at 52. This was partially ad-dressed by the Cantor Survey discussed below, and is specifically addressed by our survey of HR professionals.

44. See CANTOR SURVEY, supra note 10.

45. Id. at 3-6 tbl. 3-3.

46. Id. at 4-9. The Cantor Surveys went to those "most knowledgeable about the establish-ment's policies," CANTOR SURVEY, supra note 10, at 2-2, which could be either an owner, man-ager, or human resource professional. Their survey sought information on existing family and medical leave policies, the costs, benefits and impact on productivity of FMLA regulations, the impact of the FMLA on the provision of employee benefits provided by employers, and the recovery by employers of health benefits under section 2614( c)( 2). Id. While our survey was narrower in scope, it extends their research by: a) covering a longer time period (i. e., the first 24 versus 18 months the act was in effect), b) looking at both public and private sector usage of the FMLA, c) including more quantitative and interval-level data than their survey did, d) emphasiz-ing HR professionals exclusively as the target of our surveys, and e) including a number of variables which may be helpful to organizations in predicting usage of the FMLA. No previous studies have sought to predict FMLA usage by employees.

47. See Scharlach et al., supra note 32, at 66.

48. See id.

49. See id. at 66-67.

50. See STEFFEN T. KRAEHMER, QUANTITY TIME (1990).

51. See Scharlach et al., supra note 32, at 67.

52. See Eileen Trzcinski & William T. Alpert, Handling Work During Leave: Strategies and Costs, 3 J. MANAGERIAL ISSUES 403 (1991).

53. Scharlach et al., supra note 32, at 68.

54. See AMA Survey Reveals Perceptions of FMLA, HR FOCUS, Oct. 1993, at 11 [hereinaf-ter AMA Survey].

55. See Michelle Neely Martinez, FMLA: Headache or Opportunity?, HRMAGAZINE, Feb. 1994.

56. See CANTOR SURVEY, supra note 10, at 1-1.

57. See E. Patrick McDermott & Marilyn B. Katz, The Family and Medical Leave Act of 1993, 1993 LABOR L. J. 673.

58. See Morgan & Milliken, supra note 27, at 236-241.

59. See CANTOR SURVEY, supra note 10, at x.

60. See generally JANICE M. BEYER & HARRISON M. TRICE, IMPLEMENTING CHANGE (1978); Kossek et al., supra note 24; Stuart H. Milne et al., Factors Influencing Employees' Pro-pensity to Use an Employee Assistance Program, 47 PERSONNEL PSYCHOL. 123 (1991).

61. See Michael M. Harris & Mary L. Fennell, Perceptions of an Employee Assistance Pro-gram and Employee's Willingness to Participate, 24 J. APPLIED BHVRL. SCI. 423 (1988).

62. See Milne et al., supra note 60, at 140-43.

63. See VIOLA M. LECHNER & MICHAEL A. CREEDON, MANAGING WORK AND FAMILY LIFE (1994).

64. MCGONAGLE SURVEY, supra note 21, at 17.

65. See Goodstein, Institutional Pressures, supra note 24, at 376; Jerry Goodstein, Employer Involvement in Eldercare: An Organizational Adaptation Perspective, 38 ACAD. MGMT. J. 1657 (1995) [hereinafter Goodstein, Employer Involvement]; Morgan & Milliken, supra note 27.

66. See Robert L. Rose, After Turning Around Giddings & Lewis, Fife is Turned Out Him-self, WALL ST. J., June 22, 1993, at A1.

67. See Milne et al., supra note 60, at 140-43.

68. See Kossek et al., supra note 24.

69. We used a stratified random sample for our study. Within each state, equal percentages of mailing labels were selected from each of the following categories: 50-125, 126-250, 251-500, 501-750, and 750 employees and above. Organizations employing up to 2,500 individuals were included in the sample. Mailing lists were then generated randomly within each division for each state. This division was used primarily to allow us to compare our results with those reported in the Cantor Survey. See CANTOR SURVEY, supra note 10.

70. 2,780 surveys were sent out, as labels were discarded if individuals could be identified as holding a position other than Human Resource professional (e. g., lawyer or professor). Addi-tionally, when multiple members of the same organization were identified, the survey was sent to the individual with the highest ranking title, and other labels from that organization were dis-carded. 30 surveys were returned to us with no forwarding address. Approximately 100 organi-zations were excluded from our analyses, as they reported fewer than 50 employees, and are not subject to the FMLA.

71. PAMELA L. ALRECK & ROBERT B. SETTLE, THE SURVEY RESEARCH HANDBOOK 45 (1985). In Morgan and Milliken's survey of HR executives, their response rate was 18%. See Morgan & Milliken, supra note 27.

72. See generally CANTOR SURVEY, supra note 10; MCGONAGLE SURVEY, supra note 21; Scharlach et al., supra note 32.

73. To control for the impact of organizational size on leave usage, usage was calculated as the ratio of employees who had used the FMLA in an organization divided by that company's total number of employees. This variable was then used as the dependant variable for the statis-tical analyses reported below.

74. Specifically, two questions asked respondents to rank order various options concerning specific benefits and costs associated with the Act. The options for each question were drawn from previous literature. See AMA Survey, supra note 54; Michelle Neely Martinez, supra note 55; McDermott & Katz, supra note 57. Based on this literature, we compiled a list of potential benefits and costs associated with the FMLA. Respondents first ranked the following benefits pertaining to the FMLA (1= most significant benefit, 2= second most significant, etc.): 1) to use as a recruiting tool (e. g., vis-a-vis firms not covered by the Act), 2) to increase employee satisfac-tion, 3) to provide a more positive work atmosphere, 4) to retain employees who otherwise might quit, 5) to increase loyalty and commitment, 6) to decrease stress and increase the focus of employees, and 7) to increase employee reciprocity (i. e., "If we're good to them, they'll be good to us"). Respondents then ranked the following costs associated with the FMLA using the same scoring system: 1) lost productivity, 2) operational instability, 3) difficulty getting temporary help, 4) administrative concerns, and 5) training managers to understand and implement the FMLA.

75. Using a 5-point Likert scale with anchors from strongly agree to strongly disagree, they responded to statements concerning interpretation, costs of compliance, employee abuse, top management support of the Act as well as more general family-related policies and programs, and if employees believed that they would be punished for utilizing provisions of the FMLA.

76. The mean expresses the center or the average of a set of scores and is a good represen-tation of the group's characteristics. The standard deviation is a measure of the dispersion of a set of scores and provides a description of the extent to which scores differ from each other. The amount of correlation between two variables is denoted by the letter r. If two variables are positively correlated, this means that high values of one variable correspond to high values of the other, and low values with low values. Conversely, if low values of one variable are associated with high values of the other, the variables are described as negatively correlated. The range of r is from -1 (i. e., a perfect negative correlation) to +1 (a perfect positive correlation).

77. A t-test addresses the question of whether two means differ significantly from each other, i. e., whether they differ beyond what would be expected by chance alone. If the differ-ence is not statistically significant, the t-value is followed by the abbreviation n. s.

78. See CANTOR SURVEY, supra note 10, at 3-6.

79. If all respondents viewed an option as most important, it would receive a mean ranking of 1.0.

80. Multiple regression analyzes the separate and common influence of two or more in-dependent variables on a single dependent variable. The results tell us how well the independent variables predict the dependent variable. In this case, we are determining the impact of five independent variables (top management support, existing paid leave policies, interpretation problems, gender, and employee penalty perceptions) on a single dependent variable (usage of the FMLA). In Table 3, the column marked "b" displays the regression weights, i. e., the strength of the relationship between each variable and FMLA usage. The second column, sr 2 , shows the amount of unique variance in FMLA usage that is accounted for by each variable.

81. See generally MCGONAGLE SURVEY, surpa note 21; Scharlach et al., supra note 32, at 71 tbl. 3.

82. P values denote whether something is statistically significant or not. If a p-value is less than .05, it is generally regarded as statistically significant, i. e., there is less than a one in twenty chance that this occurred randomly. In this case, the likelihood that this occurred randomly is less than 1 in 100.

83. The Model R 2 refers to the amount of variance in the dependent variable that is ex-plained by the independent variables. In this case, 24% of the usage rate of the FMLA can be accounted for by the 5 independent variables in this model. Adjusted R 2 is a more conservative estimate of the amount of variance explained by the independent variables, in that it is adjusted for the number of variables used compared to the sample size. In our case, the sample size is large relative to the number of independent variables, thus, the adjusted R 2 is almost identical to the Model R 2 . The F statistic is used to express the amount of explained variance to the amount of unexplained variance. The larger the F value, the more variance that has been explained by the independent variables.

84. See COMMISSION ON LEAVE, supra note 6. Most organizations reported "no noticeable effect" of the FMLA on turnover or on other variables. However, in companies where employ-ees had taken leave, positive effects on turnover were reported from 13.9% of respondents, compared to only 0.7% reporting positive effects in organizations where no one had taken leave. Id. at 295. Overall, 5.2% reported positive effects, 0.5% reported negative effects, and 94.4% reported no noticeable effect of the Act on employee turnover.

85. See Sue Shellenbarger, If You Want a Firm That's Family Friendly, the List is Very Short, WALL ST. J., Sept. 6, 1995, B1.

86. See Michael Davis, Final Regulations Kick in April 6; Recordkeeping Remains a Con-cern, 13 BNA'S EMPLOYEE RELATIONS WEEKLY 339 (1995).

87. See COMMISSION ON LEAVE, supra note 6.

88. See Barnett, supra note 15 (quoting 29 U. S. C. º 2612 (a)( 1)( D)).

89. See CANTOR SURVEY, supra note 10, at 1-1.

90. See JOHN E. BUTLER ET AL., STRATEGY AND HUMAN RESOURCES MANAGEMENT (1991).

91. See Milne et al., supra note 60, at 140-43.

92. See Kossek et al., supra note 24, at 141-42.

93. See Morgan & Milliken, supra note 27.

94. See generally BURTON T. BEAM, JR. & JOHN J. MCFADDEN, EMPLOYEE BENEFITS (1996).

95. See COMMISSION ON LEAVE, supra note 6, at 94 fig. 5.2. In contrast, the percentages are much lower for other covered reasons for taking FMLA leave. For example, 13.3% reported taking leave to care for a newborn, adopted or foster child, 3.8% reported taking maternity-disability, 7.6% took leave to care for an ill child, 3.7% to care for an ill spouse, 8.6% to care for an ill parent, and 3.1% to care for an ill relative. Id.

96. In the McGonagle Survey data, for example, 22.1% of the employees they surveyed responded that they might lose their job if they took leave, and 21.9% felt that taking leave might hurt their job advancement. See COMMISSION ON LEAVE, supra note 6, at 300 tbl. 7. D.

97. See CANTOR SURVEY, supra note 10, at 3-6. These are not annual figures, as the Cantor Survey reported a percentage for the first 18 months the Act was in effect, and ours is for the first 24 months. Given this difference, our figure is actually slightly smaller than that reported by the Cantor Survey.

98. See id.

99. See Morgan & Milliken, supra note 27, at 236-42.

100. See generally JOHN A. FOSSUM, LABOR RELATIONS: DEVELOPMENT, STRUCTURE, PRO-CESS 48 (1995).

101. BARRY HIRSCH & DAVID MACPHERSON, UNION MEMBERSHIP AND EARNING DATA BOOK: COMPILATIONS FROM THE CURRENT POPULATION SURVEY 26 (1996).

102. Id. at 20.

103. See COMMISSION ON LEAVE, supra note 6. For example, 74% of the largest organiza-tions reported that it was either somewhat or very difficult to maintain additional records, com-pared to 43.7% for organizations with 251-500 employees, 33.8% for organizations with 100-250 employees, and 17.2% for organizations with 50-99 employees. Similarly, over 50% of the larg-est organizations reported difficulties in coordinating their leave policies with other federal laws. In contrast, difficulties were reported by 36.9% of firms with 251-500 employees, 32.8% of firms with 100-250 employees, and 12.9% of firms with 50-99 employees. Id. at 285 tbl. 6. A.

104. See CANTOR SURVEY, supra note 10.

105. See Gillian Flynn, The Final FMLA Regulations -What They Mean to HR, PERSONNEL J., July 1995, at 97.

106. See generally Linda Haas, Gender Equality and Social Policy, 11 J. FAM. ISSUES 401, 411-17 (1990); Allan Halcrow, Should Business Alone Pay for Social Process, PERSONNEL J., Sept. 1987, at 69-73.

107. 29 U. S. C. º 2612( a)( 1)( D)

108. 29 U. S. C. º 2612( a)( 1)( C)

109. See Joann S. Lublin, Family-Leave Law Can be Excuse for a Day Off, WALL ST. J., July 7, 1995, at B1.

110. See generally JOAN KOFODIMOS, BALANCING ACT (1993); Gregory K. Stephens & Steven M. Sommer, Linking Work-Based Social Support and Work Group Trust with Job In-volvement and Organizational Citizenship Behavior: A Test of a Path Analytic Model, Paper Presented Before the Academy of Management, Dallas (1994).

111. See generally Cheryl R. Saban & Dana Sacco, An FMLA Compliance Update: What Every Employer Should Know about the Final Rule, 21 EMPLOYEE RELATIONS L. J. 145 (1995).

112. DEPARTMENT OF LABOR, supra note 35. Total monetary damages received by employ-ees under the FMLA through June 30, 1997 totaled $7,947,977. Id.

113. See 42 U. S. C. º 2000e-2000e-17 (1997).

114. See Expansion of the Family and Medical Leave Act (FMLA) Under Debate, WORK-PLACE VISIONS (Society for Hum. Res. Mgmt., Alexandria, Va.) May/ June 1997, at 3-4. [hereinaf-ter WORKPLACE VISIONS]. For example, some of the bills proposed to expand the FMLA are: the Family Friendly Workplace Act of 1996, S. 4, 105th Cong. (1997); the Family and Medical Leave Improvements Act of 1997, H. R. 109, 105th Cong. (1997); the Parental and Community Involvement Leave Act, H. R. 191, 105th Cong. (1997); the Time for Schools Act of 1997, S. 280, 105th Cong. (1997); and the Battered Women's Employment Protection Act, S. 367, 105th Cong. (1997).

115. See WORKPLACE VISIONS, supra note 114.

116. Senator Larry Craig (R-ID), a member of the Commission on Family and Medical Leave, wrote in an additional view at the end of the Commission's Final Report that the "hall- mark finding was that the FMLA seems not to have had a major impact, for good or ill," COM-MISSION ON LEAVE, supra note 6, at 226.

117. Experiences with the FMLA appear to be analogous to what happened with the Work-ers' Adjustment Retraining and Notification or WARN Act of 1989. 29 U. S. C. º 2101-09 (1997). The WARN Act requires organizations with 100 or more employees to give 60 days notice before closing a facility. Prior to its passage, then-President Ronald Reagan called it a "ticking time bomb" for American business. See Matthew Cooper & Allan Holmes, The Disaster that Never Happened, U. S. NEWS & WORLD REP., Feb. 26, 1990, at 47. However, subsequent experi-ence has shown that the Act has had very minimal effects on business.

118. See WORKPLACE VISIONS, supra note 114.

119. See COMMISSION ON LEAVE, supra note 6, at 259 tbl. 4. f.

120. See generally COMMISSION ON LEAVE, supra note 6; RICHARD A. SWENSON, MARGIN: RESTORING EMOTIONAL, PHYSICAL, FINANCIAL, AND TIME RESERVES TO OVERLOADED LIVES (1992); Holly B. Tompson & M. Audrey Kosgaard, Understanding the Impact of Multiple Life Roles on Work Attitudes and Intentions, Paper Presented Before the Academy of Management, Vancouver (1995).