Gender effect on the perceived value of human capital in management.

Brahmasrene, Tantatape ; Chaiprasop, Ruangthong

ABSTRACT

This study examined earnings inequality in Thailand. There is a relatively higher ratio of men than women in high level management positions in manufacturing. Women are typically employed in lower level positions such as bookkeeping, human resources and secretarial staff. They are the main force in the service sector such as banking, retail trades, education, hospitality and tourism. The trend has shifted in recent years when more women than men enter into a managerial level position. However, earnings for women are relatively lower than that of their male counterparts.

A human capital model was employed to assess factors affecting earnings among men and women at managerial levels. These factors included education, work experience, marital status, job satisfaction, number of hours at work and at home, willingness to engage in frequent business trips and gender. The empirical analysis found that education, work experience, marital status and willingness to engage in frequent business trips had a significant effect on all samples in the study. Education and work experience profoundly affected earnings for both men and women. However, only women's earnings were affected by marital status and willingness to engage in frequent business trips.

INTRODUCTION

Thailand is currently undergoing reforms and adjustments in economic development planning aimed at bolstering market confidence and achieving economic recovery and stability. Amidst the past financial crisis, female employment has played an important role in the Thai economy, especially in the private sector. Table 1 provides selected statistics to compare males and females in the labor force. The Thai labor force is about 57 percent male and 43 percent female. More than twice the number of administrative, managerial and professional workers reside in an urban area. The percentage of women in administrative, managerial and professional positions increased from 49.7 in 1996 to 56.8 percent in 2000. There have been more job opportunities and better career advancement for women in the modern private sector. However, women with college degrees continue to earn less than their male counterparts. Wage and earnings inequality may arise due to heterogeneous jobs and workers. The initial motivation behind this research was to assess the impact of crucial factors affecting earnings at the entry to middle managerial levels in Thailand in order to control for some consequences of heterogeneities. Previous research has shown gender effect on employment and earnings inequality mainly among low educated and unskilled workers. Higher educated women at the managerial level have rarely been investigated.

REVIEW OF RELATED LITERATURE

The US historical trend in the female-male ratio indicated a rapid increase during the 1980s and stood at .72 by 1990 (O'Neill & Polachek, 1993). American women's economic status improved significantly in the 1980s. Understanding the gender gap in pay is important because even in the absence of any labor-market discrimination, it is unlikely that the wage rates of women and men would be equal (O'Neill, 2003). The relative improvement in women's wages can be attributed to an increase in labor market experience and work attachment (Blau & Kahn, 1997; O'Neill & Polachek, 1993). Not only can changes in the structure of the economy affect the wage distribution, but also human capital investments can create income inequality within a particular population.

Earnings dispersion generally occurs due to differences in job characteristics and worker skills. Conversely, earnings inequality may occur among equally skilled workers in the same job because of gender, race and other irrelevant characteristics. It is widely accepted that differences in human capital between men and women do matter (Borjas, 2000). Men tend to acquire more human capital than women. The human capital that women acquire depreciates somewhat during the child bearing years when they engage in household production. According to Polachek (1981), women who wish to maximize the present value of lifetime earnings will not enter occupations where their skills will depreciate rapidly during the child bearing years. The human capital explanation of gender wage differential states that because women have shorter payoff periods, they invest less in on-the-job training and other forms of human capital, and hence have lower wages. Becker (1985) pointed out that women remained largely responsible for child care and household production with high effort intensive. Thereby, married women seek occupations that require less effort. They segregate themselves into less demanding occupations. In addition, wage depends not only on hours worked but also work effort. Hence, women's earnings are lower. Lower wages lead to less incentive to work. Gronau (1988) and Neumark (1995) examined whether a woman weaker work's attachment resulted in a lower wage or whether the lower wage led to less work attachment. Some studies (England, Farkas, Kilbourne and Dou, 1998; Macpherson & Hirsch, 1995) investigated the relationship between women's employment and earnings in a particular occupation. After holding constant the human capital and other socioeconomic characteristics, these studies found that female jobs pay lower wages.

Michael (1973) further investigated how human capital influenced various aspect of a worker's behavior. Macpherson and Hirsch (1995) found that women in a profession where at least seventy five percent of the coworkers were women earned fourteen percent less than comparable women in a profession where less than twenty five percent of the coworkers were women. In contrast, a man in a predominantly female profession earned fourteen percent less than a man in a predominantly male profession.

HYPOTHESIS

The above literature review leads to the hypothesis that earnings are affected by three categories of independent variables: human capital variables, worker characteristics and working conditions.

Human capital variables include educational attainment (ED) and years of work experience (WORK). Mincer (1974), employing earnings power function, found the direct effect of years of education and work experience on earnings. Educational attainment augments a worker's productivity via enhancement of his/her abilities or skills, especially cognitive skills. A more educated and better trained worker is able to offer a greater productive potential than one with less education and training. Additional years of education will increase an individual's productivity. Johnson, Palermo, and Asgary (2002) found positive effects of increased literacy on wages in manufacturing in their international study. Mincer (1974) indicated that once the postschooling on-the-job training investment was included in the human capital definition, one-half to two-thirds of the personal earning variations were explained. Over time, workers may acquire human capital through work experience and thereby increase earnings (Mincer, 1974; Naderi & Mace, 2003). Acquired additional job skill increases productivity. Increase in marginal productivity results in higher earning. Thus, education and work experience positively affect earnings (Y).

Worker characteristics may include marital status (MSTATUS), weekly hours worked at the office (HRWORK) and at home (HRHOME) and gender (GENDER). Schumann, Ahlburg and Mahoney (1994) examined the influence of both the worker characteristics and job characteristics on earnings. Data were collected specifically for this purpose. They found that both worker characteristics and job characteristics were important determinants of pay. In addition, women held jobs that were of lower value to the firm. This explains part, but not all, of the gross male-female pay differential.

Marital status is expected to have either positive or negative influence on earnings. Married persons may have a greater incentive to increase earnings to support their families. On the other hand, married women may engage in occupations that require less effort in order to spend more time providing for child care and household production. This leads to lower earnings. Time spent in the office increases productivity if it is time well spent, but the fact that most women continue to take the primary responsibility for child care is a cause of distraction, diversion, anxiety and absenteeism (Schwartz, 1989). Hill (1983) investigated allocation of weekly hours to various activities by gender and marital status. Women allocated more hours to the non-market sector than men. Married women worked about 35 hours per week as compared to 14 hours for married men. However, the total hours at work and at home were above 50 hours per week for both genders. While weekly hours worked at the office is positively related to earnings, weekly hours worked at home is expected to have a negative impact on earnings.

Working conditions are reflected in job satisfaction (JOBSAT) and willingness to engage in frequent business trips (FREQUENT). Job satisfaction is conceived in terms of a worker's general reaction to the job without reference to any specific job facets (Hamermesh, 1977). Higher job satisfaction results in higher productivity and thereby increases earnings. Schumann, Ahlburg and Mahoney (1994) indicated that the characteristics of both workers and jobs determine pay. The introduction of job characteristics to the pay function reduces the male-female pay differential. Willingness to engage in business trips frequently (FREQUENT) is a newly introduced factor in this paper that represents job characteristics and preference or tastes. If women prefer jobs that require fewer business trips, then the exercise of this preference may result in lower earnings. This should directly relate to earnings.

For empirical analysis, the model has been constructed as shown below:

Y or LnY = CONSTANT + b1ED + b2WORK + b3MSTATUS + b4JOBSAT + b5HRWORK + b6HRHOME + b7FREQUENT + b8GENDER + ui

Description of variables is summarized in Table 2. ui is a stochastic error term or disturbance term.

First, two models are estimated by controlling the gender variable. In effect, the first model was estimated based upon men's earnings, the second model on women's earnings. The last model includes gender variables, along with other independent variables, to determine whether the gender variable is statistically significant.

DATA AND METHODOLOGY

The data source used in this study is an original survey conducted in 36 parts of the Bangkok Metropolitan Area in 2002. A stratified simple random sampling was employed to obtain sample managers from two groups of establishments, which included (1) manufacturing, handicraft, construction, transportation, and (2) retail and service industries. The population was restricted to entry to middle level men and women managers in establishments of more than twenty employees. The number of returned surveys totaled 203. Of those surveys completed and returned, some had one or more items on the survey left blank. Therefore, 198 observations were analyzed. Members of the sample surveyed ranged in age from 21-58 years, and 63 percent were married. Education ranged from high school diplomas to Master's degrees.

Further data exploration is shown in Table 3, Subpopulation Differences. By and large, women managers' monthly earnings were about 8 percent less than men's earnings. Monthly earnings are shown in Thai baht (approximately 40 Baht = 1 US dollar). Married persons earned 54 percent more than singles. Over all, when compared with women, men managers showed slightly higher education (16.30 VS 15.83 years), years of work experiences (6.82 VS 6.32), job satisfaction (.81 VS .69) and willingness to engage in frequent business trips (.65 VS .55). They also reported fewer numbers of hours worked in the office (44.76 VS 46.04) and at home (6.43 VS 9.83).

The ordinary least square (OLS) method was employed to test the above hypotheses. One of the tasks in performing regression analysis with several independent variables was to calculate a correlation matrix for all variables. Table 3 reports the Pearson Correlations for all variables. Monthly managers' earnings were highly correlated with education attainment, years of work experience and marital status (< 0.01). There were no particularly large intercorrelations among independent variables. However, a measure of multicollinearity among independent variables would be performed.

EMPIRICAL RESULTS

The assumption of linear multiple regression and the fitness of the model was tested with no apparent violation. According to the computed values of a multiple regression model, the null hypothesis was rejected at a significant level of less than 0.01 (F test) in all three earnings models shown in Table 5. This means that among these estimated equations, there existed a relationship between earnings and the explanatory variables; education, years of work experience, marital status, job satisfaction, weekly hours worked at the office and home, willingness to engage in frequent business trips and gender (for third model only). The variance inflation factor (VIF) is presented in Table 5 to detect multicollinearity among independent variables. A value of VIF less than 10 generally indicates no presence of multicollinearity. It appears that the observed dependencies did not affect their coefficients.

The coefficient of multiple determination (R Square) of earnings model for men, women and both were 0.596, 0.515 and 0.527, respectively (see Table 5). These indicated that 51.5 to 59.6 percent of variation in the earnings could be explained by the variations of variables included in each model. Furthermore, significant test (t-test) for men's earnings indicated that coefficients of education and years of work experience had a highly significant t-value (< 0.01). Therefore, the null hypothesis of education and years of work experience was rejected. Two additional independent variables, marital status (< 0.1) and willingness to engage in frequent business trips (< 0.01) were significant in women's earnings. When combining both men and women's earnings, coefficients of education and years of work experience had a highly significant t-value (< 0.01), while marital status and willingness to engage in frequent business trip were significant at < 0.05, all with expected signs. Job satisfaction, weekly hours spent at work at the office and at home, and gender itself had no significant influence on earnings.

DISCUSSION

This study found that years of education and work experience, marital status and willingness to engage in frequent business trips were crucial factors in determining the managerial earnings of both men and women. Years of education and work experience positively affected earnings. Marital status and willingness to engage in frequent business trips seemed to have more effect on women's earnings, as these variables were insignificant under men's earnings. When marital status was coded as 0 for married and 1 for single persons, the negative coefficient indicated that single women earned less than married women. This might be because married persons strive to earn more to support their families. The result contradicts what generally is found in the USA, i.e., that married women earned less than singles due to high effort for childcare and household production (Becker, 1985).

The willingness to engage in frequent business trips directly affected women's earnings. Willingness to accept frequent business trips was more likely to enhance women manager's earnings. This frequent business trip factor may reflect that workers have different tastes or job preferences, because it was not a significant factor in estimating men's earnings. If frequent business trips become an issue for women, then the exercise of their preference may result in lower earnings than men.

Table 6 further displays the independent sample test results about differences in group means of earnings between men and women. The equality hypothesis was accepted. It appeared that gender had no significant effect on earnings. Gender was not a significant factor in determining manager's earnings in this study. This may be due to job characteristics and not the individual characteristics that determine earnings. The introduction of job characteristics such as frequent business trip to the earning function reduces the male-female earning inequality. Moreover, responsibilities and effort required in managerial positions are the same regardless of gender. A human resource management approach, therefore bases pay on the job requirements, rather than individual characteristics. Furthermore, there was no statistical difference in means between men and women managers in terms of education, years of work experience, number of hours at work and willingness to engage in frequent business trips.

CONTRIBUTIONS

This paper makes three important contributions in the literature on gender effect on earnings inequality. First, no existing research has studied earnings inequality at management level in Thailand. By controlling for similar observable dimensions, such as management level, educational attainment, work experience and race, it demonstrates that gender is not a significant factor in earnings inequality. Second, a newly created job preference factor, such as frequent business trips, was a significant factor in the women's earnings model. The introduction of job characteristics such as frequent business trips to the earning function reduces the male-female earning inequality. On the other hand, the exercise of this preference may contribute to lower earnings for women. Third, the results of testing for equality of means between men and women managers showed that the two populations were indifferent in terms of earnings, educational attainment, years of work experience, work hours at office and willingness to engage in frequent business trips. However, men and women managers were significantly different in terms of marital status, job satisfaction and hours work at home.

PRACTICAL IMPLICATIONS

For practical implication, a semi-logarithm of monthly manager's earnings (LnY) was estimated. Table 7 relates the predicted log on earnings to the independent variables, so that their coefficients can be expressed in percentage terms. Several studies (Becker & Chiswick, 1966; Mincer, 1974; Hanoch, 1967) estimated the rate of return to education in this fashion. For example, the education coefficient of 0.185 men's earnings (Model 4) indicated an 18.5 percent internal rate of return. If men increase education by 1 year, their earnings will increase by 18.5 percent. The education coefficient of women's earnings (Model 5) was 0.178 reflecting a 0.7 percent return below their male counterparts. In Beker's work (1975), the estimated rate of return in 1939, 1949 and 1958 was 14.5, 13.0 and 14.8 percent, respectively. Angrist and Krueger (1991) found a rate of return of 8 percent in 1980. A more recent study by Kane and Rouse (1995) reported a 9 percent rate of return for higher education in 1986.

However, work experience coefficients of men and women's earnings were 0.046 and 0.048, respectively. An increase in one year work experience among men will result in 4.6 percent higher in earnings. Women had slight advantage of 0.2 percent over men at 4.8 percent. In the USA, Duncan and Hoffman (1979) reported that the rate of return to a year of experience was between 5.4-8.5percent. Rosen (1982), using the same data set, found that after correcting for inflation, the real return on investment in on-the-job training (OJT) as measured by experience for college graduates was 12.6 percent and for non-college graduates was 19 percent. Groot and Mekkelholt (1995) employed a 1985 labor market survey of the Organization of Strategic Labor Market Research in the Netherlands with subset of male wage earners (1,339 observations). They discovered that the rate of return to OJT increased with the level of education. The rate of return to OJT for the least educated workers was 13 percent. At the intermediate level the rate of return to OJT was 36 percent, while at the highest education levels the rate of return to OJT was over 80 percent.

As found in Table 3, women managers overall spent more time than men taking care of domestic chores (9.83 VS 6.43 hours) and have less job satisfaction (0.69 VS 0.81). In particular, married women spent 17.94 hours as compared to 9.99 hours for married men on average to take care of work at home. This finding suggests more support for women as described below.

POLICY IMPLICATIONS

Improvements in women's labor work status coexists with deterioration in their relative economic well-being such as inhibiting fertility and dissolving marriage (Smith and Ward, 1989). Balancing efforts in the labor market and the family should be taken into account in formulating public policy on women's issues. In particular, policies should address how well they promote gender equity in the work force and what impact they will have on family. Nevertheless, it is imperative to address important policy implications of this result for business in the developing countries. In general, there appears to be a need to combine two types of policy action for women that promote progressive social change (Beneria, 2003). The first is a focus on the transformation of gender relations with policies that will enhance gender equality. These policies and actions can rely on a wide range of possibilities, from educational policies to the many dimensions of cultural transformation. Secondly, more structural policies aimed at socioeconomic change and the promotion of development models capable of incorporating progressive social change should be an integral part of this transformation. Both types of policy can complement each other and are interconnected in the organizational culture.

Our findings in this study provide a rationale for policy makers to encourage business organizations to support women managers, especially the married. O'Neill (2003) examined the trends and reasons of the gender-gap in wages in the United States. Gender gap was largely due to non-discriminatory factors that were unlikely to change unless the roles of men and women in the home become more nearly identical. The dual role of women workers in Thai society should be recognized and every effort should be taken to facilitate women's contributions. Flexible working schedules and management practices, in-plant childcare nurseries, daycare operations and other nontraditional benefits are some examples found in a conventional work setting. A study by a multinational corporation showed that the rate of turnover in management position was two and a half times higher among top performing women than it was among men. A large producer of consumer goods reported that one half of the women who took maternity leave return to their jobs late or not at all (Schwartz, 1989). Thus, it is important to provide support for women who want to both raise children and pursue a career, and to promote women's freedom to pursue a career and a family.

Also noteworthy are the advances in technology that continue to increase competition in the job market. Automated technology enhances a firm's ability to accommodate employees in a non-conventional work setting, such as working from home. Thai business organizations should embrace the idea of flexible work schedules, including allowing employees to do some work from home. Women should equip themselves with computer skills and communication technologies to fully utilize home-based work environments. With these non-conventional work opportunities, government and business organizations should emphasize the importance of education and training in dealing with the highly competitive, global business, and provide financial and technical support to women. Equality in education is a necessary, but insufficient, condition to couteract gender-based and other forms of inequality due to discriminatory practices. Anti-discriminatory policies must also be implemented.

CONCLUSIONS

Much of the earnings inequality literature has attempted to evaluate the impact of gender on earnings. This paper investigated the gender effect on earnings inequality at the administrative and managerial level in the private sector of Thailand. Factors that are important in explaining earnings inequality were explored. Contradictory to the popular findings, this study found no gender influence on earnings and no earnings inequality between men and women managers in Thailand. An avenue for further research might be to investigate other dimensions of gender inequality that likely exist in hiring, promotion and training. In addition, there is room for future research to include studies in other countries for generalization of these results. Such research will contribute significantly toward our understanding of the perceived value of human capital in different organizational cultures.

Note: The authors acknowledge financial support from National Research Council of Thailand. The authors wish to thank Journal Editor, referees and their colleagues for comments on the earlier draft.

REFERENCES

Angrist, J.D. & Alan B. Krueger (1991). Does compulsory school attendance affect schooling and earnings? Quarterly Journal of Economics, November, 979-1014.

Becker, Gary S. (1985). Human capital, effort, and the sexual division of labor. Journal of Labor Economics, January, 33-58.

Becker, Gary S. (1975). Human capital, New York: National Bureau of Economic Research.

Becker, Gary S. & B.R. Chiswick (1966). Education and the distribution of earnings. American Economic Review, 56, 358-369.

Beneria, Lourdes (2003). Gender, development and globalization: economics as if people mattered, New York, NY: Routledge.

Blau, F. and L.F. Kahn (1997). Swimming upstream: trends in gender wage differential in the 1980s. Journal of Labor Economics, 15, 1-42.

Borjas (2000). Labor Economics, Crawfordsville: Irwin-McGraw-Hill

Duncan, G. & Hoffman, J. (1979). On-the-job training and earning differences by race and sex, The Review of Economics and Statistics, 61, 594-603.

England, P., G. Farkas, B.S. Kilbourne & T. Dou (1998). Explaining occupational sex segregation and wages: findings from a model with fixed effects. American Sociological Review, 53, 544-558.

Gronau, Reuben (1988). Sex-related wage differentials and women's interrupted labor careers-the chicken or the egg. Journal of Labor Economics, 6, 277-301.

Groot, Wim & Eddie Mekkelholt (1995). The rate of return to investments in on-the-job training, Applied Economics, 27 (2), 173-181.

Hamermesh, D. (1977). Economics aspects of job satisfaction. In O.Ashenfelter and W.Oates (Eds), Essays in Labor market analysis. New York: John Wiley & Sons.

Hanoch, Giora (1967). An economic analysis of earning and schooling, Journal of Human Resources, 2, 310-329.

Hill, Martha S. (1983). Pattern of time use. In F. Thomas Juster & Frank P. Stafford (Eds.), Time, Good, and Well Being. Ann Arbor: University of Michigan Survey Research Center.

Johnson, Peter, Tia Palermo & Nader Asgary (2002). Effects of increased literacy on wages in manufacturing: an international study, Proceedings of the Academy for Economics and Economic Education.

Kane, T. J. & Rouse, C. (1995). Labor market returns to two-and four-year colleges. American Economic Review, June, 600-613.

Macpherson, D. & B.T. Hirsch (1995). Wages and gender composition: why do women's job pay less? Journal of Labor Economics, 13, 426-471.

Michael, Robert T. (1973). Education in non market Production. The Journal of Political Economy, 81, 306-327.

Mincer, J. (1974). Schooling, Experience, and Earnings, New York: Columbia University.

Naderi, A. & J. Mace (2003). Education and earnings: a multilevel analysis (a case study of the manufacturing sector in Iran), Economics of Education Review, 22, 143-156.

Neumark, David (1995). Sex discrimination and women's labor market outcomes. Journal of Human Resources, 30, 713-740.

O'Neill, June (2003). The gender gap in wages, circa 2000. American Economic Review, 93 (2), 309-314.

O'Neill, J. & S. Polachek (1993). Why the gender gap in wages narrowed in the 1980s. Journal of Labor Economics, 11, 205-228.

Polachek, S.W. (1981). Occupational self selection: a human capital approach to sex differences in occupational structure. Review of Economics and Statistics, 63, 60-69.

Rosen, H. S. (1982). Taxation and on-the-job training decisions, The Review of Economics and Statistics, 64, 442-449.

Schumann, Paul L.; Dennis A. Ahlburg & Christine Brown Mahoney (1994). The effects of human capital and job characteristics on pay. Journal of Human Resources, Spring, 29 (2), 481-503.

Schwartz, Felice N. (1989). Management women and the new facts of life. Harvard Business Review, January-February, 65-79.

Smith, James P., & Michael Ward (1989). Women in the labor market and in the family. The Journal of Economic Perspectives, 3 (1), 9-23.

Tantatape Brahmasrene, Purdue University North Central Ruangthong Chaiprasop, Ramkhamhaeng University, Thailand

Table 1: Thailand Labor Force Survey
Selected Categories by Gender

Year 1996

Gender Male Female

Labor force in millions (percent) 17.2 (57.34) 12.8 (42.66)

Administrative, Managerial & Professional Workers

Urban Area (percent) 170,850 (50.3) 168,700 (49.7)
Rural Area (percent) 82,400 (44.7) 101,800 (55.3)

Year 2000

Gender Male Female
Labor force in millions (percent) 17.4 (57.24) 13.0 (42.76)

Administrative, Managerial & Professional Workers

Urban Area (percent) 248,600 (43.2) 326,200 (56.8)
Rural Area (percent) 96,100 (40.7) 139,800 (59.3)

Year 1994 N/A

Avg. Monthly Earnings (Baht) of College Graduates

Urban Area 10,192 9,662
Rural Area 9,618 8,390

Avg. Weekly Hour Work of College Graduates

Urban Area 38.6 37.9
Rural Area 38.6 37.9

Source: Thai National Statistical Office

Table 2: Description of Variables

 Dependent variables

Y Monthly earnings
LnY Natural logarithm of monthly earnings
 Human capital variables
ED Years of education
WORK Years of work experience
 Worker characteristics
MSTATUS Marital status: 0 = married, 1 = otherwise
HRWORK Weekly hours work at the office
HRHOME Weekly hours work at home
GENDER 0 = Men, 1 = Women
 Working conditions
JOBSAT Job satisfaction: 0 = unsatisfied to unsure, 1 = satisfied
FREQUENT Willingness to engage in business trip frequently: 0 = No,
 1 = Yes

Table 3: Subpopulations Differences

Gender Mstatus Y Ed

Men .00 24,747.7273 16.4667
 1.00 16,470.0000 16.1667
 Total 20,372.3571 16.3030

Women .00 24,849.0789 16.0000
 1.00 15,978.9024 15.7595
 Total 18,787.7917 15.8333

Total .00 24,801.9718 16.2154
 1.00 16,131.5966 15.8870
 Total 19,371.5789 16.0056

Mstatus Work Jobsat Hrwork

 .00 9.2059 .9062 45.7000
 1.00 4.7462 .7250 43.9630
 Total 6.8233 .8056 44.7620

 .00 9.3925 .8250 46.4406
 1.00 4.8027 .6235 45.8306
 Total 6.3200 .6880 46.0383

 .00 9.3068 .8611 46.1309
 1.00 4.7843 .6560 45.2640
 Total 6.5094 .7310 45.5951

Mstatus Hrhome Frequent

 .00 9.9853 .7187
 1.00 3.4075 .6000
 Total 6.4297 .6528

 .00 17.9359 .3750
 1.00 6.1059 .6353
 Total 9.8266 .5520

 .00 14.2329 .5278
 1.00 5.2424 .6240
 Total 8.5571 .5888

Note Marital status: 0 = married, 1 = otherwise

Table 4: Correlations

 Y Edu Work

Y 1 .285(**) .371(**)
EDU .285(**) 1 -.042
WORK .371(**) -.042 1
MSTATUS -.305(**) -.173(*) -.389(**)
JOBSAT .122 -.051 .171(*)
HRWORK .028 -.004 .023
HRHOME .157 -.038 .187(*)
FREQUENT .052 .079 -.096
GENDER -.056 -.249(**) -.044

 Mstatus Jobsat Hrwork

Y -.305(**) .122 .028
EDU -.173(*) -.051 -.004
WORK -.389(**) .171(*) .023
MSTATUS 1 -.223(**) -.050
JOBSAT -.223(**) 1 -.119
HRWORK -.05 -.119 1
HRHOME -.474(**) -.002 .036
FREQUENT .094 -.083 -.008
GENDER .142(*) -.125 .066

 Hrhome Frequent Gender

Y .157 .052 -.056
EDU -.038 .079 -.249(**)
WORK .187(*) -.096 -.044
MSTATUS -.474(**) .094 .142(*)
JOBSAT -.002 -.083 -.125
HRWORK .036 -.008 .066
HRHOME 1 -.256(**) .190(*)
FREQUENT -.256(**) 1 -.095
GENDER .190(*) -.095 1

NOTE ** Correlation is significant at the 0.01 level (2-tailed).

* Correlation is significant at the 0.05 level (2-tailed).

Table 5: Earnings Model Coefficients

Coefficients Men

 (1) Y VIF

CONSTANT -33856.264 * (-1.743)
ED 2764.699 *** (2.800) 1.376
WORK 900.398 *** (3.288) 1.309
MSTATUS -5747.549 (-1.555) 1.800
JOBSAT 3108.047 (.776) 1.220
HRWORK 107.260 (.714) 1.038
HRHOME -79.442 (-.426) 1.326
FREQUENT 134.501 (.043) 1.095
GENDER
R Square 0.596
F Statistics 6.736 ***

 Women

 (2) Y VIF

CONSTANT -33638.461 ** (-2.250)
ED 2808.614 *** (3.697) 1.084
WORK 1054.021 *** (4.989) 1.275
MSTATUS -5118.356 * (-1.712) 1.745
JOBSAT 3724.927 (1.332) 1.327
HRWORK -33.708 (-.280) 1.205
HRHOME 55.742 (.584) 1.590
FREQUENT 7099.004 *** (2.796) 1.354
GENDER
R Square 0.515
F Statistics 9.716 ***

 Both

 (3) Y VIF

CONSTANT -34758.282 *** (-3.018)
ED 2812.594 *** (4.819) 1.162
WORK 952.408 *** (5.943) 1.222
MSTATUS -4820.381 ** (-2.160) 1.700
JOBSAT 3305.612 (1.543) 1.184
HRWORK -3.100 (-.035) 1.082
HRHOME 25.588 (.318) 1.519
FREQUENT 4532.556 ** (2.437) 1.173
GENDER 2244.466 (1.182) 1.168
R Square 0.527
F Statistics 14.344 ***

Notes t statistics are in parentheses. Significant
level : * 0.10, ** 0.05, *** 0.01

VIF = Variance inflation factor, a measure of collinearity

Table 6: Independent Samples Test

Variables Variances Assumed Levene's Test for Equality
 of Variances
 F Sig.

Y Equal .260 .611
 Not Equal

ED Equal .169 .681
 Not Equal

WORK Equal 1.568 .212
 Not Equal

MSTATUS Equal 9.86 .002
 Not Equal

JOBSAT Equal 14.415 .000
 Not Equal

HRWORK Equal .007 .932
 Not Equal

HRHOME Equal 4.989 .027
 Not Equal

FREQUENT Equal 7.725 .006
 Not Equal

Variances t-test for Equality of Means
 Assumed t Sig. (2-tailed)

 Equal .764 .446
Not Equal .793 .429

 Equal 1.287 .200
Not Equal 1.249 .214

 Equal .617 .538
Not Equal .610 .543

 Equal -2.019 .045
Not Equal -1.981 .049

 Equal 1.764 .079
Not Equal 1.839 .068

 Equal -1.035 .302
Not Equal -1.070 .287

 Equal -2.201 .029
Not Equal -2.424 .017

 Equal 1.338 .183
Not Equal 1.352 .178

Table 7: Model Coefficients for Natural Logarithm Of Monthly Earnings

Coefficients Men

 (4) LnY VIF

CONSTANT 6.087 *** (7.063)
ED .185 *** (4.229) 1.376
WORK .046 *** (3.762) 1.309
MSTATUS -.212 (-1.291) 1.800
JOBSAT .216 (1.213) 1.220
HRWORK .009 (1.307) 1.038
HRHOME -.003 (-.328) 1.326
FREQUENT -.084 (-.602) 1.095
GENDER
R Square 0.685
F Statistics 9.930 ***

Coefficients Women

 (5) LnY VIF

CONSTANT 6.536 *** (9.399)
ED .178 *** (5.032) 1.084
WORK .048 *** (4.884) 1.275
MSTATUS -.223 (-1.600) 1.745
JOBSAT .090 (.690) 1.327
HRWORK .000 (-.009) 1.205
HRHOME .004 (.794) 1.590
FREQUENT .190 (1.605) 1.354
GENDER
R Square 0.515
F Statistics 9.716 ***

Coefficients Both

 (6) LnY VIF

CONSTANT 6.346 *** (12.034)
ED .180 *** (6.750) 1.162
WORK .044 *** (6.046) 1.222
MSTATUS -.208 ** (-2.035) 1.700
JOBSAT .123 (1.250) 1.184
HRWORK .003 (.632) 1.082
HRHOME .002 (.573) 1.519
FREQUENT .106 (1.245) 1.173
GENDER .082 (.943) 1.168
R Square 0.590
F Statistics 18.520 ***

Notes t statistics are in parentheses.

Significant level : * 0.10, ** 0.05, *** 0.01

VIF = Variance inflation factor, a measure of collinearity