DO WOMEN STILL EARN LESS THAN MEN?
An empirical study, using open data, about the effect of education on the gender income gap in the Netherlands
Duco de Wit
Student number: 10092617 First Supervisor: Vladimer Kobayashi Second Supervisor: Dr. Gábor Kismihók
ABSTRACT
In spite of the fact that more women are graduating from colleges and universities than men, in the US women still receive 82 cents for every dollar men earn. Using earlier research, I investigated whether there is a gender income gap in the Netherlands, and if it is possible to overcome that gap with educational attainment. More specifically, I have conducted a search through the different open databases available on the Internet about Education and Income to find out if there is an income difference between male and female workers in the Netherlands, particularly if men earn more than women. Next, whether higher education leads to higher earnings and lastly if the difference of income between male and female workers decreases as the education level increases. These propositions will be tested using an existing database about education and labour with a sample of 666 workers from the Netherlands. The first three propositions are supported, meaning that there is a gender income gap in the Netherlands and higher education leads to a higher income. However, the last proposition was not supported, thus the gender income gap cannot be narrowed by getting a higher educational degree.
TABLE OF CONTENTS
1. INTRODUCTION 1
1.1. Background of the Study 1
1.2. Objectives 1
1.3. Scope and Limitations 2
1.4. Impact and Contribution 2
2. LITERATURE REVIEW 3
3. CONCEPTUAL FRAMEWORK 5
3.1. Model 5
3.2. Propositions 6
4. MATERIALS AND METHOD 7
4.1. Data source 7
4.2. Methodology 8
4.3. Data Descriptions 9
5. RESULTS AND DISCUSSION 11
5.1. Descriptives 11
5.2. Results of Analysis 13
6. CONCLUSION 18
6.1. General Conclusion 18
6.2. Implication and Future Research 18
7. REFERENCES 19
1. INTRODUCTION
1.1 Background of the Study
A trend that has been going on for years is that more women are graduating from colleges and universities than men, and the workforce consists for almost half of women. In 2013, the ratio of women’s to men’s median weekly full-time earnings was 82.1 percent (Hegewisch, Williams, Hartmann, Hudiburg, 2014). In other words, women receive only 82 cents for every dollar earned by men in the United States.
The Netherlands is recognized worldwide for its equality. Moreover, according to the Gender Inequality Rank Index of the United Nations of 2012, the Netherlands is seen as on of the most gender equal country among the 148 countries all over the world (Gender Inequality Index, 2012). But is the gender income gap in the Netherlands really that narrow? And does the educational success of women have effect on their earnings in later life?
To answer these questions, I have conducted a search through the different open databases available on the Internet about Education and Income. For this search, Quandl1 came out to be very useful. Quandl is a data search engine that contains millions of datasets, such as data from EuroStat2 or LabourStata3. When someone has found a dataset that could be useful, it is possible to explore this dataset graphically.
1.2 Objectives
The first objective for this paper is to explore open databases, which contain data about education and the labour market industry. Using these databases, I will try to investigate whether there is an income gap between male and female workers in the Netherlands, and whether that difference in income is affected by education level, using the data about annual earnings and educational attainment. Therefore, my research question is: What is the extent of the gender income gap in the Netherlands and does the income gap vary at different levels of education?
1 Quandle.com 2
http://epp.eurostat.ec.europa.eu/portal/page/portal/statistics/search_database
3 http://laborsta.ilo.org/
To answer this research question, first previous studies were explored in a literature review. After that, the propositions and conceptual framework were explained. Next, the materials and methodology that were used were stated, followed by the results and discussion. Lastly, I provided give the conclusion and ended with recommendations for future research.
1.3 Scope and limitations
The study will include all male and female workers in the Netherlands. It will deal with their annual earnings, and the highest level of education they have attained. The study will exclude male and female workers in other countries. The data that will be used to test the propositions and at the end the research question, is retrieved from an existing dataset that is online and publicly available.
The first limitation of this study is that it cannot be used to generalize to countries other than the Netherlands. Considering the limited budget and time for this study, the preference goes out to an existing database. This is another limitation, because the data is not collected especially for this study.
Another limitation which is a consequence of the fact that I do not collect my own data is that according to many scholars it is important when studying the income differences to ensure that the sample consist mostly out of relatively young workers because their earnings are more likely to reflect current labor market conditions than older workers' earnings, which embody historical changes as well (Loury, 1997). Looking only at the beginning of careers probably provides an accurate picture of differences further into the life cycle (Blackburn, Mckinley, Bloom, Freeman, 1990).
1.4 Impact and contribution
Many studies have tackled the gender income gap. Although data shows that the gap is narrowing, there is still a difference in income between male and female workers. The most gender equal country in the world is thought to be the Netherlands. However, there are still statements from institutions like WOMAN Inc. about the gender income gap, like the recent campaign ‘Where is my €300.000,-?’ stating that women earn in their lifetime a total of €300.000,- less than men. (WOMAN Inc., 2014). The purpose of this paper is to focus especially on the income difference in the Netherlands between male and female workers, and whether it is possible to overcome that difference by getting a higher educational degree.
2. LITERATURE REVIEW
In this section, existing literature and prior research will be examined and compared to understand the current status of this field of research. Different studies will be used to compare their findings.
Human capital consists of different qualifications, namely, knowledge acquired through formal education and skills, competencies, and expertise acquired through work experience. In the standard economic model, the accumulation of human capital is seen as an investment decision, where an individual person gives up some proportion of his income during the period of education and training in return for increased earnings in the future (Blundell, Dearden, Meghir and Sianesi, 1999). So an individual will only want to spend time and money on a study or additional training if expected higher earnings compensate the costs of that study in the future and the time invested.
Whether that person is a male or female that follows an education or training should be indifferent for the increase in earnings that follow. However, many studies have indicated that women earn significantly less than men, although they perform the same jobs (e.g. Loury, 1997 & Bobbit-Zeher, 2007), which is still as mentioned earlier, almost 20% less.
A consequence of the gender socialization is that men and women have different values, expectations and choose different occupations. In the 80s, Daymont and Andrisani (1984) stated that when men select a job or career, for them it is more important to make a lot of money than for women who are selecting a job. Furthermore, men want a job that gives them an opportunity to be a leader, while women think it is more important to be helpful to others or to society. This suggests that because of different values and expectations women and men choose different studies and different careers, which leads to different earnings.
Education does not only compensate for the time and money spend, according to different studies it would also reduce the difference in gender incomes (Gill & Leigh 2000; Loury 1997). Research has shown that the gender wage gap has narrowed significantly since 1985 (Gill & Leigh, 2000). According to O’neill and Polachek (1993) this change can be explained by the following facts: the level and return to women’s years of labour market experience has increased, the discrimination towards women in the labour market has declined, the gap in the proposition of
college graduates is closing and the return to education for women has improved relatively. Loury (1997) also states that the narrowing income gap is caused by the success of women’s education. Another reason Loury mentioned as cause of the declining gender gap in the 80s is the change of perceptions about the appropriate role of women in the labor market, in other words, the declining discrimination towards women.
Blackburn and Neumark (1993) showed in their study that the difference in income due to education only occurred for workers with relatively high levels of academic ability, and thus only affects the gender income gap for higher educated workers. This is emphasized by Loury, who says that the grades that are earned at high school do not have the same effect on the productivity as college grades (1997).
Nowadays women perform much better in education and more women are enrolling in college than men. Giving the importance of educational attainment for labour market success (Farley, 1995) the increasing participation and success of women in higher education suggests that young women who are entering careers might finally have earnings comparable to men. However, college-educated men still earn in their mid-20s already, on average, about €7,000 more per year than do college-educated women. When they have the same college major, cognitive skills, selectivity of the college from which they graduated, women still earn about €4,400 less per year than men (Bobbit-Zeher, 2007).
So on the one side, studies state that following education is a way to narrow the gender income gap, but on the other side studies show that although men and women had the same education and perform the same job, they still earn significantly less than their male colleague. In this study I will try to find out which scholars are right.
3. CONCEPTUAL FRAMEWORK 3.1 Model
Using the knowledge acquired from the literature review, in this section the relationships between the different variables and the propositions used to answer the research question will be shown in a conceptual model.
Figure 1. The conceptual model of this study
Figure 1 shows the propositions graphically in the conceptual model. The connection between Gender and Income represents the first two propositions. The right-hand arrow indicates the third proposition. The bottom-hand arrow represents the moderation effect of education on the gender income gap.
3.2 Propositions
Because prior research shows evidence that there is still a difference in the annual earnings between male and female workers in the Netherlands, my first proposition is that there is a difference in income level between male and female workers in the Netherlands. Secondly, I expect that male workers earn on average more than female workers. So, my second proposition is that the average income level of male workers is greater than female workers in the Netherlands.
Because studies show that the time and money spend on education, leads to higher compensation in return, my third proposition is that education has a positive relationship with income.
My fourth proposition is that the income difference between men and women narrows down at higher education levels, because I expect that getting a higher educational degree can narrow the gender income gap.
Proposition 1a: There is a difference in income between male and female workers in the Netherlands.
Proposition 1b: The average income of male workers is greater than female workers in the Netherlands.
Proposition 2: The average income increases as the education level increases.
Proposition 3: The difference of income between male and female workers decreases as the education level increases.
4. MATERIALS AND METHODOLOGY
4.1 Data source
As mentioned earlier, the first objective for this paper is to explore open databases about education and the labour market. There are thousands of databases that contain data about education or about the labour market. Quandl, a data search engine, offers free and unlimited access to 8 million datasets from 400 sources e.g. the United Nations, Worldbank and the World Health Organization, with data about e.g. finance, economics and health (Quandl, 2014). For instance, when you search for datasets containing data about education and labour, it shows more than 200,000 different datasets. When searching for data about gender and income, at the site of EuroStat, the data set cannot be found at Income and Living Condition but at Labour market. However, with Quandl, it is easier to search for datasets per subjects.
To test my propositions, and find an answer to my research question, I have used a database from EuroStat4 called The Structure of Earnings Survey (SES). This is the most up to date dataset, containing income levels measured at different education levels and gender in the Netherlands.
It is a survey, which provides EU-wide, harmonised structural data on gross earnings, hours paid and annual days of paid holiday leave. It is collected every four years, using eDAMIS. eDAMIS (electronic Data files Administration and Management Information System) is a web portal that Eurostat use and support to collect and transfer data files in the European Statistical System. The EuroGroups Register uses eDamis to send and retrieve its country specific datasets to Member State (EGR, 2011).
The latest version is 2010, which I will use. Here the annual earnings are gathered and sorted by personal characteristics like sex, age, economic activity and educational attainment for the 27 countries in the European Union. (EuroStat, 2012). For my study I will only use the data from the workers in the Netherlands.
4 http://epp.eurostat.ec.europa.eu/portal/page/portal/labour_market/earnings/database
4.2 Methodology
For the analysis, SPSS 21 was used. First I showed a summary of statistics with a descriptive table including for example the means and SD of the variables. To answer the propositions, I have run a factorial ANOVA. Factorial ANOVA is used to measure whether a combination of independent variables predict the value of a dependent variable. Secondly, with this test it is possible to test if the effect of one independent variable on the dependent variable is the same across all level of the other independent variable. (Maths Statistics, 2010). Thus analysing if there is any interaction between the independent variables. The data will be presented in a table and a graph, followed by a discussion.
Aside from the output provided by SPSS, to test the propositions properly, it is also important to analyse how strongly the two independent variables are related with the dependent variable, and if there is a difference between the two. This is measured with the eta squared (η2) (Levine & Hullett, 2002). In the output from the Factorial ANOVA, SPSS only produces the partial eta squared as estimated values. Because when having a more complex ANOVA, e.g. a Factorial ANOVA, the partial eta squared does not give the correct estimated value (Levine & Hullett, 2002), I have calculated the correct Eta Squared separately.
Table 1.
Design Table for variable ‘Group’
Male Female Edu lvl. ED 0 & 1 1 2 ED 2 3 4 ED 3 & 4 5 6 ED 5A 7 8 ED 5B 9 10 ED 6 11 12
As seen in Table 1, this factorial ANOVA model represents 2 x 6 design, because the independent variable Gender has two levels and the moderating variable Education level has six levels.
However, factorial ANOVA only indicates whether there is an interaction between the variables, and whether it is of significance. To state the direction of the interaction effect, and therefore support my last prediction, the simple main effects approach is used to examine the interaction effect. This is done by creating a new variable, named ‘group’, representing the twelve different factors mentioned in Table 1. With this variable, it is possible to test the direction of the interaction effect. The Post Hoc Scheffe test is a method for adjusting significance levels in a linear regression analysis to account for multiple comparisons (Scheffe’s Method, 2013). This test is used to analyse the significance of the difference in means between the variables, and therefore to answer the last proposition.
4.3 Data description
Table 2.
Summary of the variables
N Minimum Maximum Mean Std. Deviation
Gender 666 1 2 1,48 ,500
Edu. lvl 666 0 5 2,42 1,670
Income 666 €19,297.00 €117,974.00 €47,039.2252 €19,833.86118 Valid N (listwise) 666
Income is the dependent variable (DV) and was constructed as individual annual earnings, measured in Euro. In this study, the effect of tax e.g. was ignored. The sample consisted of a total of 696 participants. However 30 participants had no income level, so those were deleted.
Education level is the moderating variable (M) and is measured according to the standard of the United Nations, the International Standard Classification of Education (ISCED). This allows comparisons of education statistics and indicators across countries on the basis of uniform and internationally agreed definitions (UNESCO, 2014). The version used here is the second version, created in 1997.
There are seven levels within the ISCED97: Pre-primary and primary education (levels 0 and 1), lower secondary or second stage of basic education (level 2), upper secondary and post-secondary non-tertiary education (levels 3 and 4), first stage of tertiary education - programs that are theoretically based/research preparatory or giving access to professions with high skills requirements (level 5A), first stage of tertiary education - programs which are practically oriented and occupationally specific (level 5B) and second stage of tertiary education leading to an advanced research qualification (level 6).
The independent variable (IV), Gender, has a straightforward definition. According to the Oxford Dictionary (2014), gender means “The state of being male or female”.
5. RESULTS AND DISCUSSION
5.1 Descriptives Table 3.
Descriptive Statistics of the design table
Gender Edu. lvl Mean Std. Deviation N
M ED 0 & 1 €32,294.5862 €5,722.70085 58 ED 2 €34,861.5172 €7,276.25076 58 ED 3 & 4 €42,621.1552 €7,261.92700 58 ED 5A €64,224.2586 €€12,449.78182 58 ED 5B €58,733.7414 €13,108.97724 58 ED 6 €83,790.8519 €9,653.14347 54 Total €52,393.4622 €20,342.40974 344 F ED 0 & 1 €26,229.9444 €3,884.25453 54 ED 2 €28,276.9138 €4,406.22472 58 ED 3 & 4 €34,279.5000 €5,760.29854 58 ED 5A €53,904.2069 €19,247.20741 58 ED 5B €42,553.2778 €5,713.57720 54 ED 6 €70,894.0750 €9,239.59288 40 Total €41,319.1708 €17,585.01551 322 Total ED 0 & 1 €29,370.5625 €5,769.49684 112 ED 2 €31,569.2155 €6,840.92266 116 ED 3 & 4 €38,450.3276 €7,754.47171 116 ED 5A €59,064.2328 €16,949.89239 116 ED 5B €50,932.4464 €13,030.32545 112 ED 6 €78,302.8617 €11,401.81094 94 Total €47,039.2252 €19,833.86118 666 Note: Dependent variable is Income
Table 2 presents the descriptives of the design table. The average income of the workers is €47.039,23 (SD = €19.833,86). The education levels 2, 3 & 4 and 5A had equal samples, all 17,4% (N = 116). Levels 0 & 1 and 5B consist of 16,8% of the
sample (N = 112) and level 6 has 94 participants. 51,7% of the workers was male (N = 344) and 48,3% female.
Figure 2. Boxplot of education level on income
Figure 2 shows that the education level has a positive influence on the income, which could indicate support for proposition 3. Education level 5B stands out. This education level is the first stage of tertiary education, more practically oriented and occupationally specific. This is called in the Netherlands het Hoger Beroepsonderwijs. According to the International Standard Classification of Education this is a subcategory that is lower than the 5A level, e.g. the theoretically based education form.
Figure 3. Boxplot of gender on income
Figure 3 shows that the male workers in the Netherlands earn overall on average more than the female workers. However it is uncertain whether the proposition is supported. Figure 3 also indicates that there are some women who fall outside the boxplot, and earn a lot more than the average female worker.
5.2 Results of Analysis
In this study, I tried to examine whether there is a gender income gap in the Netherlands and if attaining education can narrow that income gap by answering the question what the extent of the gender income gap in the Netherlands is and if the income gap changes at different levels of education.
To find an answer to the research question, I had four predictions. Because there are still many sounds in popular media that women earn still less than men, I expected that there is a difference in income between male en female workers and moreover, male workers in the Netherlands earn on average more than female
workers in the Netherlands. Next, I expected that when someone, male or female had a higher education level, he or she would earn more than a worker who had a lower
education level. Lastly I expected that the difference of income between male and female workers decreases as the education level increases. In other words, attaining higher education can narrow the gender income gap.
Table 4.
Tests of Between-Subjects Effects
Source Type III Sum of Squares df Mean Square F Sig. Partial Eta Squared Corrected Model 200166755131,977a 11 18196977739,271 193,723 ,000*** ,765 Intercept 1500980400435,380 1 1500980400435,380 15979,233 ,000*** ,961 Gender 16690881578,152 1 16690881578,152 177,689 ,000*** ,214 Edu. lvl 174240762018,924 5 34848152403,785 370,989 ,000*** ,739 Gender * Edu. lvl 2114659543,564 5 422931908,713 4,502 ,000*** ,033 Error 61432307634,239 654 93933192,101 Total 1735249743486,000 666 Corrected Total 261599062766,217 665 a. R Squared = ,765 (Adjusted R Squared = ,761)
Note: Dependent variable is Income, N = 666 *p < .05, **p < 0.01, ***p < .001
Table 5.
Measures of Association
Variables η η2
Income * Gender 0,279 0,078 Income * Edu. Lvl 0,833 0,694
Table 4 shows the factorial ANOVA. This test reveals that there is a significant main effect of Gender on the Income level (F(1,654) = 177,689, p <.001). This suggests that
there is a gender income gap, as predicted earlier. More precisely, it shows that male workers in the Netherlands (M = €52,393.46, SD = €20,342.41) earn overall more than the female workers in the Netherlands (M = €41,319.17, SD = €17,585.02). As seen in Table 5, the eta squared is 0,078. Thus difference in income between male and female workers in the Netherlands, regardless the education level is about 7,8%. This gives evidence to support the first two propositions, e.g. that there is a difference in income between male and female workers in the Netherlands and that the average income for male workers is greater than female workers.
There was also a significant main effect of Education Level on the Income Level (F(5,654) = 370,989, p <.001), this means that when a worker in the Netherlands
has an Education level of 0 or 1 (M = €29,370.56, SD = €5,769.50) he earns a lot less than when he has a higher level of Education, for example 5A (M €59,064.23, SD = €16,949.89). More precisely, the eta squared is 0,694. This means that the difference in income – without the gender difference taken into account – is for 69,4% explained by the education level. This supports the third proposition, namely that the average income increases as the education level increases.
There was also a significant interaction between the two factors (F(5,654) =
4,502, p <.001). So there is a significance effect of Gender and Education Level on Income. However, with this test, it is impossible to say anything about the direction of the interaction effect, thus whether the gender income gap increases or decreases when the education level increases.
Figure 4. Interaction between Gender and Education Level on Income
Figure 4 shows the effects from the tables above. First you can see that for every level of education, male workers earn more than female workers. Secondly the income for both female and male workers rises when the level of education increases (As mentioned earlier, the 5B level is lower than 5A). Lastly, the interaction effect can be seen. All the slopes are negative, but there are not parallel. This means that there is an ordinal interaction. The slopes of the lower lines are less steep than the higher education level lines.
Table 6.
Multiple Comparisons, Scheffe’s Method
Groups Mean Difference Sig.
1 2 €6,064.64176 0,449 3 4 €6,584.60345 0,272 5 6 €8,341.65517* 0,031 7 8 €10,320.05172* 0,001 9 10 €16,180.46360* 0,000 11 12 €12,896.77685* 0,000 Note: *p < .05
In Table 6 the important outcomes of the simple main effects approach using the Scheffe’s method are shown. All the outcomes of this test can be found in de Appendix. As seen in the Table, the first two mean differences (€6,064.64, €6,584.60, ns) do not indicate that there is an interaction effect. So the Education Levels 0, 1 and 2 there is not a significant difference of income level for male and female workers.
However, the next four comparisons are significant (€8,341.66, €10,320.05, €16,180.46, €12,896.78, p < .05). This mean, as mentioned before, that there is a significant interaction effect. Table 6 indicates that, in contrast to my prediction, the difference in income level between male and female workers in the Netherlands who have a lower education level (€8,341.66, p < .05) is smaller than those with a higher education level (€12,896.78, p < .05). This indicates that the gender income gap does not decreases when the Education Level increases.
Thus the last proposition was not supported, the difference between income of the male and female workers in the Netherlands did not reduce when the education level rises. On contrary to the predictions, the income difference increased when the education level increased.
Although prior research indicated that it is possible for women to overcome the gender income gap by following higher education (e.g. Blackburn and Neumark, 1993; Loury, 1997; O’neill and Polachek, 1993) my findings support the study of Bobbit-Zeher (2007), that is to say that although women follow higher educational programs, the difference in income still increases.
A possible explanation is that women major in fields that are less rewarded with higher income (Bobbit-Zeher, 2007). In that study he shows that women are more likely to graduate from education, arts, humanities, social sciences and law, and men are more likely to graduate from natural sciences, mathematics and engineering. And beta majors tend to lead to higher incomes than the humanity majors.
Other factors that have a great influence in the height of the income level are for instance the years of work experience, choice of the major and grade point average (Loury, 1997).
6. CONCLUSION 6.1 General Conclusion
In this paper I tried to answer the question whether there is a gender income gap in the Netherlands and if getting a higher educational degree can narrow that income gap. To answer this question I used four propositions. Although I found a positive relationship between male and income and between education and income, the income gap cannot be narrowed with education. Moreover, I found evidence that suggest that the income gap grows as the education level increases.
6.2 Implications and Future Research
The comparison between the income of male and female workers for the first two levels was not significant. This can be explained by the fact that the first two education levels – ED 0 & 1 and 2 – are only basic education. At this level, education does not make such a great difference.
The fact that my last prediction appears to have been proven invalid could be explained by the fact that my own prediction was a bit biased. Prior research showed evidence both convincing and counters my prediction that it was possible to overcome the gender income gap trough educational attainment.
Another reason that my prediction was falsified could be because the sample consisted of an equal age distribution. According to Bobbit-Zeher, when someone wants to study the effect on the gender income gap, the first years of a career are the most important. Because then there are minimal differences between employment histories, life experiences and accumulated skills, and the educational credentials and school experiences are likely to matter the most (Bobbit-Zeher, 2007). Collection own data, or using data that is more focused on the younger workers could solve this problem.
As mentioned before, the income difference could be explained by the fact that women tend to study and subsequently work in sectors that are less rewarding (Bobbit-Zeher, 2007). A suggestion for future research is therefor that the different sectors are added to the analysis. Then, it is possible to check whether the income difference depends on the sector, and if these different sectors explain more of the income gap.
7. REFERENCES
Bobbitt-Zeher, D. (2007) The Gender Income Gap and the Role of Education, Sociology of Education, Vol. 80, No 1 (Jan. 2007), pp. 1-22
Blackburn, McKinley, Bloom, and Freeman. (1990) "The Declining Position of Less Skilled American Men." Gary Burtless, ed. ,A Future of Lousy Jobs? Washington, D.C.: Brookings Institution, pp. 31-76.
Blackburn, McKinley, and Neumark, D. (1993) Omitted-Ability Bias and the Increase in the Return to Schooling, Journal of Labor Economics, Vol. 11, No. 3 (July), pp. 521-44
Blundell, R., Dearden, L., Meghir, C., Sianesi, B., (1999) Human Capital Investment: The Returns from Education and Training to the Individual, the Firm and the Economy Fiscal Studies (1999) vol. 20, no. 1, pp. 1–23
Daymont, T. and Andrisani, P. (1984) Job Preferences, College Major, and the Gender Gap in Earnings. Journal of Human Resources 19:408-28. EGR (2011) http://egr.istat.it/?q=node/231, visited on 25/6/2014
EuroStat (2012) http://epp.eurostat.ec.europa.eu/cache/ITY_SDDS/EN/earn_ses 2010_esms.htm, visited on 16/6/2014
Farley, R. (1995) State of the Union: Americain the 1990s: Volume One. Economic Trends. New York: Russell Sage Foundation.
Gender Inequality Index (2012) https://data.undp.org/dataset/Table-4-Gender-Inequality-Index/pq34-nwq7, visited on 19/5/14
Gerber, P., and Schaefer D. (2004) Horizontal Stratification of Higher Education in Russia: Trends, Gender Differences, and Labor Market Outcomes, Sociology of Education 77:32–59.
Gill, A., Leigh, D. (2000) Community College Enrolment, College Major, and the Gender Wage Gap, Industrial and Labor Relations Review 54:163-81.
Hegewisch, A., Williams, C., Hartmann, H., Hudiburg S.K., (2014) The Gender Wage Gap: 2013; Differences by Race and Ethnicity, No Growth in Real Wages for Women http://www.iwpr.org/publications/pubs/the-gender-wage-gap-2013-
Levine, T., Hullett, C. (2002) Eta Squared, Partial Eta Squared and Misreporting of Effect Size in Communication Research, Human Communication Research, Vol. 28, No. 4, October 2002, 612-625
Loury, L., (1997) The Gender Earnings Gap Among College-educated Workers, Industrial and Labor Relations Review, Vol. 50, No. 4 (July 1997)
Maths Statistics (2010) http://www.maths-statistics- tutor.com/two_way_factorial_ANOVA_pasw_spss.php, Visited on 5/7/2014 O’neill, J., Polachek, S., (1993) Why the Gender Gap in Wages Narrowed in the
1980s, Journal of Labor Economics, Vol. 11, Part 1 (January) pp 205-28. Oxford Dictionary (2014)
http://www.oxforddictionaries.com/definition/english/gender, visited on 19/5/14
Quandl (2014) http://www.quandl.com/, visited on 5/7/2014
Scheffe’s Method (2013) http://en.wikipedia.org/wiki/Scheff%C3%A9%27s_method, visited on 7/7/2014
UNESCO (2014) http://www.uis.unesco.org/Education/Pages/international-standard-classification-of-education.aspx, visited on 14/5/’14
8. APPENDIX
8.1 Multiple Comparisons Table 6.
Dependent Variable: Income Scheffe
(I) group (J) group Mean Difference (I-J)
Std. Error Sig. 95% Confidence Interval Lower Bound Upper Bound
1,00 2,00 €6,064.64176 €1,832.76877 ,449 -€2,098.0969 €14,227.3805 3,00 -€2,566.93103 €1,799.74320 ,998 -€10,582.5813 €5,448.7193 4,00 €4,017.67241 €1,799.74320 ,931 -€3,997.9779 €12,033.3227 5,00 -€10,326.56897* €1,799.74320 ,001 -€18,342.2193 -€2,310.9187 6,00 -€1,984.91379 €1,799.74320 1,000 -€10,000.5641 €6,030.7365 7,00 -€31,929.67241* €1,799.74320 ,000 -€39,945.3227 -€23,914.0221 8,00 -€21,609.62069* €1,799.74320 ,000 -€29,625.2710 -€13,593.9704 9,00 -€26,439.15517* €1,799.74320 ,000 -€34,454.8055 -€18,423.5049 10,00 -€10,258.69157* €1,832.76877 ,001 -€18,421.4303 -€2,095.9529 11,00 -€51,496.26564* €1,832.76877 ,000 -€59,659.0044 -€43,333.5269 12,00 -€38,599.48879* €1,991.95070 ,000 -€47,471.1879 -€29,727.7897 2,00 1,00 -€6,064.64176 €1,832.76877 ,449 -€14,227.3805 €2,098.0969 3,00 -€8,631.57280* €1,832.76877 ,025 -€16,794.3115 -€468.8341 4,00 -€2,046.96935 €1,832.76877 1,000 -€10,209.7081 €6,115.7694 5,00 -€16,391.21073* €1,832.76877 ,000 -€24,553.9494 -€8,228.4720 6,00 -€8,049.55556 €1,832.76877 ,059 -€16,212.2943 €113.1832 7,00 -€37,994.31418* €1,832.76877 ,000 -€46,157.0529 -€29,831.5755 8,00 -€27,674.26245* €1,832.76877 ,000 -€35,837.0012 -€19,511.5237 9,00 -€32,503.79693* €1,832.76877 ,000 -€40,666.5356 -€24,341.0582 10,00 -€16,323.33333* €1,865.20967 ,000 -€24,630.5565 -€8,016.1102 11,00 -€57,560.90741* €1,865.20967 ,000 -€65,868.1306 -€49,253.6842 12,00 -€44,664.13056* €2,021.83910 ,000 -€53,668.9459 -€35,659.3152 3,00 1,00 €2,566.93103 €1,799.74320 ,998 -€5,448.7193 €10,582.5813 2,00 €8,631.57280* €1,832.76877 ,025 €468.8341 €16,794.3115 4,00 €6,584.60345 €1,799.74320 ,272 -€1,431.0468 €14,600.2537 5,00 -€7,759.63793 €1,799.74320 ,072 -€15,775.2882 €256.0124 6,00 €582.01724 €1,799.74320 1,000 -€7,433.6330 €8,597.6675 7,00 -€29,362.74138* €1,799.74320 ,000 -€37,378.3917 -€21,347.0911 8,00 -€19,042.68966* €1,799.74320 ,000 -€27,058.3399 -€11,027.0394 9,00 -€23,872.22414* €1,799.74320 ,000 -€31,887.8744 -€15,856.5739 10,00 -€7,691.76054 €1,832.76877 ,094 -€15,854.4992 €470.9782 11,00 -€48,929.33461* €1,832.76877 ,000 -€57,092.0733 -€40,766.5959 12,00 -€36,032.55776* €1,991.95070 ,000 -€44,904.2569 -€27,160.8586
4,00 1,00 -€4,017.67241 €1,799.74320 ,931 -€12,033.3227 €3,997.9779 2,00 €2,046.96935 €1,832.76877 1,000 -€6,115.7694 €10,209.7081 3,00 -€6,584.60345 €1,799.74320 ,272 -€14,600.2537 €1,431.0468 5,00 -€14,344.24138* €1,799.74320 ,000 -€22,359.8917 -€6,328.5911 6,00 -€6,002.58621 €1,799.74320 ,434 -€14,018.2365 €2,013.0641 7,00 -€35,947.34483* €1,799.74320 ,000 -€43,962.9951 -€27,931.6945 8,00 -€25,627.29310* €1,799.74320 ,000 -€33,642.9434 -€17,611.6428 9,00 -€30,456.82759* €1,799.74320 ,000 -€38,472.4779 -€22,441.1773 10,00 -€14,276.36398* €1,832.76877 ,000 -€22,439.1027 -€6,113.6253 11,00 -€55,513.93806* €1,832.76877 ,000 -€63,676.6768 -€47,351.1994 12,00 -€42,617.16121* €1,991.95070 ,000 -€51,488.8603 -€33,745.4621 5,00 1,00 €10,326.56897* €1,799.74320 ,001 €2,310.9187 €18,342.2193 2,00 €16,391.21073* €1,832.76877 ,000 €8,228.4720 €24,553.9494 3,00 €7,759.63793 €1,799.74320 ,072 -€256.0124 €15,775.2882 4,00 €14,344.24138* €1,799.74320 ,000 €6,328.5911 €22,359.8917 6,00 €8,341.65517* €1,799.74320 ,031 €326.0049 €16,357.3055 7,00 -€21,603.10345* €1,799.74320 ,000 -€29,618.7537 -€13,587.4532 8,00 -€11,283.05172* €1,799.74320 ,000 -€19,298.7020 -€3,267.4014 9,00 -€16,112.58621* €1,799.74320 ,000 -€24,128.2365 -€8,096.9359 10,00 €67.87739 €1,832.76877 1,000 -€8,094.8613 €8,230.6161 11,00 -€41,169.69668* €1,832.76877 ,000 -€49,332.4354 -€33,006.9580 12,00 -€28,272.91983* €1,991.95070 ,000 -€37,144.6190 -€19,401.2207 6,00 1,00 €1,984.91379 €1,799.74320 1,000 -€6,030.7365 €10,000.5641 2,00 €8,049.55556 €1,832.76877 ,059 -€113.1832 €16,212.2943 3,00 -€582.01724 €1,799.74320 1,000 -€8,597.6675 €7,433.6330 4,00 €6,002.58621 €1,799.74320 ,434 -€2,013.0641 €14,018.2365 5,00 -€8,341.65517* €1,799.74320 ,031 -€16,357.3055 -€326.0049 7,00 -€29,944.75862* €1,799.74320 ,000 -€37,960.4089 -€21,929.1083 8,00 -€19,624.70690* €1,799.74320 ,000 -€27,640.3572 -€11,609.0566 9,00 -€24,454.24138* €1,799.74320 ,000 -€32,469.8917 -€16,438.5911 10,00 -€8,273.77778* €1,832.76877 ,043 -€16,436.5165 -€111.0391 11,00 -€49,511.35185* €1,832.76877 ,000 -€57,674.0906 -€41,348.6131 12,00 -€36,614.57500* €1,991.95070 ,000 -€45,486.2741 -€27,742.8759 7,00 1,00 €31,929.67241* €1,799.74320 ,000 €23,914.0221 €39,945.3227 2,00 €37,994.31418* €1,832.76877 ,000 €29,831.5755 €46,157.0529 3,00 €29,362.74138* €1,799.74320 ,000 €21,347.0911 €37,378.3917 4,00 €35,947.34483* €1,799.74320 ,000 €27,931.6945 €43,962.9951 5,00 €21,603.10345* €1,799.74320 ,000 €13,587.4532 €29,618.7537 6,00 €29,944.75862* €1,799.74320 ,000 €21,929.1083 €37,960.4089 8,00 €10,320.05172* €1,799.74320 ,001 €2,304.4014 €18,335.7020 9,00 €5,490.51724 €1,799.74320 ,594 -€2,525.1330 €13,506.1675 10,00 €21,670.98084* €1,832.76877 ,000 €13,508.2421 €29,833.7196 11,00 -€19,566.59323* €1,832.76877 ,000 -€27,729.3319 -€11,403.8545
12,00 -€6,669.81638 €1,991.95070 ,427 -€15,541.5155 €2,201.8828 8,00 1,00 €21,609.62069* €1,799.74320 ,000 €13,593.9704 €29,625.2710 2,00 €27,674.26245* €1,832.76877 ,000 €19,511.5237 €35,837.0012 3,00 €19,042.68966* €1,799.74320 ,000 €11,027.0394 €27,058.3399 4,00 €25,627.29310* €1,799.74320 ,000 €17,611.6428 €33,642.9434 5,00 €11,283.05172* €1,799.74320 ,000 €3,267.4014 €19,298.7020 6,00 €19,624.70690* €1,799.74320 ,000 €11,609.0566 €27,640.3572 7,00 -€10,320.05172* €1,799.74320 ,001 -€18,335.7020 -€2,304.4014 9,00 -€4,829.53448 €1,799.74320 ,782 -€12,845.1848 €3,186.1158 10,00 €11,350.92912* €1,832.76877 ,000 €3,188.1904 €19,513.6678 11,00 -€29,886.64496* €1,832.76877 ,000 -€38,049.3837 -€21,723.9062 12,00 -€16,989.86810* €1,991.95070 ,000 -€25,861.5672 -€8,118.1690 9,00 1,00 €26,439.15517* €1,799.74320 ,000 €18,423.5049 €34,454.8055 2,00 €32,503.79693* €1,832.76877 ,000 €24,341.0582 €40,666.5356 3,00 €23,872.22414* €1,799.74320 ,000 €15,856.5739 €31,887.8744 4,00 €30,456.82759* €1,799.74320 ,000 €22,441.1773 €38,472.4779 5,00 €16,112.58621* €1,799.74320 ,000 €8,096.9359 €24,128.2365 6,00 €24,454.24138* €1,799.74320 ,000 €16,438.5911 €32,469.8917 7,00 -€5,490.51724 €1,799.74320 ,594 -€13,506.1675 €2,525.1330 8,00 €4,829.53448 €1,799.74320 ,782 -€3,186.1158 €12,845.1848 10,00 €16,180.46360* €1,832.76877 ,000 €8,017.7249 €24,343.2023 11,00 -€25,057.11047* €1,832.76877 ,000 -€33,219.8492 -€16,894.3718 12,00 -€12,160.33362* €1,991.95070 ,000 -€21,032.0328 -€3,288.6345 10,00 1,00 €10,258.69157* €1,832.76877 ,001 €2,095.9529 €18,421.4303 2,00 €16,323.33333* €1,865.20967 ,000 €8,016.1102 €24,630.5565 3,00 €7,691.76054 €1,832.76877 ,094 -€470.9782 €15,854.4992 4,00 €14,276.36398* €1,832.76877 ,000 €6,113.6253 €22,439.1027 5,00 -€67.87739 €1,832.76877 1,000 -€8,230.6161 €8,094.8613 6,00 €8,273.77778* €1,832.76877 ,043 €111.0391 €16,436.5165 7,00 -€21,670.98084* €1,832.76877 ,000 -€29,833.7196 -€13,508.2421 8,00 -€11,350.92912* €1,832.76877 ,000 -€19,513.6678 -€3,188.1904 9,00 -€16,180.46360* €1,832.76877 ,000 -€24,343.2023 -€8,017.7249 11,00 -€41,237.57407* €1,865.20967 ,000 -€49,544.7973 -€32,930.3509 12,00 -€28,340.79722* €2,021.83910 ,000 -€37,345.6126 -€19,335.9819 11,00 1,00 €51,496.26564* €1,832.76877 ,000 €43,333.5269 €59,659.0044 2,00 €57,560.90741* €1,865.20967 ,000 €49,253.6842 €65,868.1306 3,00 €48,929.33461* €1,832.76877 ,000 €40,766.5959 €57,092.0733 4,00 €55,513.93806* €1,832.76877 ,000 €47,351.1994 €63,676.6768 5,00 €41,169.69668* €1,832.76877 ,000 €33,006.9580 €49,332.4354 6,00 €49,511.35185* €1,832.76877 ,000 €41,348.6131 €57,674.0906 7,00 €19,566.59323* €1,832.76877 ,000 €11,403.8545 €27,729.3319 8,00 €29,886.64496* €1,832.76877 ,000 €21,723.9062 €38,049.3837 9,00 €25,057.11047* €1,832.76877 ,000 €16,894.3718 €33,219.8492
10,00 €41,237.57407* €1,865.20967 ,000 €32,930.3509 €49,544.7973 12,00 €12,896.77685* €2,021.83910 ,000 €3,891.9615 €21,901.5922 12,00 1,00 €38,599.48879* €1,991.95070 ,000 €29,727.7897 €47,471.1879 2,00 €44,664.13056* €2,021.83910 ,000 €35,659.3152 €53,668.9459 3,00 €36,032.55776* €1,991.95070 ,000 €27,160.8586 €44,904.2569 4,00 €42,617.16121* €1,991.95070 ,000 €33,745.4621 €51,488.8603 5,00 €28,272.91983* €1,991.95070 ,000 €19,401.2207 €37,144.6190 6,00 €36,614.57500* €1,991.95070 ,000 €27,742.8759 €45,486.2741 7,00 €6,669.81638 €1,991.95070 ,427 -€2,201.8828 €15,541.5155 8,00 €16,989.86810* €1,991.95070 ,000 €8,118.1690 €25,861.5672 9,00 €12,160.33362* €1,991.95070 ,000 €3,288.6345 €21,032.0328 10,00 €28,340.79722* €2,021.83910 ,000 €19,335.9819 €37,345.6126 11,00 -€12,896.77685* €2,021.83910 ,000 -€21,901.5922 -€3,891.9615 *. The mean difference is significant at the 0.05 level.
