Abstract
As shown in previous research a relationship between tuition fee and student enrollment exists, this study focusses on this relationship moderated by government aid and career expectations. By manipulating these variables in a questionnaire with vignettes all other variables able to influence this relationship are controlled as much as possible. Results suggest that a variation in both government aid and career expectation influence students’ decision to enroll both for their first study as for a further study after graduation. The effect of career expectations seems to be the most crucial on a student’s enrollment choice.

How career expectations, tuition fees and government aid impact student enrollment
Education in all its forms is one of the fundamental elements of societies. By educating people a society can function and develop. This is especially the case in western countries in which knowledge, which can be acquired through education is of crucial importance. The need for high educated persons is of great importance for western countries such as member states of the European Union. The importance of higher education is reflected in strategies launched by the European Union (Rodriguez, al, 2010; European Commission, 2010), aiming towards a knowledge economy and becoming more competitive in future years. This vision is also being shared outside the European Union (Yang, 2011). The tuition fees in the Netherlands are rising and government aid is decreasing, which might cause a serious decrease in students, either continuing for a Master after a Bachelor degree or opting for a second study. This could have a negative impact on the Netherlands and its aim for a knowledge economy.
In recent years, as the financial crisis of Europe turned to a “peak”, the Netherlands have been under enormous budgetary pressure. The government has adopted various measures to reduce its spending. One of these measures consists of lowering the government aid for students. At the same time, there is a significant rise in tuition fees of higher education institutes in the Netherlands; it is assumed that this situation has caused increasingly more financial pressures on families and/or students.
In this research, academic theories and practical cases will be combined to research and examine this topic and the possible effects of career expectations, rising tuition fees, decreasing government aid on the enrollment behavior of Dutch students. By combining academic theories and published research, further data can be collected which should be of practical importance for universities and the government. In addition policy makers will better understand what drives students to their choices. The theoretical use will be a contribution to the field of education to provide a clear picture of students´ decision process and cost-benefit analyses.
The core research question to be addressed in this research is: whether there is a moderating effect of student’ career expectations and government aid on increasing tuition fees and student enrollment? Questions to be answered are:
- Do the published research findings, that students’ enrollment is affected by the gap between tuition fees and government aid, also apply to the Netherlands?
- Do changes in the gap between tuition fee and government aid have an impact on students’ ability to enroll?
- Do (positive) career expectations of students lower the effect of the gap between tuition fees and government aid on enrollment?
Several investigators have suggested that raising the tuition fees in higher education has caused problems and pressures for students and families. Simon Gompertz (2011) suggested that many adolescents are questioning whether they should go to university. The raising tuition fees are forcing the parents to save more, which results into pressures for the students and their family, potentially causing a decrease in enrollment. Christine Neill (2006)’s article discusses the effect of the tuition fee on students´ work. She concluded that there is evidence that rising student employment is strongly related to the rising in tuition fees. The German Council of Economics Expert, Malte Hübner (2004), did research on this topic and found that the tuition fees affect the enrollment behavior of the applicants (2004).
This is a cross-sectional lab experiment based on students as units of analysis and on deductive reasoning. This research will be done through vignettes and therefore is quantitative in nature. This research proposal consists of a problem statement, research question, literature review, conceptual framework, methodology and a reference list. The part following is the literature review.
Literature Review
Background
Tuition fees are raising in the Netherlands due to new government regulations. This has impacted students who wish to follow a second degree program or are opting for a Master degree. For instance at Maastricht University (the Netherlands) a Master study has a tuition fee of € 1,771 for Dutch students in the academic year 2012-2013, this is similar to other institutions. However if a student would like to follow a second Master study €12,000 is needed for enrollment only. This is based on article 7.46 of the Law on Higher Education and Scientific Research (Wet op het hoger onderwijs en wetenschappelijk onderzoek). Also diversification in tuition fee is allowed (Wet Versterking Besturing).
Another fact is that new plans are being discussed by the government to change the financial aid programs for students in two ways: transportation advantages and grants. At present students have a right to free public transportation. This allows them to travel without cost from their homes to university, to other universities for research purposes, minors and their internship for a maximum period of 7 years (based on enrollment) (Utrecht University, 2012). In addition every Dutch student can get a government grant and lend money from the government. The grant amount depends on parents’ income.
Some plans suggest that the government is aiming to change the system to a loan system in which students can lend money from the government for their study while at the same time grants are removed from the higher education system.
Government Aid
Literature differentiates between the different types of government aid. The most common types mentioned in literature are need-based aid which focusses on the student that are in financial need, merit-based aid which focusses on the results of students, loans, work-study aid is another option that government can choose to help students find part-time jobs next to their study and aid packages which are a combination of different types of government aid (Kim, DesJardins & McCall, 2009). Studies demonstrate that there is a significant relationship between government aid and student enrollment (Braunstein, Mcgrath & Pescatrice, 1999; Doyle, 2010; Kim, DesJardins & McCall, 2009). Both need-based and merit-based aid is type of grants from the government, which students do not have to repay if they finish their study. Previous research has shown that both grants and loans appear to have a positive effect on student enrollment in higher education. Work-study aid on the other hand appears to have a minimal impact on enrollment. When incorporated in an aid package it contributes to the packages’ attractiveness, this is based on a cross-sectional study at a single institution (Braunstein, Mcgrath & Pescatrice, 1999). Loans are a type of government aid that is often mentioned in literature. Its effectiveness is praised by some (Braunstein, Mcgrath & Pescatrice, 1999) and questioned by others (Long, 2010) although both sides recognize that there is a positive relationship between loans and enrollment. The reason why loans are not favored by all can be explained by some research stating the long-term indirect effects that loans can have on students. Such as, the fact that loans have to be repaid after graduation and consequences of dropping out of a study for students with a loan (Long, 2010). All above mentioned studies have the similarity of being from the United States of America. The Netherlands currently have a type of need-based aid with a minimum amount given to each student enrolling in higher education.
Other factors to be discussed when researching government aid are the variables which influence the type of aid most suited in a country. In literature several factors seem to have influence on students, first of all family income seem to be of importance. The level of family income seems to have a relationship with the response to the different types of government aids; this is especially true for the low-income and high/upper-income families. High-income families appear to be the least influenced by government aid (Braunstein, Mcgrath & Pescatrice, 1999). In this study we will not take family income/ parental aid into account, as the present Dutch grant system does differentiate for this. A variable that needs to be taken into account are the racial and ethnical differences (Kim, DesJardins & McCall, 2009). A different reaction on the types of aid can be expected when looking at various types of ethnical backgrounds. This variable will not be taken into account in this study, as most students receiving aid are Dutch or have the Dutch nationality. Another variable is the expectation of students. Some argue that there is a direct relationship between enrollment and the expectations of students about getting financial aid, and that it needs to be taken into account when opting for a type of government aid, this is based on a sample of over 80,000 students at the University of Iowa (Kim, DesJardins & McCall, 2009). Government aid might not only have an impact on enrollment and financial aspects (Braunstein, Mcgrath & Pescatrice, 1999; Long, 2010) but also on academic success. It appears that if students get more aid they may get higher grades and better prospects, this could be explained by the fact that they would have to work less in part-time jobs than they would have without the aid and thus have more time to study (Yang, 2011; Neill, 2006).
Tuition Fee
Tuition fees in higher education are an important factor in the cost-benefit analysis of students, becoming an increasingly important factor in recent times. Although worldwide there has been an overall rise in tuition fees, there are many inconsistencies that can be explained by regional differences, the population size of the state, economic development, public private higher education mix and state planning agencies (Hearn, Griswold & Marine, 1996). The effects of tuition fees on enrollment are ambiguous: On the one hand, high fees may represent important academic heritage and prestige, offering greater potential future earnings to its graduates. On the other hand, high tuition raises the cost of attending certain universities (price effect) (Dwenger, Storck & Wrohlic, 2009).
Literature shows that rising tuition fees affect the students’ work, and both of them are related to the enrollment decision. In both Canada and the US, the number of students’ working has increased substantially in the last two decades by facing the increasing tuition fee (Neill, 2006). Ahituv and Tienda (2004) found that the students increasing work is associated with a reduction in the probability of continuing with an additional year of education.
Student Enrollment
A major part of the literature on student enrollment focuses on the analysis and discussion of variables affecting enrollment. Amongst other macroeconomic factors discussed, research suggests that a macroeconomic factor such as inflation seem to have a relationship with school enrollment. This relationship is mostly explained by the effect that inflation has on total costs of attending school, which would have a negative impact on enrollment, but also the positive impact on enrollment that appears to be caused by the perception of students that studying may protect them from the negative impact of inflation risks in the future. This last mentioned factor refers to the perception of higher future disposable income. The impact of these macroeconomic factors related on enrollment seems to differ between males and females (Ewing, Beckert & Ewing, 2010).
As discussed earlier in the variable of government aid, family income appears to be of influence on enrollment (Braunstein, Mcgrath & Pescatrice, 1999) as does the variable of government aid itself. Government aid in the various forms of its existence has been proved to be related to school enrollment (Kim, DesJardins & McCall, 2009; Long, 2010). Tuiton fees impact on enrollment seems to be certain; some argue that it is perceived as being important for students (Hearn, Griswold & Marine, 1996; Dwenger, Storck & Wrohlic, 2009).
Career Expectations
The literature shows that the outcome expectations are considered to be the critical factors for the career interests and choices of students. The outcome expectation is defined as the consequence of a behavior or action, which may be specific to the academic performance. For instance, if a student has good grades in a study, this student would have a high career expectation (Feldt & Woelfel, 2009). A Chinese study showed that parental value systems are also an important factor that influences students’ career expectations. Parents encourage their children to pursue higher education in order to fulfill their career aspirations (Ma & Yeh, 2010). In addition, counselors also play a significant role in shaping students’ career goals. A good counselor would provide valuable information that could guide the students on shaping their career goals (Mau, Hitchcock & Calvert, 1998). While, the key factor of choosing a future career is self-efficacy, this is defined as the perceived level of confidence in one’s ability to perform various activities related to career planning and development (Feldt & Woelfel, 2009). It has been shown that self-efficacy has an impact on the future career, in research various instruments have been created to self-efficacy expectations associated with career choices (Bacanli, 2006).
Studies show that employability is a key issue for graduate students. However, it appears to be difficult to find a desirable career as easily as expected. A high career expectation would make the graduated students acquire a deeper knowledge or further study; therefore, they would like to enter a higher education system to fulfill their expectations for their further career. It is also indicated that in order to get a desirable career, the graduated students would like to continue studying, which has increased the enrollment as well (King, 2003).
Conceptual Framework
Figure 1 Conceptual framework
Methodology
This is a cross-sectional study based on students as units of analysis and on deductive reasoning. It is indicated in previous research that government aid and tuition fee have a significant impact on student enrollment. Therefore, this research will focus on the moderating effect of students’ career expectations and government aid on increasing tuition fees and student enrollment. In this case, the research will be conducted through a series of manipulations of the variables, which will be in a contrived study setting. Therefore, a lab experiment will be conducted in order to prove the causal relationship between the variables. Ethical considerations and anonymity are taken into account, the latter guaranteed.
The following series of manipulation will be used:
1. High tuition fee and low government aid.
2. High tuition fee and high government aid.
3. High tuition fee, low government aid and high career expectations.
4. High tuition fee, low government aid and low career expectations.
5. High tuition fee, high government aid and low career expectations.
Sampling
The population of this research is all Dutch university students. To obtain a proper representation of this population we will opt for combination of proportionate stratified sampling and random sampling. The population will be divided into strata; Universities of Applied Sciences and Universities of Science. A simple random sample will be picked out of each strata, this sample will be given vignettes to fill in. Tilburg University and Fontys University of Applied Sciences are representatives of the two types in the Netherlands. Therefore, representatives of the students from these two universities will be used as a sample. The sample size is determined by a series of statistical steps, which are showed below:
Determine continuous data according to Cochran’s formula assume the alpha level at .05, by using a five-point scale and set the acceptable level at 3%. Then get the first result of n。, which is 118. As the online figures show that the main population of Dutch university students is 91,035. Therefore, by using the second formula, the final sample size will be get, which is 118. The response rate of the online survey is predicted as 65%, therefore, the sample size will be 118/65%=181. To ensure the accuracy of the data from the respondents we would like to distribute the questionnaires face-to-face and the online survey. Around half of the sample will get from Fontys and Tilburg University and the other half will be gathered by online survey.

Figure 2 Cochran’s formula

Data Collection
In order to know if a relationship exists between the research variables we need to know the reaction of students towards a change in them. This change occurs over time and for this reason different scenarios (vignettes) will be used to measure the possible different reactions towards them. A disadvantage of using vignettes is that we cannot claim that the responses given are exactly the same as they would be in actual situations (Gliner, Haber & Weise, 1999). However this method provides standardized stimuli to all respondents, which enhances measurement reliability, internal validity, and makes it easier to replicate (Wason, Polonsky & Hyman, 2002). Due to its controlled scenarios the variables can be manipulated in such manner that the required data can be collected. By creating such a controlled scenario in which only certain variables are manipulated, noise is reduced.
The respondents getting the vignettes will be briefed and asked to participate. Each vignette will start with an introduction stating that participation is voluntary and that answers of the vignette are completely confidential and will never be associated with the respondent as an individual. Upon completion the respondent will be debriefed about the research and its purpose. Each respondent will get all vignettes to fill in; with this method it is possible to compare how an individual reacts in different situations. It allows analyzing within respondent and across respondent variance. The questionnaire will include 20 closed questions. Each of the variables will be measured by different scale types. For example dichotomous scales, ranking scales and a 5-point Likert scale running from: 1 Strongly Disagree, 2 Disagree, 3 Neutral, 4 Agree and 5 Strongly Agree.

Vignettes
Each vignette represents a different scenario in which one or more variables are being manipulated. This is of importance to measure students’ reactions in different situations. In total five vignettes will be filled in per respondent. First a series of demographic questions will be asked followed by the vignettes. The manipulation of these is as follows:
1. High tuition fee and low government aid.
2. High tuition fee and high government aid.
3. High tuition fee, low government aid and high career expectations.
4. High tuition fee, low government aid and low career expectations.
5. High tuition fee, high government aid and low career expectations.
All vignettes will be followed by the same questions to make the results comparable. This sequence of questions will always start with whether or not they would continue their current study in the mentioned situation and whether or not they (still) consider studying after graduating from their current study.
Operationalization
In order to process the acquired data of the vignettes the variables need to be operationalized and measurable. Tuition fee, government aid and career expectations will be expressed in a certain amount of euros. A distinction is made between levels; for instance career expectation is defined as a starter income of €1,200 per month at the low level and €2,400 at the high level, this is based on the minimum wage of 2012 in the Netherlands. The variable student enrollment will be explained as whether or not a student would enroll in a certain situation.

Data Analysis
Several types of analysis will be performed. Firstly, a sample description will be included. Secondly, to determine the relationship between tuition fee and student enrollment according to government aid and career expectation, a Chi-squared test will be used. This test will be repeated for the two types of questions in each vignette, being if the students would start their first study and if they would continue for further study in the situations.
The specific null hypotheses to be tested are shown below in null form:
1. H0: Assuming equal tuition fee, a variation in government aid does not influence the student enrollment. This will be tested by applying Chi-squared.
2. H0: Assuming equal tuition fee and government aid, a variation in career expectations does not influence the student enrollment. This will be tested by Chi-squared.
3. H0: Assuming equal tuition fee and career expectations a variation in government aid does not Influence student enrollment. This will be tested by Chi-squared.

Results
Descriptive statistics
This survey was done by combining online survey and field questionnaires. The questionnaires include two parts, general information and the vignettes. After receiving the questionnaires, the data of the respondents were analyzed.
General questions:
Gender Online TU Fontys Total Percent
Male 49 2 34 85 47.222%
Female 31 8 56 95 52.778% Age
20 or < 34 3 41 78 43.333%
21-30 46 7 49 102 56.667%
>30 0 0.000% Education
HBO 14 90 104 57.778%
WO 66 10 76 42.222% Aid
0-100 61 10 78 149 82.778%
101-200 3 3 1.667%
201-300 15 9 24 13.333%
301 - > 4 4 2.222% Highest level
No degree 0 0.000%
Bachelor 50 55 105 58.333%
Master 24 10 31 65 36.111%
Post-master 6 4 10 5.556% Starting income
1200-1600 7 7 3.889%
1601-2000 15 4 54 73 40.556%
2001-2400 47 5 20 72 40.000% higher 18 1 9 28 15.556%
Figure 3 General information of the respondents
Figure 2 provides the background information of the respondents, and those questions are closely related to the following vignettes. From the above tables we see that there are 180 respondents in total of the questionnaires, including 85 male students and 95 female. 56.7% of the respondents are aged from 21 to 30, the rest of the respondents are younger than 20. It is shown that 82.7% Dutch university students are getting the basic monthly government aid from 0 to 100 euro, whereas, others are getting more. In addition, approximately 58% of the Dutch students want to achieve a Bachelor degree and the rest of the students are wishing to achieve a higher degree of study. From the choice of the their expecting income we see that only 3.9% of the students expect to gain a income between 1,200 to 1,600 euro, while the others expect to gain a higher income.
The above information not only provides the background information of the respondents, but also provides the first choice of the university students after their graduation, whether they would opt for further study. Remarkably, these data changed after manipulating the tuition fee, government aid and career expectation in the vignettes which shows significant influence on the enrollment.
Vignettes:
The Vignettes provides questions under manipulation of the government aid and career expectation. As the questionnaires were executed by online and printed version, the results were accumulated separately. Under the conditions of each vignette, we calculated the number of students who wanted to start their first study and who opted for further study. In this questionnaire, the second question was a Likert scale, therefore, we decided to combine the ‘strongly agree’ and ‘agree’ together as an ‘Yes’ answer, ‘strongly disagree’ and ‘disagree’ as ‘No’. Meanwhile, half of the ‘Neutral’ data belonged to ‘Yes’, and the other half belonged to ‘No’. This approach has been repeated for the other vignettes and the final data is shown in the following tables.

Vignette 1 Online TU Fontys Total Percent
Start studying?
Yes 45 9 57 111 61.7%
No 35 1 33 69 38.3%
Continue studying?
Yes 26.5 3 52.5 82 45.6%
No 53.5 7 37.5 98 54.4%
Vignette 2 Online TU Fontys Total Percent
Start studying?
Yes 65 10 82 157 87.2%
No 15 0 8 23 12.8%
Continue studying?
Yes 54.5 3.5 41 99 55.0%
No 25.5 6.5 49 81 45.0%
Figure 4 Result of vignette 1 and 2
Vignette 1 provides the information in a situation with a high tuition fee of €3,000 and a low government aid of €50, 61.7% of the respondents would like to start their first study, and 45.6% students would like to continue their study after graduating from current study in this situation. In comparison with the data of vignette 2, after the government aid increased to €200, the number of respondents who would like to start their first study increased with approximately 25%; and the number of students who chose for further study also slightly increased.
Vignette 3 Online TU Fontys Total Percent
Start studying?
Yes 61 9 69 139 77.2%
No 19 1 21 41 22.8%
Continue studying?
Yes 42 6 47 95 52.8%
No 38 4 43 85 47.2%
Vignette 4 Online TU Fontys Total Percent
Start studying?
Yes 36 5 50 91 50.6%
No 44 5 40 89 49.4%
Continue studying?
Yes 28.5 1 39.5 69 38.3%
No 51.5 9 50.5 111 61.7%
Figure 5 Result of vignette 3 and 4
Vignette 3 provides the information in a situation with a high tuition fee of €3000, a monthly government aid of €50 and an expectation of a starting salary of €2,400. In this situation, 77.2% respondents would like to start their first study and 52.8% of them would like to study further. Compared with the data of vignette 4, with the same conditions of tuition fee and government aid, with an expected starting salary of €1,200, 50.6% of the respondents would like to start their first study and around 38.3% of students would like to study further. The above information shows that the enrolment would change significantly under different level of career expectation.
Vignette 5 Online TU Fontys Total Percent
Start studying?
Yes 43 6 63 112 62.2%
No 37 4 27 68 37.8%
Continue studying?
Yes 37.5 3.5 45 86 47.8%
No 42.5 6.5 45 94 52.2%
Figure 6 Result of vignette 5
Vignette 5 provides the information in a situation with a high tuition fee of €3000, a monthly government aid of €200 and an expected starting salary of €1,200. Compared with vignette 4, with a higher government aid 12% more respondents would like to start their first study and around 10% more students would like to go for further study.
Hypotheses
All the hypotheses are tested with a Chi-squared test for both first study and study after graduation. O E O-E (O-E)²/E

Hypothesis 1 : Government aid ( vignette 1 + 2)
Yes 157 111 46 19.06306306
No 23 69 -46 30.66666667 χ² 49.72972973
Hypothesis 2: Career expectation ( vignette 3+4)
Yes 139 91 48 25.31868132
No 41 89 -48 25.88764045 χ² 51.20632177
Hypothesis 3: Government aid with all other variables (vignette 4+5)
Yes 112 91 21 4.846153846
No 68 89 -21 4.95505618 χ² 9.801210026
Figure 7 Chi-squared test of for first study choice
Figure 6 provides the result of the Chi-squared tests for the first study choice. The first hypothesis is to test whether a variation in the government aid influences the student enrollment under the same amount of tuition fee. The observed vignettes are vignette 1 and 2. With a different amount of government aid, the study choice differences were tested by Chi-squared. As vignette 2 had a higher government aid €200, the data of vignette 2 was regarded as the observation data, the expected data should be the total number of the respondents in vignette 2 multiplied by the percentage of the study choices in vignette 1. The rejection region were set based on the formula v=(r-1)*(c-1), and the significant level were set as 5%. As the row total is three, and the column total is two, the rejection region is 3.84. The test statistic is 49.73, therefore, the null hypothesis can be rejected, there is enough evidence to infer that a variation of government aid influences the student enrollment under the same amount of tuition fee. The other tests following the same procedure, as the above table shows that the test statistic of the second Chi-squared test is 51.21, therefore, the null hypothesis can be rejected, there is enough evidence to infer that a variation of career expectation influences the student enrollment. The test statistic of the third test is 9.80; the null hypothesis can be rejected as well, therefore, there is enough evidence to infer that with the same amount of tuition fee and same level of career expectation, a variation in government aid influences the student enrollment. O E O-E (O-E)²/E

Hypothesis 1 : Government aid ( vignette 1 + 2)
Yes 99 82 17 3.524390244
No 81 98 -17 2.948979592 χ² 6.473369836
Hypothesis 2: Career expectation ( vignette 3+4)
Yes 95 69 26 9.79710145
No 85 111 -26 6.09009009 χ² 15.88719154
Hypothesis 3: Government aid with all other variables (vignette 4+5)
Yes 86 69 17 4.188405797
No 94 111 -17 2.603603604 χ² 6.792009401
Figure 8 Chi-squared test of for study after graduation
The same test procedure was applied for the study after graduation choice. For hypothesis 1, the test statistic result is 6.47, as the rejection region is the same as the above test 5.991, therefore, the null hypothesis can be rejected, there is some evidence to infer that a variation in government aid influences the student enrollment under the same amount of tuition fee. The same procedure follows, the test statistic result of the second hypothesis is 15.89, therefore, the null hypothesis can be rejected, there is enough evidence to infer that a variation of career expectation influence the student enrollment. The test statistic of the third test is 6.79; the null hypothesis can be rejected as well, therefore, there is some evidence to infer that with the same amount of tuition fee and same level career expectation, a variation of the government aid influence the student enrollment.
Discussion
To sum up the findings it is important to look at each hypothesis separately and to give a possible explanation for the results. First we will start with the questions whether or not the students would start their first study. In the first hypothesis we only look at a situation with high tuition fees and a variation in government aid, the results of this variation is obvious when looking at chi-squared (49.7) there is thus overwhelming evidence to infer that government aid does influence the enrollment of students in this situation. Hypothesis two looks at a situation with tuition fee, government aid and a variation in career expectations. The chi-squared result of this test was even higher (51.2) resulting in a conclusion that career expectations also influence students choice to enroll. Remarkably when looking at the third hypothesis in which government aid is manipulated in a situation with high tuition fee and low career expectations the chi-squared result is much lower (9.8) although this is still enough to reject the null hypothesis we can try to explain this phenomenon. When understanding the conclusion of the second hypothesis, which gave overwhelming evidence that a variation in career expectations does influence student enrollment, it is easier to understand that the variation of government aid in the third hypothesis is of lower influence, caused by the presence of the variable; low career expectations. We might assume by this result that career expectations are of greater importance to whether or not a student chooses to enroll than government aid. The results of the second series of questions, whether or not students would opt for further study are significantly lower. A possible explanation why all chi-squared results are lower than in the first series of questions could be that the choice to continue to study is less uniform and is more of a personal choice, while the students asked are all doing a study and most likely value doing at least their first study. Important to mention is that the same phenomenon is occurring as in the first series of questions; the highest chi-squared is observed when manipulating career expectations and in the third hypothesis when manipulating government aid in a situation with high tuition fee and low career expectations chi-squared more that halves.
Limitation
As part of the questionnaires was executed online and there were 20 questions in the questionnaire, respondents might not have been very focused throughout the entire questionnaire. In addition to that, there are 5 vignettes in the questionnaire and each of vignettes with a short description, the respondents need to read it carefully, and they might have felt some confusion resulting in a misinterpretation of the questions. Another limitation caused by a lack of time and resources is that the research was only conducted in the province of North-Brabant which might influence the generalizability of the results.
Recommendation for Further Research
- The effect of racial and ethnical differences.
- The influence of gender.
- The influence of different living standards.
- The influence of a part-time job.