Great expectations: students’ educational attitudes upon the transition to post-secondary vocational education

In this study, we examine students’ educational attitudes upon the transition to Dutch senior vocational education (SVE), a transition associated with high dropout rates in the first year. Prior studies have identified differences in educational attitudes between sociodemographic groups. However, the mechanisms underlying those differences remain topic of debate: some studies point at differences in the school orientation and support in students’ social communities outside school, others focus on differences in educational experiences between sociodemographic groups. Multilevel sequential regression analyses on a diverse sample of 1438 students in urban SVE schools reveal that students have very positive educational attitudes upon their transition to SVE. Ethnic minority students express particularly positive attitudes. School-related encouragement and support at home plays an important role in students’ attitudes, but the attitudes of students from lower educated or ethnic minority communities are less related to this support. Prior school experiences play an essential, but occasionally counterproductive, role in students’ attitudes upon transition, depicting the transition as a fresh new start for some, and an unwelcome threshold for others.

Education Inspectorate 2002). While a transition can offer a welcome fresh start for some, transitions are difficult for all youth, as students need to integrate into a new social and academic environment (Alexander et al. 2001;Tinto 1993;Langenkamp 2010). Students from disadvantaged backgrounds in particular may encounter difficulties when making a transition (Roderick 1993). This picture holds especially true for the transition to senior vocational education (SVE) in the Netherlands. SVE is the post-secondary continuation of the vocational track in Dutch secondary education, which serves about 55 % of the Dutch secondary school population (Dutch Ministry of Education Culture and Science 2010). More than half of all dropout in SVE occurs in the first year after the transition (Dutch Ministry of Education Culture and Science 2011). Like in many Western countries, dropout rates in SVE are socially and ethnically patterned: students from lower socio-economic backgrounds drop out more often, as do students with less educated parents, and ethnic minority students (OECD 2006;Dutch Ministry of Education Culture and Science 2009;Alexander et al. 2001). Students from those backgrounds have been found to drop out in the first semester after the transition to SVE in particular (Elffers 2011).
The attitudes with which students enter a new institution have been found to play a decisive role in their attainment after the transition (Bers and Smith 1991;Hausmann et al. 2007;Tinto 1993). Therefore, it is relevant to examine whether differences in educational attainment between socio-demographic groups are mirrored in the educational attitudes among those groups upon the transition to SVE. Prior studies have indeed identified differences in the educational attitudes between socio-demographic groups, most prominently between ethnic groups. However, the nature of those differences remains topic of debate. Fordham and Ogbu (1986) founded their renowned 'acting White' thesis on reports of oppositional attitudes towards school among Afro-American students in the US, which could offer an explanation for their lower achievement in education. However, a growing number of studies indicates that the educational attitudes among ethnic minority students resemble the attitudes of majority students or are even more positive. These findings point towards a gap between the educational attitudes and achievement among ethnic minority students (Alexander et al. 1994;Khattab 2003;Downey et al. 2009;Kao and Tienda 1998;Mickelson 1990; Van der Veen and Peetsma 2006).
Differences in how educational attitudes are defined, and at what stage in students' academic careers they are measured, may account for some of the contradictions in the findings of the abovementioned studies (Downey et al. 2009;Mickelson 1990;Tyson 2002). In addition, a range of explanations is offered in the literature to account for the differences in educational attitudes between socio-demographic groups. Whereas some studies attribute those differences primarily to socio-cultural differences, and particularly to the educational attitudes prevailing in students' social communities (Davies and Kandel 1981;Fordham and Ogbu 1986;Wigfield et al. 1998;McMillan and Reed 1994;McCarron and Inkelas 2006;Ryan 2000;Steinberg et al. 1992;Wentzel 1998), other studies suggest that they are mainly related to differences in students' educational experiences, and in the extent to which students evaluate their attitudes in the light of those experiences (Alexander et al. 1994;Eccleston et al. 2010;Eccles 1983;Hossler and Stage 1992;Tinto 1993;Tyson 2002).
In this study, we look into the educational attitudes of students upon the transition to post-secondary vocational education in the Netherlands. Dutch schools for SVE, especially those in urbanized areas, serve a diverse population of students aged 16+, with a substantial proportion of students from lower socio-economic and ethnic minority backgrounds. As students transferring to SVE have spent at least 12 years in education, their educational attitudes carry the traces of long-range experiences in education. Thus, the context of the transition to SVE offers excellent opportunity to study differences in the educational attitudes of students from various sociodemographic backgrounds, and the potential role of the social and academic context in those attitudes. By looking at the relevant yet understudied setting of the transition to post-secondary vocational education, we aim to expand the existing body of research on educational attitudes and school transitions, which is primarily based on research in Anglo-Saxon secondary and higher education. The research questions guiding our study are: • Do the educational attitudes of students upon the transition to Dutch SVE differ according to socio-demographic background characteristics? • What is the role of the school orientation and support in students' home environment in their educational attitudes upon the transition to SVE? • What is the role of students' prior academic pathway and experiences in their attitudes upon the transition to SVE? • Do the abovementioned factors interact in their relationship to students' educational attitudes?
2 Theoretical background

Defining educational attitudes
In previous works, the term 'educational attitudes' has been used to describe various dimensions of students' pro-or anti-school sentiments, their personal educational aspirations, and their expectations of their educational career. The term is often used to refer to students' general orientation towards school and the importance that students attach to education for their personal lives or careers (Downey et al. 2009;Mickelson 1990). Such attitudes can be defined in more and less contextualized ways: students have both context-specific attitudes towards their particular school and degree program (Anderman and Kaplan 2008), as well as a more universal orientation towards education or their role as a student (Anderman and Freeman 2004). Educational aspirations refer more specifically to students' personal plans or goals in education, which appears to be one of the most powerful predictors of academic achievement (Downey et al. 2009). Such aspirations can take the form of more abstract, or idealistic, goals about the educational level that students plan or hope to attain in their lives, or of more concrete, or short-term, goals, for instance regarding the completion of a class or degree program (Mickelson 1990;Buchmann and Dalton 2002;Deil-Amen and Lopez Turley 2007;Teachman and Paasch 1998). Last, when students transfer to a new school, they have certain expectations of their new school environment and how they will fit into this environment, both in social as well as academic terms (Tinto 1993). The above mentioned dimensions of students' educational attitudes may overlap and affect each other, and are often studied together (Deil-Amen and Lopez Turley 2007). The term 'educational attitudes' can serve as an umbrella for those various dimensions, and we will use this umbrella term accordingly. In our study, we distinguish four dimensions of students' educational attitudes upon the transition to SVE, capturing both general and contextualized dimensions of educational attitudes. We examine students' expectations of their new school environment (expected institutional fit and expected academic fit), their aspirations with respect to the program they have enrolled in (educational aspirations) and their attitudes towards education in general (general educational attitudes).

Socio-demographic differences in students' educational attitudes
Educational attitudes predict educational attainment: if students report high aspirations, optimistic expectations, and positive feelings towards their education, they are more likely to succeed in school (Ekstrom et al. 1986;Schoon 2008;McMillan and Reed 1994;Rumberger 1995;Sewell and Shah 1968;Maehr and Meyer 1997;Morgan 2005;Tinto 1993). School dropouts generally report less positive school attitudes (Ekstrom et al. 1986). As students from lower socio-economic and ethnic minority backgrounds drop out more often, we may expect that those students report more negative attitudes towards school. While some studies indeed found more negative educational attitudes among those groups of students (Fordham and Ogbu 1986), other studies revealed an inverse pattern: students from lower socioeconomic and immigrant backgrounds generally report more positive attitudes towards their education (Alexander et al. 1994;Downey et al. 2009;Van der Veen and Peetsma 2006). Irrespective of the exact nature of the differences found, scholars point to both social and academic context factors to explain differences in the educational attitudes across various socio-demographic groups, as we will set out below.

The social context
The educational attitudes of close relatives or significant others in students' lives affect students' own attitudes (Khattab 2003;Davies and Kandel 1981;Sewell and Shah 1968). In particular, the influence of parents has been studied extensively in research on educational attitudes. Not only do parents act as "expectancy and value socializers" for their children (Eccles 1983), their important role as sources of social and practical support and guidance for the educational careers of their children is emphasized in various studies (Alexander et al. 2001;Ekstrom et al. 1986;Hossler and Stage 1992;Linnenbrink-Garcia and Fredricks 2008;McCarron and Inkelas 2006;McMillan and Reed 1994;Rumberger 1995;Schoon 2008;Trusty 1998). Parental influence on students' attitudes seems to be stronger than peer influence (Davies and Kandel 1981). Yet, peer attitudes appear to shape students' educational attitudes considerably as well (Chang and Le 2005;Ryan 2000;Steinberg et al. 1992), especially among adolescent students (Davies and Kandel 1981;Cotterell 2007;Wentzel 1998;Wigfield et al. 1998). Differences in the school orientation and support for students' school careers in their community may account for differences in their attitudes. While most parents have high expectations of their children's academic careers regardless of the family's socio-economic status (Khattab 2005), socio-economically disadvantaged communities, especially recent immigrant communities, may attach additional value to education as a means to pursue upward social mobility (Rothon et al. 2010). However, there seems to be more incongruity between the emotional and practical support provided by parents in disadvantaged communities (Alexander et al. 1994). Families with higher socio-economic status have better knowledge of the educational system, and better access to human and financial resources, to not only support their children's educational careers emotionally, but in more practical ways as well (Khattab 2005). We can conceptualize the emotional and practical resources in students' social networks supporting their academic career as a school-related dimension of their social capital (Stevens et al. 2007).

The academic context
The school environment has been found to affect educational attitudes substantially (Eccles 1983;Khattab 2005;Goodenow and Grady 1993;Kao and Tienda 1998;Maehr and Meyer 1997). Not only do teachers and classmates affect students' educational attitudes through expressing, directly or indirectly, their personal academic values and expectations, but students' experiences in education also shape their attitudes towards their current and future educational pathways (Eccles 1983;Hossler and Stage 1992;Nichols 2008;Tinto 1993;Goodenow and Grady 1993). Students have been encouraged or discouraged by prior interactions with teachers and classmates before entering a new school (Finn 1993), and some students make the transition in the course of gradual disengagement from school (Finn 1993;Rumberger 1995). Earlier experiences of failure in school can be an important reason for students to enter a new school with less positive attitudes (Tyson 2002). Students from lower socio-economic or ethnic minority backgrounds generally encounter more academic or behavioral difficulties in their school careers (Alexander et al. 1994;Finn 1993;Finn and Rock 1997), and have been found to report lower levels of emotional engagement in school (Marks 2000;Voelkl 1995). Such experiences could press down students' optimism about the next step in their educational careers. Thus, students' academic pathways can play an important role in shaping their educational attitudes upon the transition to a new school.

The case of the transition to Dutch post-secondary vocational education
The vocational track in Dutch post-secondary education offers degree programs with an explicit vocational orientation and strong ties to the labor market. SVE offers programs in engineering, health and social care, economics, and agriculture at four levels: assistant level (1), basic vocational level (2), full professional level (3), and specialist level (4), with program durations from one year (level 1) to four years (level 4).
The Dutch post-secondary school system allows for moving up to subsequent degrees: graduation at one level in SVE enables students to continue to the next level, a level 4 SVE diploma gives access to higher professional education, which in turn can lead to access to university. The educational attitudes of students in school systems like the Dutch system, which sorts students into different educational trajectories at an early age, have been found to be more determined by the type of school they attend than in more comprehensive school systems (Buchmann and Dalton 2002). Students may experience a stigma of attending the lower track, perceiving that they are less worthy and need to lower their future aspirations (Arum and Shavit 1995;Khattab 2005). However, while SVE programs are primarily intended to prepare students for a direct transfer to the labor market after graduation, many SVE students aspire to move on to higher levels of post-secondary education instead, and numerous students indeed do so. In contrast, other students may transfer to SVE with the desire to leave school as soon as the legal obligation to attend school ends at their 18th birthday. Yet, irrespective of their personal goals, the transition to SVE is a major step for all students. They leave their familiar, and usually smaller, school behind, and need to integrate into a new institutional environment, which may cause both anxiety and excitement.

Methods
In this study, we look at students' educational attitudes upon the transition to SVE. We examine whether there are differences in educational attitudes according to students' socio-demographic background that may or may not mirror socio-demographic differences in academic achievement, and we assess the role of social and academic context factors in students' educational attitudes. Below, we provide an outline of the methods used to address these questions.

Sample
We conducted our study with a sample of 1438 first year students in SVE in the Netherlands. From the total number of 40 SVE school boards in the Netherlands, 10 boards were invited for participation in the study, based on (1) their urban setting, to assure socio-demographic diversity of the sample, and (2) the range of degree programs offered, to enable data collection in various degree programs and levels within one school board. The five school boards that agreed to participate represent the initial selection satisfactorily. Within each school board, first year classes were selected from the three main vocational sectors (engineering, economics, and health and social care) at all four SVE degree levels. Classes were selected during the summer break, before student enrollments were finalized, and before individual students were assigned to individual classes. A total of 61 classes participated in the study, with an average class size of 24 students. Participation in the study was voluntary for all students. As questionnaires were completed in a classroom setting under supervision of a teacher, in practice all students in a class completed a questionnaire, and we received no reports of students not willing to participate. Hence, it is unlikely that specific groups of students are over-or underrepresented in our study. This is confirmed by the descriptive statistics in

Questionnaire
A new self-report questionnaire was developed to suit the specific level and nature of SVE and its students. The questionnaire has been subjected to a pilot study before being finalized. Its psychometric properties were been assessed through reliability analyses and principle component analyses. Cronbach's alphas were satisfactory to good. All variables had a 5-point Likert-type scale (strongly disagree to strongly agree), except for the items on students' socio-demographic background and characteristics of their new school and program (school board, sector, level).

Outcome variables
We defined four outcome variables to measure students' attitudes upon transition. Two measures assessed students' expectations of their fit with their new educational environment: expected institutional fit (e.g. I think I will feel at home at this school, 3 items, α = .87), referring to students' expected fit with the institution in general, and expected academic fit (e.g. I think this degree program suits me well, 7 items, α = .80), referring to students' expected fit with the specific vocational program and sector. The variable educational aspirations (e.g. I really wish to graduate from this degree program, 3 items, α = .74) reflects students' ambitions with respect to the degree program they have enrolled in. General educational attitudes (e.g. I think it is important to go to school, 7 items, α = .80) refers to students' general valuing of education as a means to pursue their career goals.

Explanatory variables
In line with the theoretical framework and the related research questions, we distinguished three groups of explanatory variables, that were entered in sequential steps to the statistical models: (1) socio-demographic background characteristics, (2) social context variables, and (3) academic context variables. As indicators of students' sociodemographic background, we used measures of students' ethnic and socio-economic background. The ethnic background measure was based on self-report by the student of his/her ethnic identity. As indicators of students' socio-economic background, we included student statements about their parents' job status, highest educational level of the parents, and financial circumstances at home. Moreover, we took into account students' birth year and gender.
The social context variables comprised reports on students' school-related social capital: the amount of encouragement and support regarding their school careers that The academic context variables included the educational program that the student was enrolled in prior to entering SVE, with additional information on whether the student graduated in that program or not, the vocational sector that the student is about to enter (engineering, economics, health and social care), the school board (3 school boards in highly urbanized areas, and 2 school boards in intermediate urban areas), and the level of the degree program (level 1 to 4). Moreover, students' prior school experiences were measured using four scales: prior emotional engagement (e.g. I felt at home at my previous school, 3 items, α = .88), prior academic engagement (e.g. I worked hard at my previous school, 6 items, α = .77), prior contact with classmates (e.g. I had many friends at my previous school, 5 items, α = .81), and prior contact with teachers (e.g. I had good relations with my teachers at my previous school, 6 items, α = .87).

Data collection
Data were collected in the first week of the academic year 2008/2009, at the moment students entered SVE. This scheduling prevented students' reports of their educational attitudes upon transition from being colored by actual experiences in their new school environment. Students filled in a questionnaire in the classroom under supervision of their teacher. All teachers were informed about the purpose of the study, and received an instructional handout. Students could hand in their questionnaire in a blank envelope. A short introductory statement about the purpose of the study was printed on the questionnaire, as well as a short privacy statement in which we guaranteed that no identifiable personal data would be reported.

Statistical analysis
Data were screened for meeting the assumptions of regression analyses. As missing values were limited and missing at random, we applied the expectation maximization algorithm to impute missing values (SPSS Inc. 2010). The intra-class correlation coefficient (ICC) indicated that a substantial part of the variance could be attributed to between-class differences, with ICC's ranging from .04 to .10. Therefore, multilevel regression analyses were carried out, with class and student included as two separate levels (Snijders and Bosker 1999). To facilitate the interpretation of the coefficients, we standardized variables to zero mean and unity variance. We conducted sequential regression analyses. First, we fitted a model which included students' sociodemographic characteristics only (Model 1). Then, we added the social context variables (Model 2). Third, we added the academic context variables (Model 3). In addition, we checked for first order interactions of all socio-demographic student characteristics with the social and academic context variables, by adding each interaction as a fixed variable to the third model one by one. Table 2 shows the results of the multilevel sequential regression analyses.

Results
Our data indicate that, on average, students have positive educational attitudes upon the transition to SVE, and especially high aspirations (M = 4.47 on a scale from 1 to 5, with 5 signifying the most positive score, SD = .51). A model that only includes socio-demographic characteristics explains small proportions of variance for students' expectations (R 2 = .02 for both expected institutional fit and expected academic fit) and aspirations (R 2 = .04), and slightly more for their general attitudes (R 2 = .11). Including the social context variables increases the amount of explained variance of all four outcome variables substantially (R 2 = .11 to .22), and adding academic context measures increases the proportion of explained variance even more (R 2 = .15 to .29).

Socio-demographic patterns in students' educational attitudes
Ethnic minority students report significantly more positive aspirations and general attitudes than Dutch majority students. There is no significant difference for students with an Antillean background, but this exception may be caused by their smaller representation in our sample. Students with an ethnic background other than the largest five ethnic groups (native Dutch, Moroccan, Turkish, Surinamese, Antillean) have equal or more positive aspirations and general attitudes, but expect less academic fit. The ethnic background coefficients remain significant after controlling for social and academic context variables in our models. We do not find a significant role of the level of parental education, except for a positive association between having parents who completed secondary education at most and students' expected institutional fit. Students with unemployed parents have more positive general attitudes towards education than students with working parents. Students who don't know if their parents work, which is likely an indication that they are not closely involved with their parents at the moment of measurement, expect a better academic fit. Male students report more negative educational attitudes, but this difference decreases when taking social and academic context variables into account. Financial problems in the family have no significant influence on students' educational attitudes in the models that include social and academic context variables. Only when we control for students' school-related social capital, we find an effect of year of birth, indicating that younger students have less positive educational attitudes upon transition.

The role of the social context
The encouragement and support of parents, peers and the community plays an important role in students' educational attitudes. Especially parental encouragement and support is important for all four dimensions of students' attitudes upon transition to SVE. Peer support particularly influences students' general attitudes, while peer support is less related to students' aspirations. School-related support from parents, peers and the broader community all remain substantial and significant in models with academic context variables added.  .14 (.15) .25 (.14) .10 (.14) Other ethnic background .08 (.10) .10 (.10)  Academic context variables have significant influence on students' educational attitudes. In particular, students' prior academic engagement, but also their contact with classmates at their former school, associates with their educational attitudes. The perceived connection with teachers in the prior school is of less importance. Prior emotional engagement seems to affect students' educational attitudes negatively. The zero-order correlation between emotional engagement and attitudes is positive, but in a model that also includes other prior school experiences, the effect of prior emotional engagement on educational attitudes turns negative. Emotional engagement is a measure of students' overall perceived fit with the prior school, which may comprise various social and academic aspects of students' prior school experiences, such as the fit with classmates, teachers, or the curriculum. If we control for those specific aspects, a negative correlation between students' prior emotional engagement and their educational attitudes remains. Students attending schools in highly urbanized areas expect less fit with the institution than students transferring to schools in less highly urbanized areas. Students enrolling in an economics program are less certain of their academic fit with the program, while engineering students value education in general less. Some differences between students in different degree levels are found for expected institutional fit and general attitudes. Students who enter SVE after dropping out of pre-vocational education are more motivated for their new academic program in SVE, while students who make a second try after earlier enrollment in SVE are more confident about their fit with the institution and program. Students transferring to SVE from other academic routes expect a better academic fit and value education in general more.

Interaction effects
Assessment of first order interactions between socio-demographic characteristics and social and academic context variables revealed two interesting patterns. Above, we reported positive associations with educational attitudes of both an ethnic minority background and the amount of school-related support from parents. However, an earlier study of our data (Elffers 2012) showed that ethnic minority students report lower levels of school-related support from their parents. Inclusion of interaction effects in our third models showed that the level of parental support plays a smaller role in the attitudes of ethnic minority students than of students with a native Dutch background. An example of this interaction effect is given in Fig. 1 for students with a Surinamese background. In the group of students with lower levels of parental school-support, Surinamese students report higher educational aspirations than Dutch majority students, and their aspirations are less affected by an increase of parental support than the aspirations of Dutch majority students. We found similar interaction effects of a Surinamese background with parental support on expected academic fit, and of Moroccan and Surinamese backgrounds with community support on expected institutional and academic fit. Moreover, similar patterns were found for the interaction of the educational level of students' parents with the school-related support from parents or the  Fig. 1 Interaction effect of Surinamese background and school support parents on students' educational aspirations upon transition to SVE. Note plot based on 4 theoretical students who score zero on all other predictors community, and of having financial problems or unemployed parents with parental support. For instance, the aspirations of students from poor families do not seem to benefit from increased parental support as much as the attitudes of students from more affluent families, and the same is true for the expected academic fit of students with unemployed parents or with parents who completed secondary education at most. The educational attitudes of those students appear to be less affected by lower levels of support. An interaction effect in the opposite direction can be found among students with higher educated parents: they expect a better academic fit if they report better access to supportive resources in their community. Figure 2 shows an example of the second pattern of interaction effects: an interaction effect of students' prior education with their personal experiences in that prior education on their educational attitudes. Whereas we find an overall positive effect of students' prior educational experiences on their educational attitudes upon transition, interactions between students' prior education and their experiences in that prior education indicate a negative relationship in some situations. Students who were earlier enrolled in an SVE program, without graduating from that program, have higher aspirations if they were less academically engaged in the prior program, and have lower aspirations if they were previously more engaged. A similar interaction was found for this particular group with the level of prior emotional engagement and with contact with teachers at the prior school, on their expected institutional fit, aspirations, and general attitudes.

Discussion
Positive educational attitudes upon the transition to post-secondary education are critical for school success after the transition. High dropout rates among particular socio-demographic groups might seem to suggest that students from those groups may enter their new school with more negative attitudes. However, this study adds to the growing number of studies that indicate that students from lower socio-economic backgrounds or from ethnic minority groups do not report more negative educational attitudes. On the contrary, while ethnic minority students are overrepresented in SVE prior education: SVE without graduation Fig. 2 Interaction effect of 'prior education: SVE without graduation' and prior academic engagement on students' educational aspirations upon transition to SVE. Note plot based on 4 theoretical students who score zero on all other predictors. PVE pre-vocational education, SVE senior vocational education dropout statistics in the Netherlands, those students report substantially higher aspirations and more positive general attitudes towards education. The same pattern is found for students from lower educated or poorer families. Thus, our results attest to the high ambitions among students hailing from those particular groups, which likely signifies a desire to pursue upward social mobility. Like students with lower educated parents, ethnic minority students, especially those growing up in first and second generation immigrant families, are often one of the first in their family to make the transition to post-secondary education, which makes them highly ambitious (Suárez-Orozco et al. 2009). Parents in those families are less experienced with the educational system, and may find it difficult to support their children in their school careers for that reason (Stevens et al. 2009). An earlier study based on our data confirmed that ethnic minority students, students with lower educated parents, and students from poor families, characteristics that are often clustered within the same families, report significantly lower levels of support for their school careers in their social communities (Elffers 2012). Inevitably, these youth have learned to navigate the educational system by themselves, causing a certain educational independence that can work out both beneficial and detrimental to school success. As students rely less on the support of relatives or significant others for their school careers, they are more autonomous and self-supportive. But this also means that they will encounter less monitoring, guidance and correction from the people who know them best. This mechanism is illustrated by our finding that, compared to other groups, the educational attitudes of students from lower socio-economic and ethnic minority communities are less related to the amount of school-related support in their environment. While in general students' educational attitudes are strongly related to the amount of school-related social capital in their environment, this is not the case for students from lower socio-economic or ethnic minority groups. The attitudes of students in those groups remain very positive, regardless of sufficient or limited access to school-related support in their social networks.
Our findings imply that the potential negative effect of limited access to supportive resources on student achievement is not mediated by students' personal educational attitudes. This implication is consistent with the notion of the attitude-achievement gap, which describes the discrepancy between positive educational attitudes and negative achievement outcomes among disadvantaged students (Alexander et al. 1994;Downey et al. 2009). Furthermore, while lower levels of parental support seem to exert little negative influence over the educational expectations and aspirations of students from lower socio-economic or ethnic minority communities, those lower levels of parental support may limit students' prospects to realize those. Lower educated parents may think they are not capable of supporting their children, simply because they haven't completed many years of schooling themselves. However, our results imply that it is not the actual educational level of the parents that counts, but the schoolrelated support that they provide to their children. This result confirms Eccles' view that it is not the power of parents as role models that matters, but their role as direct socializers of achievement beliefs and attitudes (Eccles 1983).
School transitions are critical turning points in students' school careers, and will often be met with both excitement and anxiety. A transition to a new school is no isolated event, but rather an important threshold embedded in students' educational life-course. Whether students perceive a transition as a welcome new start or a disturbing interruption, or perhaps both, depends on their personal educational history. Our findings indicate that students with negative experiences in a previous SVE program are more eager to make a new start. Conversely, students with positive experiences in a previous SVE program, who apparently did not succeed in the program and therefore are required to make a new start, are more cynical about the prospects of their second try. These students may be especially at risk, as they apparently feel discouraged by the failure of the first try. These findings stress the importance for educators to learn about the educational history of their students. Intake interviews with transferring students can be used to find out about their prior school experiences, successes and failures, and to check for indicators of progressive disengagement from school.

Conclusion
Research on the educational attitudes of students in vocational tracks is scarce. This study contributes to earlier work on students' educational attitudes by expanding the focus to students transferring to post-secondary vocational education in Dutch urban areas. However, because of this specific context, we need to be careful to generalize our findings to other educational settings. Other limitations of the current study pertain to the measures used, and the cross-sectional design. First, our data result from students' self-reports. While the information on students' graduation before their transition informs us about their success or failure in their prior school program, no objective measures of prior academic achievement were available to include in our analyses. Prior grades or standardized test scores could be powerful predictors of students' educational attitudes upon transition. Future research could benefit from inclusion of such more objective measures. Second, we used retrospective measures of students' experiences in their previous education that were assessed at the same time as their expectations, aspirations, and attitudes. Such design calls for awareness not to make causal inferences based on our data.
Earlier research on the educational attitudes of disadvantaged students has mainly concentrated on students from socio-economically marginalized communities. Our study of students in the vocational track adds to this perspective by focussing on the educational attitudes of students who may experience academic marginalization, as they do not transfer to higher education after high school. It is often suggested that placement in a vocational track could result in a 'cooling out' of academic aspirations and expectations (Alexander et al. 2008;Arum and Shavit 1995;Khattab 2005). However, our study demonstrates that students in the Dutch vocational track do not show signs of cooled attitudes towards education. They start out in SVE with great expectations. An important next step in educational research is to examine the factors that help or hinder those students in turning their great expectations into great achievements.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
Frans J. Oort is professor of Methods and Statistics of Educational Research and director of the Graduate School of Child Development and Education, University of Amsterdam. His research interests include measurement and measurement bias, and statistical modeling, especially non-standard applications of structural equation modeling.