A longitudinal study on time perspectives: Relations with academic delay of gratification and learning environment

Authors


  • This research was made possible by a grant from the Netherlands Organization for Scientific Research (NWO-PROO).

Thea Peetsma, Research Institute of Child Development and Education, University of Amsterdam, Nieuwe Prinsengracht 130, 1018VZ, Netherlands. (E-mail: t.t.d.peetsma@uva.nl)

Abstract

After they start secondary school (at age 12 in the Netherlands), students' time perspectives on school and professional career and self-regulated learning decrease, while their perspectives on leisure increase. We aimed to investigate relations in the developments in time perspectives and delay of gratification in the first years of secondary education. In addition, the dependence of these relations on students' perceptions of the extent to which teachers emphasize the relevance of what is being learned was investigated. Seven hundred and one students participated in the study. A self-report questionnaire was administered four times. The results showed the expected decreases in time perspectives on school and professional career and on delay of gratification and increases in perspectives on leisure. Furthermore, as expected, the development in short-term perspective on school and professional career was positively related, and the development in the long-term perspective on leisure was negatively related, to the development in delay of gratification. Positive relations were found between emphasized relevance of learning and developments in students' perspectives on school and professional career and delay of gratification.

The time dimension of goals has been considered a motivator or motivational determinant for a long time (Frank, 1939). Future Time Perspective (FTP) is generally described as a representation or conceptualization of a particular life domain in terms of time with a direction to the future. Lens (1986) defined this time perspective as a cognitive-motivational concept. It is characterized by “valence” and “extension” (Gjesme, 1996; Husman & Lens, 1999). The valence of the FTP indicates the value ascribed to a life domain in the future. The extension indicates the degree of remoteness of the representation in time. For students, “the time after finishing school” and “the current school year” seem to be meaningful terms in time. The motivational effect of a long-term goal seems not to be effective over very long periods (10–20 years) (Zhang, Karabenick, Maruno, & Lauermann, 2011).

Peetsma (1992) defined FTP as an attitude towards a certain life domain viewed over time, and conceptualized FTP in terms of three components: “cognition,”“affect,” and “behavioral intention.” Cognition consists of ideas or expectations with regard to the future, and of knowledge of social realities. Affect is interpreted as an expression of feeling or affection towards a particular life domain in the future. In the context of FTP, “behavior” is first seen as behavioral intention. Defined like this, FTP stands for a concept broader than instrumentality, which is mainly cognitive in nature. Including affect, FTP on school and professional career has proved to be a better predictor of school investment than the merely cognitive perceived instrumentality (Peetsma, 1992).

Time perspectives or orientations reflect a general predisposition. Perspectives are considered rather stable over time (Gjesme, 1996; Nurmi, 1989; Zimbardo & Boyd, 1999) and develop from the age of approximately 11 years on (Erikson, 1968; Piaget, 1955). The relatively late development of FTP makes it an interesting variable for research into motivation in secondary education.

Bembenutty and Karabenick (2004) embedded FTP within a self-regulated learning framework. In their view, FTP is a component of students' toolkits for learning to complete academic tasks over time. An important aspect of self-regulated learning is students' ability to delay gratification. Getting started and remaining engaged in learning tasks requires students to set aside distractions. Students' ability to delay gratification increases the likelihood of completing academic tasks (Bembenutty & Karabenick, 2004).

Research has repeatedly shown that school motivation and self-regulation, including academic delay of gratification, declines during the first years of secondary education (Midgley, Feldlaufer, & Eccles, 1989; Peetsma, 1997; Peetsma, Hascher, Van der Veen, & Roede, 2005). This decline in academic delay of gratification seems to be related to the development of FTP. FTPs on a school and professional career show positive relations with self-regulated learning behavior, such as academic delay of gratification (Bembenutty & Karabenick, 2004), while FTPs on leisure show negative relations (Peetsma, 2000; Peetsma & Van der Veen, 2011). In addition, shifts have been found in the relevance of life domains: Students' FTPs on school and professional career have become less important, while perspectives on other domains of life, specifically leisure time, have grown in importance or stayed stable from the start of secondary education. These shifts in the relevance of live domains can provide a possible explanation for the decline in motivation and self-regulation.

The decline in motivation and self-regulation has been a reason for schools to implement innovative learning environments to enhance motivation and learning (Waslander, 2004). Learning environments differ in the opportunities they give students to engage in self-regulated learning (Boekaerts & Niemivirta, 2000). It could be expected that in schools where learning is more related to students' experience and more relevant to them, the developments in FTPs and learning behavior would be more positive, but little is known about this yet. For teachers it seems important to help students understand the relevance of a certain learning task for themselves and to relate the learning content to students' prior knowledge (Assor, Kaplan, & Roth, 2002; Thoonen, Sleegers, Peetsma, & Oort, 2011). Students are supposed to be more inclined to engage in self-regulated learning, like academic delay of gratification, if they perceive what is being learned as relevant to their own lives.

Developments in perspectives on different life domains and the effects on students' learning behavior have not often been studied longitudinally. Longitudinal research is needed to fully understand the development of FTP and relations with the learning environment and self-regulated learning, particularly with regard to school. This paper reports on a longitudinal investigation of developments in students' FTPs on two life domains and on two terms in time. In addition, we examined the relations between developments in FTP, students' academic delay of gratification, and their perceptions of the extent to which their teachers emphasize the relevance of learning. Students (aged 12 years) were followed from the start of secondary education until halfway through the second year of secondary education. Based on the theory above, the following hypotheses were tested:

  • 1FTPs on school and professional career, and academic delay of gratification, decrease from the start of secondary education and perspectives on leisure increase.
  • 2Developments in FTPs on school and professional career, academic delay of gratification, and perceived emphasis on relevance of learning at school are positively related.
  • 3Developments in FTPs on leisure, academic delay of gratification, and perceived emphasis on relevance of learning at school are negatively related.

Possible differences in school levels and sex were also explored.

Method

Design and procedure

We implemented a longitudinal design with four measurements to investigate developments in students' FTP, academic delay of gratification, and their perception of the emphasis on relevance of learning tasks. The first measurement was in September–October 2009, 2–4 weeks after the start of the school year. The second was halfway through the first year in February–March 2010, the third was at the start of the second year in September–October 2010, and the last measurement was halfway through the second year in February–March 2011. The questionnaires were administered during regular class time and it took the students 30–40 min to fill in the questionnaire.

Participants

A total of 701 secondary school students participated in the study. In the education system of the Netherlands, children of approximately 12 years of age leave primary school and move on to different levels of secondary education. Students are selected for the different levels based on their primary education performance. The participating students were from all types of secondary education. In this study we differentiated three levels of secondary education. The lowest level included 195 students in prevocational secondary education (4-year study), the intermediate level included 198 students in lower general secondary education (4-year study), and the highest level included 308 students from higher general secondary education (5-year study), and pre-university education (6-year study). At the start of the present study, all the students were in year one of secondary education and their average age was 12 years. There were 362 (52%) boys and 339 girls (48%) in the study. The boys and girls were evenly distributed over the three school levels, χ2 (2) = 0.82, P = 0.66.

Measures

To investigate developments in students' FTPs, academic delay of gratification, and their teachers' emphasis on the relevance of learning, we used self-report measures, based on existing scales which have been tested for reliability and validity in earlier studies. All items were rated on a 5-point Likert-type scale from 1 (completely disagree) to 5 (completely agree).

Four scales from the Time Perspective Questionnaire (TPQ; Peetsma, 1992; Stouthard & Peetsma, 1999) were used to assess students' FTP, including scales for the “long-term and short-term FTP on school and professional career and for both perspectives on leisure time.” FTP is a complicated concept, existing of different components. A facet-design has been used to develop the questionnaire: Each of the four scales measures a combination of the three components, cognition, affect, and behavior intentions, towards particular manifestations of a life domain on a certain term of time in the future (Peetsma, 1992; α = 0.61–0.81). Examples of items for the four scales include: “I like thinking about going to college or what job I might have when I leave school,” for the long-term time perspective on school and professional career; “I am pleased that I will learn a lot of new things this year,” for the short-term time perspective on school and professional career; “I love dreaming about what I'll be able to do in my free time when I'm older,” for the long-term time perspective on leisure; and “I'm really enjoying my holidays and free time this year,” for the short-term time perspective on leisure.

We assessed the students' ability to “delay gratification” in learning situations. This variable refers to the students' capacity to postpone the satisfaction of impulses and pursue long-term academic goals. We adapted three items from the Academic Delay of Gratification Scale of Bembenutty and Karabenick (1998; α = 0.83). An example of an item of this scale is “I finish my school work, before I meet with my friends to have fun.”

To investigate students' “perceptions of the emphasis on relevance of what is being learned” from their teachers, three items from the Teacher as a Social Context questionnaire (TASC; Belmont, Skinner, Wellborn, & Connell, 1988) were combined with three items from a scale measuring “connection to students' worlds” adapted from Thoonen et al. (2011; α = 0.78). An example of an item of this scale is “The teachers talk about how I can use the things we learn in school.”

Analyses

The relations between the developments in FTPs, in students' academic delay of gratification, and in students' perceptions of the emphasis on relevance by their teachers were investigated using multivariate Latent Growth Curve Analysis (LGCA), using structural equation modeling in Mplus (Muthén & Muthén, 2004). An important advantage of LGCA is that each individual's growth pattern is represented by a unique curve. Individuals are allowed to have different initial values (at measurement 1) and to have different growth rates. In these analyses, the initial values (intercepts) and growth rates (slopes) are treated as latent variables and can be correlated with each other and with other variables. A negative correlation between the initial level and the slope means, for example, that a lower start measurement corresponds with a greater degree of growth. Because the time intervals between the four measurements were not the same, time as proportion of the year was included. As a consequence the model estimates the growth rates per year.

Not all the participants were present on all four occasions: 482 students (69%) participated in all four measurements, 134 students (19%) participated in three measurements, 64 students (9%) participated in two, and 21 students (3%) in one. Listwise deletion of students with missing data on one or more occasions has been criticized extensively (Little & Rubin, 1987), which is why in the present study all students were included in the analysis. The missing values were estimated using full-information maximum likelihood estimation (FIML). The FIML estimation is based on the assumption that missing values are missing at random (MAR), which assumes that missing values can be predicted from available data. Removing all cases with missing values (listwise deletion) is based on the more strict assumption that the missing values are completely at random (MCAR).

If possible, it is preferable to include all FTPs in one model, but for this a larger number of students is necessary. For that reason, we fitted four different models to the data. Each model included one of the four measures for FTP, students' academic delay of gratification, and students' perception of the emphasis on relevance. In addition, we analyzed whether sex and school level had an effect on the initial values and growth rates of each variable. A dummy variable for sex was added to each model (boys = 0 girls = 1). For school level two dummy variables were included, one for the lowest level and one for the highest level. The intermediate level served as the reference variable.

In constructing each model, we made path assumptions. Specifically, it was assumed that students' perceptions of emphasis on relevance would influence FTP and that FTP would influence academic delay of gratification. In addition, it was assumed that there might be a direct effect of emphasis on relevance on academic delay of gratification. Furthermore, we assumed that the initial values might influence the growth rates of that same variable or of other variables. Models were compared with each other using the χ2-test. A full model was initially developed in each case. Relations between variables were omitted if including them did not mean a significant (p < .05) improvement in the fit of the model.

To evaluate the fit of the models we used two indices in addition to the chi-square test, because the chi-square depends on sample size and is not very accurate (Browne & Cudeck, 1993). We used the comparative fit index (CFI) and the root mean square error of approximation (RMSEA). An RMSEA of approximately .08 or less indicates satisfactory fit in relation to the degrees of freedom, and an RMSEA value of approximately .05 or less indicates a close fit of the model. The CFI values should be higher than .90 and values close to 1 indicate a very good fit. By comparing how much variance in academic delay of gratification was explained by the four models (Browne & Cudeck, 1993), it was possible to determine which of the FTPs was most important in explaining differences in the development of academic delay of gratification.

Results

Before discussing the four estimated models, the mean initial values and growth rates for the four FTPs, academic delay of gratification, and emphasis on relevance are presented in Tables 1 and 2. The means for sex and the different school levels are also shown.

Table 1.  Means and standard deviations for initial value and growth rate for boys and girls
 Initial valueGrowth rate per year
BoysGirlsTotalBoysGirlsTotal
MSDMSDMSDMSDMSDMSD
  1. Note. aThe means and standard deviation for academic delay of gratification and emphasis on relevance presented here are the estimates from the model including long-term future time perspective (FTP) on school and professional career. The other models produced only slightly different estimates for these variables. bSignificant difference compared with boys, p < 0.05.

Long-term FTP on school and professional career3.950.344.06b0.364.010.36−0.160.14−0.150.14−0.160.14
Short-term FTP on school and professional career3.860.524.05b0.513.950.52−0.270.21−0.290.20−0.280.21
Long-term FTP on leisure time3.660.443.55b0.473.610.460.120.170.120.180.120.18
Short-term FTP on leisure time4.120.364.110.394.120.370.030.110.040.110.030.11
Academic delay of gratificationa3.330.633.400.603.360.61−0.300.23−0.310.24−0.300.24
Emphasis on relevancea3.220.393.210.453.210.42−0.200.22−0.200.23−0.200.22
Table 2.  Means and standard deviations for initial value and growth rate per school level
 Initial valueGrowth rate per year
Lowest levelIntermediate levelHighest levelLowest levelIntermediate levelHighest level
MSDMSDMSDMSDMSDMSD
  1. Note. aThe means and standard deviation for academic delay of gratification and emphasis on relevance presented here are the estimates from the model including long-term future time perspective (FTP) on school and professional career. The other models produced only slightly different estimates. bSignificant difference with lowest level (p < 0.05). cSignificant difference with the other two school levels (p < 0.05).

Long-term FTP on school and professional career3.930.373.930.364.10c0.32−0.19b0.14−0.070.13−0.190.13
Short-term FTP on school and professional career3.940.543.910.513.980.52−0.220.20−0.200.19−0.370.19
Long-term FTP on leisure time3.470.483.560.443.73c0.440.02c0.170.110.170.190.16
Short-term FTP on leisure time3.92c0.384.150.364.220.330.020.110.030.110.040.10
Academic delay of gratificationa3.12c0.613.380.633.500.55−0.22b0.24−0.390.23−0.29b0.22
Emphasis on relevancea3.200.403.140.433.27b0.43−0.090.19−0.080.18−0.34c0.18

As expected, long-term and short-term FTP on school and professional career, as well as academic delay of gratification, showed an average decrease over time. Also in line with our expectations, the overall long-term and short-term FTP on leisure time increased, although the growth in short-term FTP on leisure time was rather small. In addition, there was an average decline in students' perception of their teachers' emphasis on relevance of the learning tasks. Differences between boys and girls and the three school levels will be discussed separately for each model in the next sections.

FTP on school and professional career

The first estimated model including long-term FTP on school and professional career is presented in Figure 1. For easy reference, the figures present only the latent variables. The fit of the model was satisfactory, χ2(86) = 275.44, p < .001, CFI = .919, RMSEA = .056.

Figure 1.

Estimated model representing the relations between the initial values (i) and the slopes (s) of long-term future time perspective (FTP) on school and professional career (LFTP_SP), emphasis on relevance (EoR), and academic delay of gratification (ADoG), including standardized regression weights. Significance levels for all values are p < .05.

The results showed that the initial values of long-term FTP on school and professional career, academic delay of gratification and emphasis on relevance were positively related to each other. Emphasis on relevance had a positive effect on the initial values of FTP and academic delay of gratification. The initial value of long-term FTP on school and professional career also had a positive effect on academic delay of gratification. This means that at the start of their secondary education, the students who perceived more emphasis on relevance had a stronger long-term FTP on school and professional career and a higher level of delay gratification. Relations between the growth rates of these variables display a different pattern. The development in long-term FTP on school and professional career appeared not to be related to the growth of academic delay of gratification. The growth rate of students' perceptions of their teachers' emphasis on relevance was positively related to both long-term FTP on school and professional career and to academic delay of gratification. This indicates a positive effect of developments in perceptions of the learning environment on the development in both FTP and academic delay of gratification.

There was only one significant effect of sex in this model. Girls had a higher initial value of long-term FTP on school and professional career. We found several differences between school levels. The students in the highest level had a higher initial value of long-term FTP on school and professional career. The students in the lowest level had a lower growth rate of long-term FTP on school and professional career than the students in the intermediate level (no significant difference was found with the highest level) and a lower initial value for academic delay of gratification than the students in the other two levels. The growth rate of academic delay of gratification of the students in the intermediate level was lower than that of the students in the other two levels. In addition, the students in the highest level had a significantly higher initial value of emphasis on relevance than the students in the intermediate level. The students in the highest level also had a higher growth rate of emphasis on relevance. In this case the students in the highest level differed significantly from both the students in the lowest level and in the intermediate level. In this model 38% of the variance of the initial value of academic delay of gratification was explained by the other variables, as was 50% of the growth rate of academic delay of gratification.

The fit of the second model including short-term FTP on school and professional career (Figure 2) was also satisfactory χ2(87) = 307.13, p < .001, CFI = .917, RMSEA = .060. The relations between the initial values of the three latent variables in this model resembled the pattern in the model for the long-term perspective on school and professional career. However, in this model there was also a positive effect of the initial values of emphasis on relevance on the growth rate of short-term FTP on school and professional career. With respect to the growth rates, this model also differed from the first model. The growth rate of the short-term FTP on school and professional career was positively related to the students' academic delay of gratification. In addition, developments in emphasis on relevance had a positive effect on the growth rate of the short-term FTP on school, but the growth in emphasis on relevance had no direct effect on developments in student's academic delay of gratification. This indicates a mediating effect of short-term FTP on school and professional career between students' perceptions of the learning environment and the students' academic delay of gratification.

Figure 2.

Estimated model representing the relations between the initial values (i) and the slopes (s) of short-term future time perspective (FTP) on school and professional career (SFTP_SP), emphasis on relevance (EoR), and academic delay of gratification (ADoG), including standardized regression weights. Significance levels for all values are p < .05.

In this model we also found that boys and girls differed at the start of the first year in secondary education. Girls had a higher initial value of short-term FTP on school and professional career. There were no effects of school level on the initial value and growth rate of short-term FTP on school and professional career. The effects of the school level on academic delay of gratification and emphasis on relevance were similar to those in the first model; 42% of the initial value of academic delay of gratification and 67% of the growth rate of academic delay of gratification was explained by the variables in this model.

FTP on leisure time

The model including long-term FTP on leisure time is presented in Figure 3. The model fitted the data well, χ2(86) = 236.80, p < .001, CFI = .935, RMSEA = .050. As expected, long-term FTP on leisure time was negatively related to academic delay of gratification. The initial value of long-term FTP on leisure time had a negative effect on the initial value of academic delay of gratification, and the growth rate of long-term FTP on leisure time was also negatively related to the development in academic delay of gratification. Students' perceptions of the emphasis on relevance had a negative effect on long-term FTP on leisure time. The initial value and the growth rate of emphasis on relevance were negatively related to the initial value and growth rate, respectively, of long-term FTP on leisure time.

Figure 3.

Estimated model representing the relations between the initial values (i) and the slopes (s) of long-term future time perspective (FTP) on leisure time (LFTP_L), emphasis on relevance (EoR), and academic delay of gratification (ADoG), including standardized regression weights. Significance levels for all values are p < .05.

Again, we found that boys and girls differed on the first measurement. Girls had a lower initial value of long-term FTP on leisure time than boys. With respect to academic delay of gratification and emphasis on relevance, we found the same differences between the three school levels as in the first two models. In addition, we found that students in the highest level had a higher initial value of long-term FTP on leisure time than students in the other two levels. The lowest level students had a lower growth rate of long-term FTP on leisure time than the other students. The variables in this model explained 33% of the initial value of academic delay of gratification and 56% of the growth rate of academic delay of gratification.

The last model, including short-term FTP on leisure (Figure 4) also fitted the data well, χ2(92) = 208.88, p < .001, CFI = .941, RMSEA = .043. We found no significant effects of emphasis on relevance on short-term FTP on leisure time and no effects of short-term FTP on leisure time on academic delay of gratification. There was no effect of sex on short-term FTP on leisure time. There was, however, an effect of the school level on short-term FTP on leisure time. The students in the lowest level had a lower initial value on short-term FTP on leisure time than the students in the other two levels. This model explained 28% of the initial level of academic delay of gratification and 48% of the growth rate of academic delay of gratification.

Figure 4.

Estimated model representing the relations between the initial values (i) and the slopes (s) of short-term future time perspective (FTP) on leisure time (SFTP_L), emphasis on relevance (EoR), and academic delay of gratification (ADoG), including standardized regression weights. Significance levels for all values are p < .05.

Discussion

This study focused on developments in students' FTPs on two life domains, their academic delay of gratification and perceptions of their learning environment, as well as on the interrelations between these variables. The development in the FTPs and academic delay of gratification were as expected: positive developments in both perspectives on leisure time, and negative developments in the perspectives on school and professional career, and academic delay of gratification. The relations between these developments and students' perceptions of the extent to which their teachers emphasized the relevance of learning tasks also proved to be as expected; however, these expectations were based on scant research data.

Focusing on the results of the growth models, it appeared that the development in short-term FTP on school and professional career was an important mediator between the effects of perceived emphasis on relevance of learning on academic delay of gratification. Developments in emphasis on relevance had a strong positive effect on short-term FTP on school and professional career (standardized regression weight = 0.89) and the growth in this perspective had a strong positive effect on the growth in academic delay of gratification (standardized regression weight = 0.83). In contrast with the three other models, there was no direct effect of the growth in emphasis on relevance on academic delay of gratification in this model, indicating that this effect was fully mediated by short-term FTP on school and professional career. Furthermore, this model explained the most variance of the growth in academic delay of gratification (percentage explained variance = 67). Of the FTPs studied, short-term FTP on school and professional career was, therefore, the most important FTP for the development in academic delay of gratification. In the model for the long-term perspective on leisure time, the initial value and the growth rate of emphasis on relevance were negatively related to the FTP. This may indicate that the negative effect of the long-term perspective on leisure time on academic delay of gratification is diminished by students' perception of the teachers' emphasis on the relevance of their learning. This type of teaching seemed to work out positively for students' learning behavior. Furthermore, a substantial direct positive effect of emphasis on relevance of learning on academic delay of gratification was found. The model with long-term perspective on leisure time still explained more variance of the growth in academic delay of gratification (percentage explained variance = 56) than the models with long-term FTP on school and professional career (percentage explained variance = 50) and with the short-term perspective on leisure (percentage explained variance = 48).

With respect to sex differences, we found that girls were more concerned with their school and professional career than boys, whereas boys found leisure time more important than girls at the start of their secondary education. These differences did not change in the first 1.5 years of secondary education as we looked at the growth rates. Sex differences were not found in either academic delay of gratification, or in students' perceptions of emphasis on relevance of their learning.

The development of FTPs also differed somewhat between students in different school levels. The decline in the FTP on school and professional career was stronger for the lowest level students (pre-vocational education) than students at the intermediate level (lower general secondary education) and in the highest level (general secondary education or pre-university education). The students in the highest level seemed to value leisure time highly at the start of secondary education. The lowest level students seemed less focused on leisure time than the other students: they valued leisure time less in the short-term and their long-term perspective on leisure time increased less. We also found differences between school levels in academic delay of gratification and in perceptions of the learning environment. In the lowest level, students' academic delay of gratification at the start of secondary education was lower than that of the students in the other two levels. In addition, the ability to delay gratification declined more in the intermediate level than in the other levels, and students in the highest level showed the smallest decline in academic delay of gratification. The students in the highest level perceived more emphasis on the relevance of learning at the start of secondary education than the students in the other levels, but the average decline in perceived emphasis on relevance of learning in the next 1.5 years was stronger than in the other school levels.

As possible explanations for a decline in motivation and achievement in school, we mentioned the shift in the importance of life domains over the years. Students' perspectives seemed, as expected, to play both a positive and a negative role in their academic motivation. The positive effect could mainly be ascribed to the perspectives on school and professional career. The development in students' perspective on leisure, particularly the long-term perspective, showed a strong negative effect on the development of their academic delay of gratification.

Students generally start secondary education well motivated and with positive perspectives. In this study, as in other recent studies on the development of adolescents' motivation for school (Gottfried, Marcoulides, Gottfried, & Oliver, 2009), a decline in academic motivation was found. Is it wise to counteract the decline in the perspectives on school and professional career of the students? Could schools or others intervene in the developments of students' FTPs in order to reduce the decline in their academic delay of gratification or even increase it? As the development of the FTP starts from early adolescence, it could be helpful to guide students planning for a desired goal in the future. Interventions in the students' academic motivation via FTPs have been published (Peetsma & Van der Veen, 2008, 2009), showing some promising results. Furthermore, many schools are changing their education concept in order to reduce students' decline in motivation. The positive results of teachers' emphasis on the relevance of students' learning tasks found in this study are promising in this respect. The way students perceive the learning environment seems to be important for the development of learning.

As a limitation of this study, we have to mention that all measurements were self-reports. However, a number of studies aimed at validating the time perspectives questionnaire have shown good results (Peetsma, 1992). In this study we assumed that students' behavior, academic delay of gratification, would be directly and indirectly influenced, via their FTPs, by the learning environment, in this case the students' perception of the relevance of their learning. The direction of the causal relations between these variables could be different.

In this study we were able to use relatively advanced statistics for the analysis of longitudinal data. We studied the four FTPs separately. In future research, all perspectives could be included in one model, but for that higher numbers of students would be necessary.

In summary, the results showed the expected shifts in FTPs and academic delay of gratification: decreasing student perspectives on school and professional career, and decreasing academic delay of gratification, while the perspectives on leisure increased. However, the increase in leisure perspective does not seem to come at the expense of academic motivation, neither of the perspective on school and professional career nor of academic delay of gratification. An emphasis on the relevance of learning by teachers was found to have a positive effect on students' perspectives on school and professional career, and on academic delay of gratification.

Ancillary