We also inspected the entropy, which indicates the precision with which the situations are classified into the profiles. After choosing a final model, we applied a probability sampling procedure for the further analyses Sahdra et al. This procedure is an alternative to saving the most likely latent profile membership and takes into account the uncertainty that is associated with classifying learning situations into the latent profiles.
The probability sampling was based on the profile probabilities i. From these distributions, we sampled 25 data sets containing profile membership information that were analyzed separately in the following analyses. Finally, the results were aggregated appropriately across the parameter estimates obtained in each of the 25 model runs Rubin, On the within-level, latent profile membership indicates that a given learning situation belongs to a certain situational motivation profile.
To examine hypothesis 2a, we regressed dispositional motivation at the beginning Time 1 and end of the semester Time 2 on profile membership profile differences model. To examine hypothesis 2b, we modeled the change score difference from Time 1 to Time 2. The change score was then regressed on latent profile membership differential change model.
We estimated separate models for dispositional expectancies, dispositional values, and dispositional costs 6 models in total , all of which were saturated. In all analyses, we dealt with missing data using full-information maximum likelihood estimation, which uses all available data without imputing missing values Schafer and Graham, We tested for measurement invariance of all dispositional constructs across time using confirmatory factor analysis CFA. We determined measurement invariance following Cheung and Rensvold who suggested that if the decrease in CFI is not more than 0.
Table 2 shows the fit indices for the most invariant model for each construct. Full scalar invariance held for all constructs except emotional cost, where we used a partial invariance model.
For each construct, factor scores from the most invariant model were saved and used in the primary analyses. Thus, the factor scores for this scale were based on a model with partial invariance. Table 2. Descriptive statistics and model fit information for tests of measurement invariance of the dispositional motivation measures.
Before carrying out our substantial analyses, we computed intra-class correlations ICCs to determine the amount of variability that was due to different situations within-level and students between-level. Latent variable ICCs were 0. This substantial amount of variance that was due to differences between students justified the multilevel approach used for further analyses.
This implies that students on average affirmed experiencing success expectancies and values, while they on average denied the items asking about costs. Table 3 shows the model fit indices for latent profile models with up to six profiles.
Although the information criteria suggested to add more latent profiles, the best likelihood did not replicate in the 5- and 6-profile models. In contrast to hypothesis 1, we did not find the expected discrepant constellations of high expectancies and low values, and vice versa. Figure 2. Depicted are means and confidence intervals. To examine change and stability in situational profile membership, we computed transition probabilities from one motivational profile to the next within one lecture session.
Changes mostly occurred between the high motivation and the low cost motivation profiles probabilities of 0. Next, we tested whether the experiences of motivational states latent profiles were associated with dispositional expectancies, values, and costs at the beginning and end of the semester profile differences models.
The results are depicted in Table 5. Table 5. Associations between profiles of situational motivation and dispositional constructs. As predicted in hypothesis 2a, students who often experienced highly motivating situations during the lecture reported higher success expectations in the beginning of the semester Time 1 , compared to frequent experiences of all other situational profiles. Concerning task values, students who often experienced highly motivating situations high expectancies and values, and low costs during the lecture tended to hold higher intrinsic, attainment, and utility value both at Time 1 and Time 2, compared to frequent experiences of all other situational profiles, which is again in line with hypothesis 2a.
An exception was utility value at Time 1, which did not differ between students who often experienced highly motivating versus motivating but costly situations medium expectancies and values, and above average costs. In contrast, frequent experiences of low motivation low expectancies and values, and above average costs were associated with lower attainment and utility value at both Time 1 and Time 2, and with lower intrinsic value at Time 2, compared to all other situational profiles.
Finally, students who often experienced motivating but costly situations held higher attainment and utility values at Time 1, and reported higher attainment value at Time 2 than students with frequent low cost motivation situations. Also as predicted, we found that students with frequent experiences of high motivation during the lecture had lower effort, emotional, and opportunity costs than other students at both Time 1 and Time 2.
These students did, however, not differ significantly at Time 2 from students with frequent low cost motivation situations on effort cost and emotional cost. Moreover, students who often experienced motivating but costly situations reported higher opportunity cost at Time 1 and Time 2, compared to students with frequent low cost motivation situations.
Next, we examined whether experiences of situational profiles predicted changes in dispositional motivation over the course of the semester differential change models , see Table 5. We found differential change in task values, but not in regard to expectancies and costs.
The results showed that frequently experiencing highly motivating situations was associated with more positive change in intrinsic and attainment value, compared to frequent experiences of other motivational situations.
Moreover, students who often experienced motivating but costly situations tended to show more positive change in attainment value compared to students with frequent experiences of low cost motivation or low motivation situations.
Also, students with frequent experiences of motivating but costly situations increased their intrinsic value relative to students with frequently low motivation. These findings concerning task values were in line with hypothesis 2b. When teachers aim to provide a learning environment which optimally fosters engagement in learning they need to take into account that students not only bring different motivational dispositions, but also vary over time in their motivational state.
We identified four different profiles of expectancies, values, and costs within specific learning situations during a university course and examined situational change and stability. In accord with our expectations, our findings suggest that students experience most learning situations in terms of similar expectancies and values, and opposite levels of costs — regardless of their overall level of motivation high, medium, or low in that situation.
In high motivation situations , high expectancies and values occurred together with low costs; in low motivation situations , low expectancies and values occurred with above average costs; and in low cost motivation situations , medium expectancies and values occurred with low costs.
Together, these symmetric profiles accounted for Our situation-level results corroborate earlier studies on dispositional motivation reporting mainly such aligned profiles of expectancies and values e. Moreover, there was one profile in which medium values and expectancies occurred together with above average levels of costs motivating but costly situations. This finding underscores the particular merit of analyzing profiles.
The finding also aligns with existing studies describing co-occurring positive and negative aspects of academic emotions Pekrun et al. The in-the-moment profiles of motivational beliefs tended to be relatively stable from one learning situation to the next within one lecture session. However, for students in the high motivation profile there was a chance of 0. But also for students in the low motivation profile, the chances were 0. Overall, our data suggest that if a student changed her motivational profile, that change more likely occurred to a profile with similar levels of costs but higher or lower expectancies and values.
This might indicate that situational expectancies and values could be more malleable to change than situational costs. A fruitful avenue for future research would be to examine the contextual characteristics of the learning situations which prompt shifts in motivational profiles. Information from such studies could be useful in the context of adaptive teaching Corno, where teachers adapt their instruction based on the motivational needs of certain groups of students. Still it remains an open question for future studies to clarify the extent to which profile change and stability can be attributed to individual and context characteristics.
Contrary to our expectation, we did not find situational profiles with discrepant expectancies and values e. This is in contrast to previous studies that found that some students reported high levels of self-concept but low levels of interest for a given subject, or vice versa Viljaranta et al. It could be that this difference in finding is due to differences in measures situational, asking about the current learning situation versus dispositional, asking about the entire subject , or due to different age groups studied, since the correlations between expectancies, values, and costs are typically lower in children than in older students Wigfield et al.
Other differences between our study and earlier ones are that we studied motivation in a university Psychology course, while Lazarides et al.
Our findings could also be influenced by the fact that the university students in our sample were not obliged to attend the lecture, whereas school students do have to attend classes.
It is possible that the students with high expectancies but low values, who think that learning Psychology is easy but unimportant, decide to not attend lessons and only engage in self-study to prepare for the exam.
For example, students with higher dispositional success expectations, higher dispositional values, and lower dispositional costs were more likely to experience states of high motivation during the lecture.
Students who frequently experienced low motivation in the lecture showed more negative development in their intrinsic and attainment value towards the subject Psychology, compared to other students. However, this only applied to task values, but not to expectancies or costs. By contrast, typical strategies of changing expectancy beliefs through feedback and attribution e.
Students with frequent motivating but costly situations strengthened their dispositional task values attainment, partly intrinsic more strongly than most other students, except those with frequent high motivation. At the same time, these students with many motivating but costly experiences also reported higher dispositional costs of learning Psychology, compared to many other students, both in the beginning and at the end of the semester.
Not much is known so far about the long-term development of students with co-occurring positive and negative motivation. Overall, the results of this study support the notions of dynamic systems theory e. Indeed the motivational dispositions that students bring into a learning situation may affect their motivational experience during learning, and in the case of task values, such situational motivation experience may contribute to the development of inter-individual differences in more stable motivational beliefs.
However, our findings indicate that situational experience contributed mainly to pre-existing stability in dispositional motivation, and only in the case of values in increased inter-individual differences. However, some limitations of this study give suggestions for future research. First, we investigated expectancies, values, and costs in only one context and population, namely adult teacher students experiencing an introductory lecture about Educational Psychology.
It is thus unclear whether these findings can be generalized to students learning in schools or other, more active learning forms, such as group work. Replications in more diverse and representative student samples could examine whether different learning contexts elicit similar profiles as described here, or if not, which exact person and context characteristics determine the situation-level profiles of expectancies, values, and costs.
In particular, we did not find difficult but valuable situations but expect that they occur in more active and challenging learning situations. Also, the small sample precluded us from conducting profile analyses on the level of students. It would thus be interesting to compare situational within-person and dispositional between-person profiles in a larger sample in the future Voelkle et al.
Third, in our multilevel analyses, the learning situations were treated as interchangeable measurements. This ignored any developmental dynamics that might have happened during the semester.
One such dynamic might be a higher motivational variability in the beginning of a course when the subject and the teacher are new, while later on after making repeated experiences, certain motivational states might occur more easily or more rarely in certain students. Finally, while the design of this study enabled us to investigate within-lesson changes of motivational profiles, the here reported findings are bound to the small time period between the measurements 30 min.
Because some other studies suggest that more variance in state motivation and emotions can be attributed to particular learning situations than for example to different days of the week e.
Moreover, intra-individual motivational profiles revealed that in some learning situations, positive and negative aspects of motivation co-occurred in motivating but costly situations , which is easily overlooked in the typically applied correlational analyses. How such mixed motivation relates to educational outcomes in the short- or long-term see Guo et al. This study was carried out in accordance with the recommendations of the APA Code of Ethics with written informed consent from all subjects.
In other words, for what kinds of people and in what kinds of situations are expectancy effects more likely to occur? Research examining these questions indicates that, while there are individual differences that moderate expectancy effects, such as self-esteem, gender, and cognitive rigidity, situational factors such as the relative power of the perceiver and target and how long they have known each other appear to be more important predictors of expectancy effects.
An expectancy effect is more likely to occur when the perceiver is in a position of greater power than the target such as in a teacher-student relationship and when the perceiver and target have not been previously acquainted.
The longer the individuals know each other, the less likely it is that perceivers will either form or be influenced by incorrect expectancies. Relatedly, much of the recent research in this area has been dedicated to the question of determining how powerful expectancy effects are in naturally occurring contexts as opposed to the laboratory.
Laboratory experiments typically yield expectancy effects of larger magnitude. In the real world, accuracy effects i. While the specific mediating behaviors involved depend on the context of the interaction, the vast majority can be classified as falling into the dimensions of affect or effort. Affect refers to the socioemotional climate that is created by the perceiver, and it involves primarily nonverbal cues associated with warmth and friendliness.
Expectancy theory, initially put forward by Victor Vroom at the Yale School of Management, suggests that behavior is motivated by anticipated results or consequences. Vroom proposed that a person decides to behave in a certain way based on the expected result of the chosen behavior. For example, people will be willing to work harder if they think the extra effort will be rewarded. In essence, individuals make choices based on estimates of how well the expected results of a given behavior are going to match up with or eventually lead to the desired results.
Expectancy theory has three components: expectancy, instrumentality, and valence. Expectancy theory, when properly followed, can help managers understand how individuals are motivated to choose among various behavioral alternatives. To enhance the connection between performance and outcomes, managers should use systems that tie rewards very closely to performance. They can also use training to help employees improve their abilities and believe that added effort will, in fact, lead to better performance.
0コメント