Goal orientation, perceived task outcome and ... - Wiley Online Library

365

British Journal of Educational Psychology (2002), 72, 365–384 2002 The British Psychological Society www.bps.org.uk

Goal orientation, perceived task outcome and task demands in mathematics tasks: Effects on students’ attitude in actual task settings Gerard Seegers 1 * , Cornelis M. van Putten2 and Cornelis J. de Brabander1 1

Centre for the Study of Education and Instruction, Leiden University, The Netherlands 2 Department of Psychology, Leiden University, The Netherlands Background. In earlier studies, it has been found that students’ domain-specific cognitions and personal learning goals (goal orientation) influence task-specific appraisals of actual learning tasks. The relations between domain-specific and taskspecific variables have been specified in the model of adaptive learning. In this study, additional influences, i.e., perceived task outcome on a former occasion and variations in task demands, were investigated. Aim. The purpose of this study was to identify personality and situational variables that mediate students’ attitude when confronted with a mathematics task. Students worked on a mathematics task in two subsequent sessions. Effects of perceived task outcome at the first session on students’ attitude at the second session were investigated. In addition, we investigated how differences in task demands influenced students’ attitude. Variations in task demands were provoked by different conditions in task-instruction. In one condition, students were told that the result on the test would add to their mark on mathematics. This outcome orienting condition was contrasted with a taskorienting condition where students were told that the results on the test would not be used to give individual grades. Sample. Participants were sixth grade students (N = 345; aged 11–12 years) from 14 primary schools. Method. Multivariate and univariate analyses of (co)variance were applied to the data. Independent variables were goal orientation, task demands, and perceived task outcome, with task-specific variables (estimated competence for the task, task attraction, task relevance, and willingness to invest effort) as the dependent variables. *Requests for reprints should be addressed to Gerard Seegers, Centre for the Study of Education and Instruction, Department of Education, Leiden University, PO Box 9555, 2300 RB Leiden, The Netherlands (e-mail: [email protected]).

366

Gerard Seegers et al. Results. The results showed that previous perceived task outcome had a substantial impact on students’ attitude. Additional but smaller effects were found for variation in task demands. Furthermore, effects of previous perceived task outcome and task demands were related to goal orientation. Conclusion. The resulting pattern confirmed that, in general, performance-oriented learning goals emphasised the negative impact of failure experiences, whereas taskoriented learning goals had a strengthening effect on how success experiences influenced students’ attitude.

Many researchers defend the claim that students’ self-referenced cognitions and motivational beliefs influence goal directed learning. These cognitions affect how students will adapt to actual learning tasks and how they evaluate the relation between their effort and task outcome. The cognitions include, among others, attributional style (Weiner, 1986), self-worth (Covington, 1992), self-efficacy (Bandura, 1986, 1989) and achievement motivation (Heckhausen, 1991). In the last twenty years, research on motivation and learning has emphasised the role of goal orientation. Goal orientation affects how students experience learning tasks in achievement settings and give meaning to learning opportunities. Personal learning goals create a framework for how students interpret and evaluate achievement-related information and how they set their goals and effort expenditure for tasks at hand. Contrasting goal orientations have been proposed under different names, including learning versus performance orientation (Dweck & Elliott, 1983), mastery versus performance goals (Ames, 1992; Ames & Archer, 1988; Elliot & Harackiewicz, 1996), and ego versus task orientation (Nicholls, 1984). In the distinction that Nicholls (1984) proposed, task orientation motivates students to evaluate their results positively as long as they experience improved mastery. Standards for comparison are one’s own previous results and aspired achievement level. Situations are preferred where tasks are challenging and where expansion of knowledge can be expected. On the other hand, ego orientation relies on the view that performance reflects mental abilities. A standard for achievement is based on comparing one’s own results with achievement of relevant peers. In learning situations ego orientation will lead to a more competitive attitude, a diminished interest in the learning task per se, and a preference for situations where abilities can be demonstrated. As a consequence, ego orientation reduces the possibilities for expanding knowledge mainly because it intensifies anxiety for negative results and promotes effort avoidance when chances of failure are considered to be high. Avoiding effort prevents the conclusion being drawn that a negative result is caused by insufficient capacity, thus providing a means of avoiding the need to accept the negative consequences of failure. In contrast, task orientation implies that effort is valued. Task-oriented students believe that effort is the cause of task outcome (Dweck & Leggett, 1988), they report more positive feelings, and are more likely to take responsibility for task outcome. The contrast between effects of task and ego orientation lies in the context of threatening errors. When public evaluation is at stake, ego orientation will lead to damaging processes in the face of failure. Fear that this failure will be interpreted by others as confirming insufficient capacity will promote the perception of learning situations as a threat, eliciting anxiety and encouraging self-protective withdrawal. Hence, when task outcome is expected to be negative, ego orientation will have a deleterious effect on how students adapt to the task, viz., how they estimate their

Effects of goal orientation, outcome and task demands

367

competence for the task at hand, its relevance and attraction, and their willingness to invest effort. On the other hand, when confronted with errors, task orientated students can be expected to blame themselves because of insufficient preparation and effort. Nicholls (1984) stated that ego and task orientation are independent student characteristics that are based on the student’s conception of ability. However, Seegers and Boekaerts (1993) found that ego and task orientation have a weak to moderate positive correlation. This pattern is in line with the expectation that task and performance (or ego) orientation have a common conceptual base and will induce effort, a sense of challenge, and absorption in the task (Elliot & Harackiewicz, 1996). Furthermore, several researchers have proposed that these two primary aspects of goal orientation (task vs. ego) do not suffice to explain the concept adequately. Nicholls (1984) suggested that students who combine ego orientation with a high level of perceived competence will seek to demonstrate a superior capacity. However, ego orientation combined with a low perceived competence will lead to avoidance of ability judgments. It has been argued that adding this approach-avoidance distinction gives a more adequate description of the dimensions of ego orientation (Elliot, 1999). This has led to the distinction between self-enhancing and self-defeating ego orientation (Seegers & Boekaerts, 1993; Skaalvik, 1996). Self-enhancing ego orientation is found when students consider their results as a confirmation of their capacity when outcome is positive, whereas self-defeating ego orientation is at stake when students face failure and are mainly occupied by avoiding looking stupid. Seegers and Boekaerts (1993) found an additional distinction in task orientation. A negative task orientation is evidenced when students who display interest in the learning domain and confirm willingness to work hard for good results are confronted with a disappointing task outcome. In this situation, students may feel frustrated. Their frustration is the outcome of a conflict between personal goals and achievement. It is typical of a negative task orientation that students blame themselves when confronted with errors. Students may react by decreasing estimated relevance of tasks, adopting a lower aspired achievement level, and reducing intrinsic motivation, but this situation may also evoke a reversed effect such as investment of extra effort. The difference with self-defeating ego orientation is that in the latter orientation the negative impact of public evaluation is a crucial element, leading to avoidance of effort. Goal orientation in actual task settings Learning situations differ in content and social context, and they dynamically change as a function of the perception of relevant cues. It is important to take adequate account of the impact of the immediate context on students’ cognitions and strategies when they are confronte d with actual learning tasks. This calls for models that explain the dynamics of context sensitive behaviour. To explain how students construct the meaning of the situation, both stable aspects and local conditions have to be accounted for. Following Cantor (1981), Boekaerts (1994) distinguished between personality variables measured at different levels. The domain-specific level of personality variables is operationalised as students’ beliefs about or attitudes toward various subject areas. Domain-specific variables include more stable person characteristics, such as self-efficacy and goal orientation. This level is contrasted with the task-specific level, where the measurement of variables in actual task situations is at stake. For example, estimated competence for and perceived relevance of the task at hand.

368

Gerard Seegers et al.

To explain this process of adaptation, Boekaerts and her co-workers developed the model of adaptive learning. In adapting to cognitive tasks, individuals draw on several sources of information from their knowledge base, such as general and domain-specific beliefs and ideas when determining a task-specific attitude. In this adaptation process, trait-like variables and task-specific elements interact in assigning value to the task. The outcome of this appraisal process is that students determine their estimated competence for the task at hand, its personal relevance and attraction, and their willingness to invest effort in working on this task. In several studies, this model was further operationalised for the mathematics domain (Boekaerts, 1987a, 1988, 1991, 1992; Seegers & Boekaerts, 1993, 1996). The distinction in levels of measurement is reflected in the hierarchical structure of the model of adaptive learning. The domain-specific variables (e.g., goal orientation) influence task-specific variables or appraisals (estimated competence for the task at hand, perceived task attraction, personal relevance) and these in turn determine variables such as willingness to invest effort in the task and task performance. This implies that whether or not students are prepared to invest effort in a maths assignment depends largely on the way they appraise mathematics tasks in situ. In the model, task appraisals mediate the effects of the domain-specific variables on willingness to invest effort. In former studies (cf., Seegers & Boekaerts, 1993) this mediating effect has been further investigated. A global representation of the model is shown in Figure 1.

Figure 1.

Main structure of the model of adaptive learning

In earlier studies a number of different domain-specific variables were included in addition to goal orientation (cf., Seegers & Boekaerts, 1993). However, goal orientation plays a central role in determining the relevant task-specific variables. Students’ goal orientations are basic to how students adapt to learning opportunities, to how they set their learning goals and effort expenditure for the task at hand by giving meaning and value to actual learning opportunities and to the social context in which these opportunities occur. Goal of the research The experimental study is designed to contribute to a further test of the model of adaptive learning. An implicit assumption in this model is that it entails the long-term development of the students’ attitude, i.e., knowledge, beliefs, and personal learning goals. Students develop an attitude towards tasks in a domain as the result of a number of (positive and negative) experiences. How they assess previous experiences is considered crucial in the build-up of relevant knowledge. This experiential knowledge becomes part of the student’s relevant domain-specific knowledge. In this study, this development is further investigated by comparing variations in students’ attitudes on two subsequent confrontations with challenging mathematics tasks by exploring effects


369

of previous experiences on how students adapt to a learning task. To study these effects, students were asked how they estimated task outcome on a first test session. It is expected that previous perceived task outcome influences students’ task appraisals and effort investment when confronte d with a similar task on a subsequent session. Given that they perceive the outcome on the first session as good, this will have a positive effect on how they appraise the task on a subsequent confrontation. It is predicted that students will estimate their competence as being higher, that they will experience the task as more attractive and that they will report more willingness to invest effort. When previous perceived task outcome is poor, a reversed effect is expected. Goal orientation is expected to have a moderating effect on the relation between previous perceived task outcome and subsequent task appraisals. For example, ego orientation is predicted to strengthen the negative effects when the perceived task outcome is considered disappointing. Ego-oriented students who perceive the previous task outcome as negative will estimate their competence as lower and will avoid investment of effort in a subsequent task because the chances are high that they cannot demonstrate adequate capacity by working on it. For task-oriented students, no such effect is expected. As long as they see an opportunity for improved mastery, they will maintain competence, estimated relevance, and effort investment at the same level. In this study, an additional task-specific variable is included. Variations in instruction are used to manipulate perception of task demands. In one condition, instruction emphasised outcome (performance). Students were told that the result of the task counted for their grade. This outcome orienting instruction was expected to result in a high evaluative load of the task. In a second condition, students were stimulated to do their best, but they were told that results would not be awarded. In this effort orienting instruction, task demands were expected to be less considerable. Again a moderating effect of goal orientation is predicted. In general, ego-oriented students will tend to avoid the task when chances for a disappointing outcome are considered to be high. An acceptable way out is to avoid effort investment. Lack of effort offers an acceptable alternative explanation for a negative outcome. It can be expected that, in particular, ego-oriented students will reduce estimated relevance of the task and effort investment when the evaluative load of the task is low. In this situation, task outcome is less relevant and it is easy to escape a possible confrontation with failure. No such effect is expected for task-oriented students. Considering combined effects of previous perceived task outcome and variation in task demands, the pattern is even more complicated because previous perceived task outcome and task demands may interact in their effects. This implies that relations between goal orientation and task appraisals may depend on specific combinations of task demands and previous perceived task outcome. This leads to an overall formulation of the research question: how do different combinations of task demands and previous perceived outcome affect moderating effects of goal orientation on task appraisals, willingness to invest effort and task performance? In Figure 2, the elaborated version of the model of adaptive learning is depicted. In this figure, relations between goal orientation, previous perceived task outcome, task demands on task appraisals, willingness to invest effort and task performance are given. Furthermore, interaction effects of task demands and previous perceived task outcome with goal orientation are included. In line with the hierarchical structure of the model of adaptive learning, task appraisals (partly) mediate the effects of the independent variables on willingness to invest effort and task performance.

370


Figure 2.

Elaborated model of adaptive learning

Method Instruments The Goal Orientation Questionnaire consisted of 21 statements about experiences and behaviour in maths class. Students rated on a 4-point scale to what extent they usually react in the described manner. Answers ranged from ‘never’ to ‘always’. Applying Principal Components Analysis to the data, four relevant latent variables were extracted, explaining 56% of the variance. The obtained principal factor structure was subjected to an orthogonal rotation. The results are given in the Appendix. Subscales are defined on the basis of this solution. In the analyses, item scores are used. The first subscale measures task orientation (five items, e.g., ‘I feel satisfied when I have learned something in maths that makes sense to me’), error frustra tion, which indicates that failure expresses a personally experienced lack of ability (four items, e.g., ‘When I have made my assignments not as well as usually, I cannot stand myself’), selfenhancing ego orientation (in cases where ability can be demonstrated) (six items, e.g., ‘I enjoy getting a better grade in maths than my classmates’), and self-defeating ego orientation which refers to situations where pupils are afraid that others will notice their shortcomings (six items, e.g., ‘During maths tasks I am afraid that other students will notice my mistakes’). Internal consistencies (Cronbach’s alpha) were .81 for self-enhancing ego orientation, 0.76 for self-defeating ego orientation, .77 for task orientation, and .78 for error frustration. As for the correlations between these subscales, there was a moderate positive correlation between self-enhancing and self-defeating ego orientation (r = .31). There was also a positive correlation between task orientation and error frustration (r = .19). A positive correlation was found for self-enhancing ego orientation with task orientation (r = .23) and with error frustration (r = .34). As for self-defeating ego orientation, there was no relation with task orientation (r = .07), although there was a positive correlation with error frustration (r = .45). The On-Line Motivation Questionnaire (OMQ) was developed by Boekaerts (cf. Boekaerts, 1987b, 2001) to obtain learners’ judgments about relevant aspects of the


371

learning situation when confronte d with an actual learning task. The four constructs that are measured include the three appraisals (subjective competence, task relevance, and task attraction) and willingness to invest effort. Subjective competence includes eight items concerning estimated competence for the task at hand (e.g., ‘How well do you think you can do this type of task?’), outcome expectancy (e.g., ‘How well do you expect to do on this task?’), and perceived level difficulty (e.g., ‘How difficult do you consider this kind of task?’); task relevance is measured with two items (e.g., ‘How important do you find it to do well on this sort of task?’); task attra ction is measured with three items (e.g., ‘How pleasant do you find this type of task?’). Willingness to invest effort concerns student’s preparedness to invest effort in doing the task. The scale to measure this construct includes four items (e.g., ‘How much effort are you going to put into this task?’ with answers ranging from ‘very little’ to ‘my very best’). Internal consistencies (Cronbach’s alpha) in this sample were .90 for subjective competence, .76 for task attra ction, .69 for task relevance, and .68 for willingness to invest effort. There were high correlations between subjective competence and task attraction (r = .56), and between task relevance and willingness to invest effort (r = .59). A moderate correlation was found between task attraction and willingness to invest effort (r = .27). The remaining correlation coefficients were small. Mathematics test. A mathematics test was developed. This test included the released items from a national assessment study (Wijnstra, 1988). The items cover the relevant topics of the arithmetic curriculum (viz., algorithmic problems, mental arithmetic, fractions, ratios, measuring, and percent problems). Problems in all topics varied from more simple algorithmic problems to complex problems. The items were divided into two sets, each including 52 items. As estimates of the item complexity were known, the two sets were parallel with respect to topic and item difficulty. Task performance is defined as the number of correctly solved items on the test. Previous perceived task outcome. After completing the task, students were asked to assess their task outcome on a 4-point scale (ranging from ‘very poor’ to ‘very good’).

Participants Fourteen schools for primary education in the urban region of Leiden (The Netherlands) participated in the study. Complete data were collected for a total of 345 students from the sixth grade (age 11–12). The sample included 179 boys and 166 girls.

Procedure Students participated in two mathematics test sessions which were about one week apart. In each session, students filled out the Online Motivation Questionnaire. The Goal Orientation Questionnaire was administered about one week before the first test session. Two different versions of instruction were used to provoke differences in the perception of task demands. These versions emphasise an outcome orientation and an effort orientation respectively. In the outcome orienting instruction, students were told: ‘In the next 45 minutes, you will be working on a mathematics test. Work as best as you can. This is a real test, and a mark will be awarded for its result. This mark will be given to your teacher and will be used to evaluate school progress.’

372


In the effort orienting instruction, students were told: ‘In the next 45 minutes, you will be working on a number of mathematics problems. We are going to compare the results with those of other students because we want to know whether students your age are able to solve these problems. Do your best. This is not a test, and its result will not be awarded a mark.’

To compare effects of perceived task demands, whole classes were randomly assigned to one of the conditions. To vary task demands, all students worked on the test under both instructions, but in a different order. One group (N = 165) worked on the task in the outcome orienting instruction in the first test session, and received the effort orienting instruction in the second session. The second group (N = 180) started with the effort orienting instruction. Each session took about one hour. Students were allowed to work on the mathematical task for 45 minutes. Filling out the OMQ took about 15 minutes. Students remained seated in their classroom. They were asked to write their names on the answer sheets. Instruction was given by advanced students in education as part of their study requirements. Teachers were allowed to remain in the classroom, but were asked not to interfere.

Analyses Multivariate and univariate analyses of (co)variance were applied to the data. The overall research question is how specific combinations of task demands, previous perceived task outcome and goal orientation influence task appraisals, willingness to invest effort and task performance. With previous perceived task outcome and task demands as between-subjects variables (factors), interaction effects with goal orientation (covariates) were tested (testing homogeneity of slopes). Significant interaction effects would indicate that effects of goal orientation are related to (combinations of) previous perceived task outcome or task demands. In this case, effects were tested separately within the levels of the factor separately. The model of analysis is in line with the hierarchical structure of the model of adaptive learning. In earlier studies, evidence was found that the task appraisals (subjective competence, estimated task relevance, task attraction) mediate the effects of the domain-specific variables on willingness to invest effort and task performance. As a consequence, analyses were done in two steps. In the first step, task appraisals were included as the dependent variables, with goal orientation, task demands and previous perceived task outcome as the independent variables. In a second step, willingness to invest effort in doing the task and task performance were included as the dependent variables, and effects of goal orientation, task demands, previous perceived task outcome and task appraisals were tested. As for task demands, participants were divided into two groups. On the first session, the groups respectively received the effort orienting and outcome orienting instruction. On the second session, both groups received instructions that were different from the first session. Combining the data of the two test sessions, the variable task demands refers to the different combinations of instruction that students received (viz., effort+outcome and outcome+effort). This was done to investigate how students adapt to changes in task demands. In Figure 2, the structure of the elaborated model of adaptive learning is given. This model clarifies the steps that have been taken in the analysis of the data. To interpret


373

the effect of task demands in a more detailed way, the data from the first test session were analysed separately. The results of the following analyses are described: –

data first session; dependent variables were task appraisals (subjective competence, task relevance, task attraction), with task demands as the between-subjects variable, and with goal orientation as the covariates;

–

data first session; dependent variables were willingness to invest effort and task performance, with task demands as the between-subjects variable, and with goal orientation and task appraisals as the covariates;

–

data both sessions; dependent variables were task appraisals on the second session, with previous perceived task outcome and task demands as the between-subjects variables, task appraisals on the first session and goal orientation as the covariates;

–

data both sessions; dependent variables were willingness to invest effort and task performance on the second session, with previous perceived task outcome and task demands as the between-subjects variables, willingness to invest effort and task performance on the first session, task appraisals on the second session, and goal orientation as the covariates.

Results Mean scores and standard deviations on the elements of goal orientation for the two groups (outcome orienting vs. effort orienting) are given in Table 1. Applying a multivariate analysis of variance (Manova) with the four elements of goal orientation as the dependent variables, and task demands (outcome vs. effort orienting) as the between-subjects factor, no differences between the groups were found, Pillais F(4,374) = 1.24, p > .10. It can be concluded that the two groups do not differ with respect to goal orientation. Table 1. Mean scores and standard deviations for the elements of goal orientation for different task demands (outcome vs. effort orienting)

Task orientation Error frustration Self-enhancing ego orientation Self-defeating ego orientation

Outcome orienting (N = 165) M SD

Effort orienting (N = 180) M SD

3.03 2.14 2.19 1.43

2.96 2.19 2.20 1.53

.59 .71 .67 .47

.61 .68 .69 .47

First session: Task demands and task appraisals The data from the first test session were analysed to investigate effects of task demands on the task appraisals. Results are shown in Table 2. Dependent variables were the task

374


Table 2. Mean scores on the task appraisals (first sessions) separately for the two levels of task demands (outcome vs. effort orienting)

Subjective competence Task attraction Task relevance Willingness to invest effort Task performance

Outcome orienting (N = 165) M SD

Effort orienting (N = 180) M SD

2.69 2.82 3.05 3.28 25.68

2.65 2.76 2.87 3.11 22.29

.46 .57 .55 .39 7.86

.50 .62 .54 .44 8.51

appraisals (subjective competence, task attraction, and task relevance), with task demands as the independent variable, and with the elements of goal orientation as covariates. Interaction effects of task demands and goal orientation were included to test for homogeneity of slopes. The multivariate results showed a statistically significant effect of the interaction of self-defeating ego orientation and task demands, Pillais F(3,333) = 4.22; p < .01. Univariate testing showed this effect to be statistically significant on subjective competence, F(1,335) = 8.57, p < .01, but not on task attraction, F(1,335) = .13, and task relevance, F(1,335) = 2.21, p > .10. No further interaction effects were found to be statistically significant. Hence, effects of self-defeating ego orientation on subjective competence were tested separately within the two levels of task demands (outcome orienting vs. effort orienting). A survey of the results for the various dependent variables is given in Table 3. With subjective competence as the dependent variable, significant effects were found for selfenhancing ego orientation, b = .13, t = 3.48, p < .01, task orientation, b = .18, t = 4.44, p < .01, and error frustra tion, b = 7 .10, t = 7 2.65, p < .01. These results imply a positive effect of both self-enhancing ego orientation and task orientation on subjective competence for the task at hand, whereas error frustration exerts a negative influence. The effect of self-defeating ego orientation was related to differences in task Table 3. Effects on task appraisals, willingness to invest effort and task performance: first session

Task orientation Error frustration Self-enhancing ego orientation Self-defeating ego orientation Task demands Self-defeating ego orientation* task demands – Outcome orienting – Effort orienting

*p < .05; **p < . 01

Subjective competence

Task Task Willingness attraction relevance to invest effort

.18** 7.10** .13**

.24**

.14**

Task performance

.18**

72.47* .29**

7.37** 7.14


375

demands. When task demands stressed outcome, a negative effect was found for selfdefeating ego orientation on subjective competence, b = 7.37, t = 74.69, p < .01. A clearly smaller effect (lacking statistical significance) was found when task demands were task-oriented, b = 7 .14, t = 71.87, p > .05. Inspection of the data showed that a higher score on self-defeating ego orientation only resulted in a lower estimated competence for the task at hand when task demands emphasised outcome. With task attra ction as the dependent variable, an effect was only found for task orientation, b = .24, t = 4.71, p < .01. A significant effect of task orientation was also found with task relevance as the dependent variable, b = .14, t = 2.98, p < .01. First session: Task demands, willingness to invest effort investment and task performance Effects of goal orientation on willingness to invest effort and task performance were tested. In addition to the elements of goal orientation task appraisals (subjective competence, task attraction, and estimated personal relevance) were included in the model as covariates. In a first step, the interaction effects of goal orientation and task demands were included to test for homogeneity of slopes. The results showed no statistically significant interaction effects. Hence, effects of the elements of goal orientation on the dependent variables were independent from task demands. With willingness to invest effort as the dependent variable, an effect was found for the appraisal task relevance, b = .42, t = 13.34, p < .01. An additional effect was found for task demands, b = .29, t = 3.02, p < .01. When task demands emphasised outcome students reported a somewhat greater willingness to invest effort (mean scores are 3.28 and 3.11 respectively). Finally, task orientation had a statistically significant effect, b = .18, t = 6.34, p < .01. The data showed a general positive effect of task orientation on reported willingness to invest effort. With task performance as the dependent variable, subjective competence had a significant effect, b = 5.05, t = 4.39, p < .01. Furthermore, the results showed a negative effect of self-defeating ego orientation, b = 72.47, t = 72.31, p < .05. Comparison of test sessions: Previous perceived task outcome, task demands, and task appraisals As a next step, the data of the two test sessions were combined. After completing the task, students were asked to indicate how they estimated task outcome. Comparing the data on two subsequent test sessions allowed the investigation of the effects of this previous perceived task outcome. Furthermore, variations in task demands are included. Task demands here refers to the two combinations of instruction students received on the two test sessions (either a outcome+effort orienting or effort+outcome orienting sequence). Previous perceived task outcome was measured on a 4-point scale (very poor, poor, good, very good). However, most students indicated that they estimated task outcome either ‘poor’ or ‘good’. For this reason scores on previous perceived task outcome have been reduced to two values (good and poor). In the model to be tested, task appraisals as measured on the second session were included as the dependent variables. These appraisals as measured on the first session and the elements of goal orientation were included as covariates. Previous perceived task outcome and task demands were between-subjects variables. Results are given in Table 4.

Good Poor

Good Poor

Task relevance

Willingness to invest effort

3.16 3.01

2.95 2.86

2.74 2.32

2.79 2.23

.41 .51

.56 .64

.54 .59

.35 .45

SD

Effort+outcome Outcome+effort




Task demands

* Multiplied by giving all combinations of perceived outcomes and task demands

Good Poor

Task attraction

1

Good Poor

Subjective competence

Perc. outcome M

3.11 3.10

2.95 2.87

2.53 2.65

2.53 2.66

M

.46 .45

.59 .58

.59 .59

.48 .45

SD

poor

good

poor

good

poor

good

poor

good

Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort Effort+outcome Outcome+effort

Perc. outcome* Task demands1 2.77 2.82 2.18 2.32 2.72 2.75 2.25 2.41 3.00 2.90 2.89 2.81 3.16 3.15 3.05 2.99

M

.34 .36 .45 .43 .53 .56 .60 .59 .56 .57 .65 .62 .42 .41 .51 .52

SD

Table 4. Mean scores and standard deviations for the task-specific appraisals (subjective competence, task attractiveness, task relevance), willingness to invest effort and task performance on the second test session separate for task demands and previous perceived task outcome

376 Gerard Seegers et al.


377

In a first step, it was tested whether effects of goal orientation on students’ attitude were related to previous perceived task outcome and task demands. Hence, interaction effects of previous perceived task outcome and task demands with goal orientation were included in a preliminary model to test for homogeneity of slopes. In one comparison, the two-way interaction effect of task demands, perceived task outcome and error frustra tion was statistically significant, Pillais F(3,315) = 3.98, p < .01. Univariate testing showed an effect of error frustration on task attra ction, F(1,317) = 7.07, p < .01, but not on task relevance, F(1,317) = 2.79, p > .05, or subjective competence, F(1,317) = .00. Furthermore, the interaction of previous perceived task outcome with self-defeating ego orientation was statistically significant, F(3,315) = 2.49, p < .05. Univariate testing showed that this effect was significant on task relevance, F(1,317) = 8.39, p < .01, but not on subjective competence, F(1,317) = .10, or task attra ction, F(1,317) = .06. A final statistically significant interaction effect was found for perceived task outcome with task orientation, Pillais F(3,315) = 2.70, p < .05). Univariate testing confirmed this effect to be significant on ta sk relevance, F(1,317) = 4.80, p < .05, but not on subjective competence, F(1,317) = 2.65, p > .05, or task attra ction, F(1,317) = .74. As a consequence, in the definite model, effects of error frustration on task attraction and task relevance were tested within the combined levels of perceived task outcome and task demands. The effect of self-defeating ego orientation on task relevance was tested for the two levels of previous perceived task outcome separately. Finally, effects of task orientation on subjective competence and task relevance were tested within the levels of previous perceived task outcome. Results of testing the effects on the dependent variables in the definite model are given in Table 5. With subjective competence (as measured at the second session) as the dependent variable, the score was partially explained by its parallel measurement on the first session, b = .66, t = 18.51, p < .01. The results showed that perceived task outcome had a statistically significant effect, b = .10, t = 5.69, p < .01. Differences in estimated subjective competence (corrected for the effect of its first measurement) were related to perceived task outcome: observed mean scores are 2.79 and 2.23 for students who reported a positive and a negative task outcome respectively. Furthermore, the (positive) effect of task orientation was significant, b = .07, t = 2.63, p < .01. Task-oriented students were more prepared to maintain estimated competence for the task at hand at a high level. With task attra ction as the dependent variable, its parallel measurement on the first session explained a substantial part of the variance, b = .57, t = 12.93, p < .01. An additional effect was found for task relevance, b = .11, t = 2.25, p < .05. Accounting for these effects, the influence of previous perceived task outcome was found to be statistically significant, b = .23, t = 2.73, p < .01. Students who reported a disappointing outcome on the first session considered the task less attractive (mean score 2.32) than students who considered task outcome successful (mean score 2.74). Furthermore, effects of error frustration on task attraction were different for the levels of both task demands and perceived task outcome. Analysing effects of error frustration within the levels of the two factors showed this effect only to be significant in the situation that students received the effort orienting instruction on the second session whereas they perceived task outcome on the first session to be good. In this situation error frustra tion has a negative effect on task attraction, b = 7.14, t = 72.39, p < .05. Hence, students who reported a higher level of error frustration reduced task attraction in the situation where evaluative load of the task was decreased and a previous experience was considered positive.

*p < .05 ** p