In this paper first the influence of cognitive motivation (NFC) and task difficulty on the process of expectancy formation should be analyzed. We assume, that differences in task difficulty only result in differences in performance expectancies when cognitive motivation is high (high NFC) but not when cognitive motivation is low (low NFC). This should be the case because analyzing the demands of the task (including analyzing task difficulty) is a process requiring cognitive capacity. Individuals low in need for cognition should less likely engage in such cognitive endeavors. Second, in line with considerations of Marshall and Brown (2004) we predict, that expectancies should not affect actual performance when the task is easy (because higher effort or higher persistence [as an effect of high performance expectancies] are not likely to increase performance in rather easy tasks). For difficult tasks we argue that expectancies should affect performance only when expectancies were formed via intensive processing (i.e. when NFC is high). When expectancies were formed via peripheral processing (i.e. when NFC is low), the expectancies should not be related to actual performance. The results of study 1 (N = 175 university students) were in line with our hypotheses: the interaction of NFC and task-difficulty affected expectancies. Only for participants with higher NFC task-difficulty had an influence on expectancies of success (with lower expectancies for the difficult task than for easy task). Given low NFC, no effect of task difficulty on the expectancies was found. Study 2 replicates the findings of Study 1 with a different task (analogy-task) and found also support for our second hypothesis: Significant correlations of expectancy of success and actual performance were only found for participants with higher NFC and when the task was difficult. The findings from both studies have several implications for research on self-regulated learning.

Performance expectancies play an important role in explaining individual’s learning outcomes. If individuals build their performance expectancies concerning a specific task, they have to analyze the specific requirements of the task at hand, and to use their ability-related self-knowledge (self-concept). The process of analyzing the requirements of the task and the comparison of the task demands with the self-concept requires cognitive capacity. The degree of cognitive effort individuals are willing or able to invest in this process may depend on situational variables (such as distraction or personal relevance) and on dispositional differences in cognitive motivation (e.g., need for cognition). In a series of experiments, Dickhäuser and Reinhard (2006) found that the extent to which individuals like engaging in and enjoy thinking (“Need for Cognition”, NFC) determines whether their expectations of success can be more accurately predicted based on their general or their task-specific self-concept. For individuals with higher NFC, the expectancy of success (and actual performance) could be predicted from specific but not from general self-concept, whereas general self-concept was a significant predictor for expectancies of success (and performance) in individuals lower in NFC. This finding is in line with the prediction, that the use of the general self-concept requires less cognitive capacity (and therefore is preferred by individuals low in NFC) as compared to the use of the specific self-concept.

Given that Dickhäuser and Reinhard found first support for the assumption that the way people form expectancies depends on their cognitive motivation, the aim of this paper is to carry on this idea. First the influence of cognitive motivation (NFC) and task difficulty on the process of expectancy formation should be analyzed. We assume, that differences in task difficulty only result in differences in performance expectancies when cognitive motivation is high (high NFC) but not when cognitive motivation is low (low NFC). This should be the case because analyzing the demands of the task (including analyzing task difficulty) is a process requiring cognitive capacity. Individuals low in need for cognition should less likely engage in such cognitive endeavors. Second, in line with considerations of Marshall and Brown (2004) we suppose that cognitive motivation and task difficulty influence the relation of expectancies and actual performance. In accordance with Marshall and Brown we predict, that expectancies should not affect actual performance when the task is easy (because higher effort or higher persistence [as an effect of high performance expectancies] are not likely to increase performance in rather easy tasks). For difficult tasks we go beyond the prediction of Marshall and Brown. While Marshall and Brown predicted that expectancies should affect performance when the task is difficult we argue that this should only be the case when expectancies were formed via intensive processing (i.e. when NFC is high). When expectancies were formed via peripheral processing (i.e. when NFC is low), the expectancies should not be related to actual performance. We assume that individuals low in NFC did not accurately analyze the task (in particular they did not accurately analyze task difficulty). Therefore, the so formed expectancy do not fit with the requirements of the given task. These hypotheses were tested in two independent studies.

In study 1 (N = 175 university students) participants NFC was assessed. Next, participants were told that they had to work on a dexterity task (building a pyramid out of mugs). We directly manipulated the task difficulty by varying the number of mugs participants got to build the pyramid up. In the difficult-task condition participants received ten mugs. They were told, that they had to build a pyramid with four mugs in the bottom row, three mugs in the second row, two in the third and the last mug on the top of the pyramid. In the easy-task condition participants had to build a pyramid out of six mugs (three mugs in the first row, two in the second and the last mug on the top). The expectancy of success was assessed by the following measure: “Given two minutes time, I will build … pyramids.” The results were in line with our hypotheses: the interaction of NFC and task-difficulty affected expectancies. Only for participants with higher NFC task-difficulty had an influence on expectancies of success (with lower expectancies for the difficult than for the easy task). Given low NFC, no effect of task difficulty on the expectancies was found.

Study 2 was designed to replicate the findings of Study 1 with a different task (analogy-task) and to test the hypotheses that expectancies should affect performance only when the task is difficult and when expectancies were formed via intensive processing (high NFC) but not when expectancies were formed via less intensive processing (low NFC). Analogous to Study 1 we manipulated task-difficulty, and assessed NFC of participants (N = 223 university students) and expectancies of success. Depending on the task-difficulty condition, participants worked on an easy or a difficult task. Performance was recorded as the number of correct tasks. The findings supported our hypothesis: First, task-difficulty had an influence on expectancies of success only for participants with higher NFC. Second, significant correlations of performance expectancies and actual performance were only found for participants with higher NFC and when the task was difficult.

The results from the two studies clearly support the idea that factors which influence the type of information processing (like in the present studies cognitive motivation, i.e. NFC) also affect whether performance expectancies depend on task difficulty. The finding from both studies that students with a low NFC seemed to ignore task difficulty when forming their performance expectancies has several implications for research on self-regulated learning. Adaptive self-regulated learning requires accurately analysing the task characteristics (including task difficulty) and to regulate one's own behaviour according to the task demands (e.g., showing higher effort when the task is difficult). As study 2 showed, for students with a low NFC, performance expectancies were unrelated to performance in difficult tasks. Low NFC therefore may be a factor which can impede successful self-regulated learning.