Test Anxiety and Learning Potential in College Students

Nicholas Cale
Christopher Fowler*
Melisa Rempfer*

University of Missouri – Kansas City


This study examined the impact of Test Anxiety (TA) on undergraduate college students' Learning Potential (LP). It was predicted that those students high in TA would perform poorly in comparison to students low in TA on initial testing of the Wisconsin Card Sorting Test (WCST), but the students would show equal testing abilities on the WCST when presented with dynamic testing. Sixty-one students completed the Test Anxiety Inventory (TAI), and were then randomized into high- or low-threat conditions. Twenty-two students completed the State-Trait Anxiety Inventory (STAI), California Verbal Learning Test (CVLT), Letter-Number Sequencing task (LNS), and the dynamic WCST. Multiple analyses of variance (ANOVA) revealed that TA did not significantly impact learning on the WCST.


Test Anxiety

Anxiety is common among students at all levels of education. A specific type of anxiety experienced during testing situations, test anxiety (TA), has been shown to elicit emotional responses that contribute to task interference and lower intellectual testing performance (Sarason, 1961). Test anxiety is also related to fears of negative evaluation, the disliking of tests, and poor study habits (Hembree, 1988). Furthermore, high levels of test anxiety have been shown to be negatively correlated to many factors, such as IQ, academic achievement, memory, and cumulative grade point average (Cassady & Johnson, 2002; Hembree, 1988). In a meta-analysis of 562 studies on test anxiety in academic settings conducted by Hembree, results found that test anxiety is consistent in reducing performance.

Various areas of cognition are also negatively influenced by test anxiety. For instance, worrying is said to use some of the cognitive processing resources, such as attention, awareness, and working memory that could otherwise be applied toward the testing situation (Wine, 1971; Lehto, 1996; Dobson & Markham, 2001). Worry and emotionality are also commonly assumed to lead to learning deficits (Cassady & Johnson, 2002). Specifically, these learning deficits often show up in testing situations. As the demands of a testing situation increase, so does the anxiety (Meijer, 1996). It has also been found that the perception that individuals will be judged on their performance is enough to elicit increases in the anxiety they experience (Sarason, 1961; Meijer, 1996; Foos & Fisher, 1988; Jensen-Campbell, et al., 2002). Therefore, the perception of the classroom or testing environment can influence performance (Lidz & Elliott, 2000; Nelson & Knight, 2010).

This decline in performance can be attributed to many factors that influence test anxiety. College students who worry about tests are more likely to feel less self-esteem, take longer to complete tasks, spend more time studying, have higher state anxiety, and expect lower success (Alansari, 2004; Hembree, 1988; Foos & Fisher, 1988). Test anxiety is experienced at varying degrees, however. Females tend to report higher levels of test anxiety than their male counterparts, yet they perform equally well on cognitive measures (Hembree, 1988).

Still, some students show significantly more test anxiety than others. Studies such as Hancock (2001) and Zatz and Chassin (1985) have split students into groups, such as high test-anxious (HTA) and low test-anxious (LTA), based on their scores on test anxiety measures and then tested them on multiple variables to gauge performance. Students high in test anxiety devote more cognitive resources to task-debilitating thoughts, which leads to poor performance, as shown by Zatz and Chassin (1985).

When threat is brought into the equation, depending on the individual, the anxiety state can be more detrimental or beneficial to performance. Evaluative threat can be viewed as a situational pressure in which individuals assume that their performance will be compared to others (Zatz & Chassin, 1985; Hancock, 2001). Highly anxious students tend to score better without evaluative pressure, whereas students with low anxiety perform better with evaluation (Hembree, 1988). The highly anxious students perceive the threat negatively, worrying about their abilities and performing inefficiently, while the low anxiety students view evaluation as a way to showcase their talents (Cassady & Johnson, 2002). Thus, in order to maximize performance, the goal must be to find a way to reduce the anxiety while in the presence of an evaluative situation so that students can perform at their best despite external pressures.

Learning Potential and Dynamic Assessment

If the purpose of administering a measure is to find out the true capabilities of the participant, assessments should minimize the influence of extraneous factors (Lidz & Elliott, 2000). In this mindset, the concept of learning potential (LP) has been proposed as a method for measuring one's capacity for learning under optimal conditions.

One way to operationalize LP is the level of knowledge that can be achieved after receiving assistance (Meijer, 2001). A broader definition is that LP is simply the ability to acquire and apply cognitive skills (Rempfer, Hamera, Brown, & Bothwell, 2006). Thus, LP is often measured using dynamic assessment (DA) methods, which embed instruction in the evaluation procedure. Meijer (1996) noted that LP assessment is used in many different settings, such as educational testing, serious mental illness (SMI) assessment, and vocational testing. The assumption is that LP evaluations are more predictive of an individual's abilities than traditional tests, due to the greater focus on the learning process itself instead of the end results. Additionally, LP assessment benefits those who are vulnerable to misunderstanding the instructions or expectations of traditional measurements due to unfamiliarity, social disadvantages, or other learning difficulties (Barr & Samuels, 1988; Glutting & McDermott, 1990; Lauchlan & Elliott, 2001).

The challenge then is to create a testing situation in which disadvantaged students can demonstrate their ability to apply skills along with learning new ones (Lidz & Elliott, 2000; Kozulin, 2005; Kozulin, 2010). DA enables the analysis of how differences in individual test performance are affected by modifying the testing conditions (Bethge, Carlson, & Wiedl, 1982). Dynamic assessment measures consist of a pretest-instruction-posttest format, with the idea that individuals will apply the techniques or instructions learned in the second phase during the posttest, providing researchers with a better representation of learning ability. As an illustration of DA, Wiedl (1999) used these methods to examine learning potential in people with schizophrenia.

Wiedl (1999) classified individuals based on their scores on a measure of executive functioning (higher-order cognitive functions include selective attention, planning, and the manipulation of information in problem-solving). He developed a dynamic protocol for the Wisconsin Card Sorting Test (WCST; Kongs, Thompson, Iverson, & Heaton, 2000), a measure of executive function using cards. In this testing format, the WCST is administered in three trials: trial one and trial three are traditional trials, administered under the standard testing protocol; trial two is an instructional trial, where the researcher gives feedback based on the responses of the participant. This example of dynamic assessment highlights the importance of the process, more than the product, of cognition. By supporting examinees during a testing situation, the disadvantages that they encounter, such as anxiety, should be somewhat alleviated, providing a more precise measurement of performance (Meijer, 1996).

One application of LP methods has been to anxious test-takers, who might benefit from the non-traditional testing format of DA. It has been found that subjects have been less anxious after learning potential tests and that these learning tests are less biased against anxious subjects than traditional tests (Meijer, 2001). Measuring the performance of individuals when help can be obtained has been found to be a more valid predictor of future performance for HTA students than for LTA students (Meijer, 1996).

The purpose of the present study was to examine how TA and evaluative threat influence performance on LP assessment at the college level, using DA as an evaluation method. Specifically, the following primary hypotheses were examined:

H1: During an evaluative condition, students reporting higher TA would perform more poorly on the WCST than students reporting lower TA.

H2: Those with HTA would benefit more from the assistance offered during dynamic learning potential assessments than LTA.

Secondary analyses examined the relationships between TA and demographic and academic variables.



Participants were recruited from undergraduate psychology classes at the University of Missouri – Kansas City. All volunteers were given course credit for their participation at the discretion of the class instructor. Testing consisted of two phases. The first phase consisted of group sessions in which the students completed informed consent, a brief demographic questionnaire, and the Test Anxiety Inventory (TAI; Spielberger, 1980). From the initial screening, students with high and low TA were identified and invited to a second session in which they completed a battery of cognitive assessments.

Sixty-one students (68.8% female; mean age of 20.92 years, SD = 3.66), completed the first phase, and 22 (68.1% females; mean age of 21.19, SD = 3.42) completed the second phase. The majority of participants were in their first two years of study (73.8%) with an average estimated GPA of at least a B (83.6%). Psychology (21.3%) and Nursing (18.0%) were the highest reported majors, accounting for 39.3 percent. In terms of race/ethnicity, 62.3 percent self-identified as Caucasian, 13.1 percent as African American, 8.2 percent as Latino, and 16.4 percent as Other.


The UMKC Social Sciences Institutional Review Board (IRB) approved this study, and all participants completed written informed consent before each session. Participants were encouraged to take breaks as needed and were also instructed that they might withdraw from the study at any time.

In Phase 1, 61 participants completed informed consent, a brief demographic questionnaire, and the TAI. Based on their test anxiety scores, 40 students were invited to participate in the second phase of the study. Participants scoring in the top 33 percent and bottom 33 percent on the TAI were included in Phase 2.

In Phase 2, participants were randomly assigned to one of two conditions: a high-threat or low-threat condition. In both conditions, participants completed the same measures, with the only difference being how the testing was described to the participants. In the high-threat condition, the measures were described as IQ tests, and in the low-threat condition, the measures were described as puzzles. Testing for Phase 2 involved informed consent, additional anxiety assessment via the State Trait Anxiety Inventory (STAI; Spielberger, 1983), along with a battery of cognitive assessments including the California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober, 1987), Letter Number Sequencing (Wechsler, 2008), and the dynamic WCST.


Part 1:

Demographic Questionnaire

A researcher-developed demographic/background questionnaire was administered, containing questions concerning sex, ethnicity, age, and educational information (e.g., GPA, best ACT score).

Anxiety Measure

The Test Anxiety Inventory (TAI) was used to assess test anxiety. The TAI is regarded as one of the best tools for measuring test anxiety in high school and college students and includes worry and emotionality subscales (Hancock, 2010). Items on the TAI include questions such as "Thoughts of doing poorly interfere with my concentration on tests" and "During examinations I get so nervous that I forget facts I really know." These items are scored using a 4-point Likert scale ranging from "Almost Never" (1) to "Almost Always" (4). Spielberger (2009) demonstrated excellent internal consistencies (α ≥ .90) for all items.

Part 2:

40 participants were invited to enroll in Phase 2 of the study. Participants were randomized into one of two conditions: a high-evaluative condition (N = 10) and a low-evaluative condition (N = 12). In the high evaluative condition, participants were told that their performance on the measures was related to intelligence and highly correlated with academic success. These procedures were developed based on those by Dobson & Markham (2001). Participants in the low evaluative condition were told the measures were puzzles. The measures were consistently referred to as the corresponding test or puzzle, respectively, throughout the second phase of testing.

Following the testing, all participants were debriefed on the actual nature of the measures, and questions were answered.

Anxiety Measure

In order to assess participant state anxiety at the time of testing, the State-Trait Anxiety Inventory (STAI) was given. This inventory uses self-assessment in order to determine the level of anxiety of the individuals in general, as well as during the time of administration (Spielberger, 1983). Items on the State portion include "I feel calm" and "I am worried." Items on the Trait portion include "I feel like a failure" and "I wish I could be as happy as others seem." These items are scored using a 4-point Likert scale, ranging from "Not At All" (1) to "Very Much So" (4). Spielberger (2009) reported extensive validity evidence for this measure.

Cognitive Measures

The cognitive measures selected to examine LP were the dynamic WCST, the dynamic CVLT, and the LNS.

The Wisconsin Card Sorting Test (Kongs et al., 2000) is a measure of executive functioning. It requires abstract thought, conceptual flexibility, and the ability to monitor one's own behavior and learn from feedback. Participants are instructed to sort, or match, 64 cards that vary according to shape, color, and number of elements they contain to one of four "key" cards. Participants are not explicitly informed of the matching rule; they are informed only whether their response is correct or incorrect. Based on this feedback, participants are to figure out the correct sorting rule. The dynamic WCST was administered three consecutive times (pre-test, training, post-test) to assess the individual's learning. This procedure has been used in previous research to establish "learning potential" (Wiedl, 1999; Rempfer et al., 2006).

The California Verbal Learning Test (Delis, Kramer, Kaplan, & Ober, 1987) is a verbal learning test that measures immediate auditory memory span. After auditory presentation of 16 common words, participants are asked to recall as many of the words as they can remember, in any order. The list is then presented again for a second and third trial. This also facilitates the assessment of "learning potential." As a result, this test can provide information about the participant's verbal memory, conceptual ability, and use of learning strategies.

The Letter-Number Sequencing Test (Wechsler, 2008) is a measure of working memory used on the Wechsler Adult Intelligence Scale-IV. This measure consists of 15 sets of letters and numbers, arranged in a random order (e.g., Y-A-2). The participant is instructed to put the letters in alphabetical order followed by the numbers in numerical order (in this case, 2-A-Y would be the correct response). The test ends when the individual answers three sets in a row incorrectly or completes all sets. This measure shows how well individuals can use their working memory to solve problems (Wechsler, 2008).


It is important to note that the small sample size limited the power available for analysis. In order to examine H1, an analysis of variance (ANOVA) was conducted. The main effect test for anxiety on WCST Trial 1 performance showed that the high anxiety group (M = 51.40, SE = 2.37) did not score significantly higher than the low anxiety group (M = 47.29, SE = 2.19), F(1,20) = 1.84, p > .05, R2 = .09. Likewise, the main effect test for threat on WCST Trial 1 performance indicated that students in the high threat group (M = 51.60, SE = 2.36) were similar to those in the low threat group (M = 47.09, SE = 2.19). These results were not significant F(1,20) = 2.19, p > .05, R2 = .19.

Furthermore, multiple ANOVAs were conducted to examine H2. On average, students in the low threat group performed comparably from Trial 1 to Trial 3 (M = 8.66, SE = 2.39) to students in the high threat group (M = 4.30, SE = 2.58). This indicated that the main effect test for threat on improvement from Trial 1 to Trial 3 was not significant, F(1,20) = 1.96, p > .05, R2 = .08. In similar fashion, the main effect test for anxiety found that students in the high anxiety condition (M = 9.56, SE = 2.39) did not differ significantly from those in the low anxiety condition (M = 3.40, SE = 2.58), F(1,20) = 3.66, p = .07, R2 = .15.

The main effect test for threat on LNS performance was not significant F(1,20) = .08, p > .05, R2 = .004. Students in the low threat condition performed comparably on the LNS (M = 11.36, SE = .62) to those students in the high threat condition (M = 11.10, SE = .67). Likewise, students in the low anxiety condition scored similarly on the LNS (M = 11.66, SE = .62) to students in the high anxiety condition (M = 10.80, SE = .61), signifying that the main effect test for anxiety on LNS performance was not significant F(1,20) = .90, p > .05, R2 = .05.

Finally, the main effect test of anxiety on CVLT performance revealed that students in the low anxiety condition achieved scores (M = 30.51, SE = 1.62) similar to students in the high anxiety condition (M = 27.3, SE = 1.75). Again, this difference was not significant F(1,20) = 1.96, p > .05, R2 = .18. The main effect test of threat on CVLT performance was not significant F(1,20) = 3.86, p > .05, R2 = .06, with students in the high threat condition achieving statistically similar results (M = 31.20, SE = 1.75) to students in the low threat condition (M = 26.61, SE = 1.62).

Secondary analyses indicated that female students reported significantly higher anxiety (M = 69.20, SD = 21.23) than males (M = 49.14, SD = 7.17) on the STAI, t(20) =

-3.28, p < .01, d = 1.27, which is consistent with previous research (Hembree, 1988). Also, females reported significantly higher TA (M = 41.68, SD = 12.13) than males (M = 37.16, SD = 11.08) on the TAI, t(59) = -2.11, p < .05, d = .39.

Pearson correlations were utilized to examine the association between test anxiety and auxiliary factors. Interestingly, test anxiety was significantly correlated to students' reported GPA (r = .468, p < .001), which supports previous findings (Cassady & Johnson, 2002). Also, the TAI worry subscale was negatively correlated with performance on the LNS at a trend level (r = -.40, p = .07).


Summary of Findings

This study examined the effects of students' test anxiety on their cognitive performance. College students reporting high and low test anxiety were exposed to two types of evaluation – high threat and low threat. After the tests were administered, the students' performance levels were measured.

This study found that students in the high evaluative condition did not perform differently than students who were in the low evaluative condition. This indicates that all of the students, despite the evaluative condition, were able to perform at the same level of functioning. An interesting finding was that the worry aspect of the TAI was found to be trending towards having a negative influence on LNS performance. This result suggests that worrying may hinder the working memory capacity, which should be examined in future research. As indicated by Meijer (1996), the higher the demands of a testing situation, the more harmful the effects of anxiety can be to the overall performance of the individual, disturbing concentration and taking up a greater amount of the processing capacity that could be used for other cognitive processes.

Similarly, in these data, students reporting high anxiety did not perform differently than students reporting low anxiety. These findings are also contrary to previous research, which showed that students who report higher anxiety should perform more poorly on cognitive measures than students who report lower anxiety (Hembree, 1988). This demonstrates that all of the students, despite the anxiety reported, were able to perform at the same level of functioning. However, anxiety was trending towards having an effect on students' learning potential, as indicated by the dynamic WCST. This suggests that anxiety could be a factor that influences students' ability to learn and should be examined further in future research.

Further, it is possible that other DA measures could be more valuable for this population than the WCST (Lauchlan & Elliott, 2001), which may prove more representative of learning potential in this population.


The absence of differences between the two experimental groups could be due to multiple factors. Again, the small sample size limited the power available for analysis. This study showed that women report higher anxiety but do not suffer performance impairments, which may be a function of the smaller sample size. Perhaps more male subjects for this sample would provide a better representation of anxiety's influence on WCST performance. Another issue with this study was the possible lack of generalization. This sample consisted of mainly white, female, psychology students, and may not generalize to the greater population.

Future Research

The finding that TA was found to influence working memory at a trend level should lead others to further investigate this relationship with a larger and more diverse sample. Similarly, the observation that anxiety influenced performance on the WCST at a trend level provides a stepping-stone into the investigation of the effectiveness of WCST administration in college students. Thus, future studies should further examine these issues with large sample sizes.

All rights reserved by the Undergraduate Research Community.


Alansari, B. (2004). The relationship between anxiety and cognitive style measured on the stroop test. Social Behavior and Personality, 32(3) pp. 283-294.

Barr, P. M., & Samuels, M. T. (1988). Dynamic assessment of cognitive and affective factors contributing to learning difficulties in adults: a case study approach. Professional Psychology: Research and Practice, 19(1), pp. 6-13.

Bethge, H., Carlson, J. S., & Wiedl, K. H. (1982). The effects of dynamic assessment procedures on Raven matrices performance, visual search behavior, test anxiety and test orientation. Intelligence, 6, pp. 89-97.

Cassady, J. C. & Johnson, R.E. (2002). Cognitive test anxiety and academic performance. Contemporary Educational Psychology, 27, pp. 270-295.

Delis, D.C., Kramer, J.H., Kaplan, E., & Ober, E. (1987). California Verbal Learning Test. San Antonio, TX: Psychological Corporation.

Dobson, M. & Markham, R. (2001). Individual differences in anxiety level and eyewitness memory. The Journal of General Psychology, 119(4), pp. 343-350.

Foos, P .W. & Fisher, R. P. (1988). Using tests as learning opportunities. Journal of Educational Psychology, 80(2), pp. 179-183.

Glutting, J. J. & McDermott, P. A. (1990). Childhood learning potential as an alternative to traditional ability measures. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 2(4), pp. 398-403.

Hancock, D. R. (2001). Effects of test anxiety and evaluative threat on student's achievement and motivation. The Journal of Educational Research, 94(5), pp. 284-290.

Hembree, Ray. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational Research, 58(1), pp. 47-77.

Jensen-Campbell, L. A., Rosselli, M., Workman, K. A., Santisi, M., Rios, J. D., & Bojan, D. (2002). Agreeableness, conscientiousness, and effortful control processes. Journal of Research in Personality, 36, pp. 476-489.

Kongs, S. K., Thompson, L. L., Iverson, G. L., & Heaton, R. K. (2000). Wisconsin Card Sorting Test-64 Card Version Professional Manual. Lutz, FL: Psychological Assessment Resources, Inc.

Kozulin, A. (2005). Learning potential score as a predictor of sensitivity to cognitive intervention. Educational & child Psychology, 22(1), pp. 29-39.

Kozulin, A. (2010). Same cognitive performance, different learning potential: dynamic assessment of young adults with identical cognitive performance. Journal of Cognitive Education and Psychology, 9(3), pp. 273-284.

Lauchlan, F. & Elliott, J. (2001). The psychological assessment of learning potential. British Journal of Educational Psychology, 71, pp. 647-665.

Lehto, J. (1996). Are executive function tests dependent on working memory capacity? The Quarterly Journal of Experimental Psychology, 49A(1), pp. 29-50.

Lidz, C.S. & Elliott, J.G. (Eds.) (2000) Dynamic Assessment: Prevailing Models and Applications. New York; Elsevier Science, Inc.

Meijer, J. (1996). Learning potential, personality characteristics, and test performance. In Hamers, J. H. M., Sijtsma, K. & Ruijssenaars, A. J. J. M. (Eds.) Learning Potential Assessment: Theoretical, Methodological, and Practical Issues. (pp. 341 – 360). Steenwijk, Netherlands: Grafisch Produktiebedrijf Gorter.

Meijer, Joost. (2001). Learning potential and anxious tendency: test anxiety as a bias factor in educational testing. Anxiety, Stress, and Coping, 14, pp. 337-362.

Murphy, M. (2009). Everyday executive function in people with schizophrenia: Investigating the dysexecutive questionnaire. Master's thesis, University of Missouri – Kansas City, 2009.

Nelson, D. W. & Knight, A. E. (2010). The power of positive recollections: Reducing test anxiety and enhancing college student efficacy and performance. Journal of Applied Social Psychology, 40(3), pp. 732-745.

Rempfer, M., Hamera, E., Brown, C., & Bothwell, R. (2006). Learning proficiency on the Wisconsin Card Sorting Test in people with serious mental illness; What are the cognitive characteristics of good learners? Schizophrenia Research, 87, 316-322.

Sarason, I. G. (1961). Test anxiety and the intellectual performance of college students. Journal of Educational Psychology, 52(4), pp. 201-206.

Spielberger, C. D. (1980). Preliminary professional manual for the Test Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.

Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory: STAI (Form Y). Palo Alto, CA: Consulting PsychologistsPress.

Spielberger, C. D. (2009). State-Trait Anxiety Inventory. [In] I. B. Weiner & W. E. Craighead (Ed.), The Corsini Encyclopedia of Psychology. John Wiley & Sons, Inc.

Spielberger, C. D. (2009). Test Anxiety Inventory. [In] I. B. Weiner & W. E. Craighead (Eds.), The Corsini Encyclopedia of Psychology. John Wiley & Sons, Inc.

Wechsler, D. (2008). WAIS-IV administration andscoring manual. San Antonio, TX: The Psychological Corporation.

Wiedl, K. H. (1999). Cognitive modifiability as a measure of readiness for rehabilitation. Psychiatric Services, 50(11), pp. 1411-1419.

Wine, J. (1971). Test anxiety and direction of attention. Psychological Bulletin, 76(2), pp. 92-104.

Zatz, S. & Chassin, L. (1985). Cognitions of test-anxious children under naturalistic test taking conditions. Journal of Consulting and Clinical Psychology, 53(3) pp. 393-401.


©2002-2021 All rights reserved by the Undergraduate Research Community.

Research Journal: Vol. 1 Vol. 2 Vol. 3 Vol. 4 Vol. 5 Vol. 6 Vol. 7 Vol. 8 Vol. 9 Vol. 10 Vol. 11 Vol. 12 Vol. 13 Vol. 14 Vol. 15
High School Edition

Call for Papers ¦ URC Home ¦ Kappa Omicron Nu

KONbutton K O N KONbutton