Keywords: test anxiety, test taking, education, college students
*Please address correspondence to faculty advisor:
Devereaux A. Poling
Ohio University - Zanesville
1425 Newark Rd
Zanesville, OH 43701
Fax: (740) 588-1436
Examinations represent one of the most common forms of performance evaluation for college students and a major source of anxiety for many. This study examined the impact of student anxiety toward test taking, study behavior, and student beliefs about whether an upcoming test was going to be easy or hard. Fifty-seven participants completed a multiple-choice exam in either a simulated classroom context or a solitary testing context. Participants were also forewarned that the exam was going to be easy or that it was going to be hard. In contrast to previous research, beliefs about test difficulty did not influence test performance. However, testing context and test anxiety did influence student exam performance.
Easy Test or Hard Test, Does it Matter? The Impact of Perceived Test Difficulty on Study Time and Test Anxiety
Imagine you were sitting, anxiously waiting for an exam to begin, what would you do? Would you use your extra time before the exam to study? How would you study? What techniques would you use? Would your study habits change depending on the setting you were in: sitting in a classroom with other students vs. sitting alone? Would it matter if you were told the test would be easy or hard?
Foos (1992) posed some of these questions in a study where participants were told that an exam they were about to take would be either easy or hard. All participants were given fifteen minutes to study. Participants who thought the test would be hard performed better than participants who thought the test would be easy. Foos (1992) concluded that participants who thought the exam would be difficult were more motivated to study and work harder than students who thought the test would be easy.
In addition to student anticipation that a test might be hard, test anxiety has also been shown to have an impact on test performance (Seipp, 1991). In fact, Hong (1999) found that anxiety over taking a test mediated the relationship between perceived test difficulty and performance. That is, students who believed that a test would be hard experienced more worry and as a result, performed worse. These results appear to be inconsistent with the Foos (1992) study. However, test takers in Hong’s (1999) study were students enrolled in a statistics course and were taking an actual exam for the course. It is possible that student anxiety in this study was related to the amount of preparation for the exam, which may also have been related to exam performance. For example, a student who had not studied for the test might have high anxiety and a low performance due to lack of preparation.
To further address the complex relationship between anxiety and perceived difficulty on test performance, Weber and Bizer (2006) proposed that the amount and type of anxiety experienced by test-takers may be beneficial in some cases and detrimental in others, depending on expectations that the test would be easy or hard. In Weber and Bizer’s study, participants’ anxiety levels were assessed and then participants were given an exam with instructions indicating that the test they were about to take would be either easy or hard. Test performance interacted with anxiety levels such that low levels of anxiety facilitated better performance when participants believed the test would be hard and high anxiety resulted in poorer performance in the hard test condition. The study showed that anticipated test difficulty did not have a consistent relationship to test performance. Weber and Bizer’s study also left open the question of whether students are more likely to study harder for a test they believe will be difficult as compared to a test they believe will be easy.
Our purpose in the current study was to address the question of whether anticipated test difficulty actually results in longer study time and whether longer study time underlies higher test scores. In addition, we were interested in the role of test anxiety on both study time and test scores. Finally, because it was not made clear in all of the previous research whether participants had completed their exams individually (i.e., in a laboratory setting) or collectively, in a simulated classroom environment, we manipulated the testing context to determine if the presence of peers might also impact test performance.
Seventy-six undergraduate psychology students participated, but only 57 were used in analyses because the remaining 19 participants reported knowing the details of the study prior to participation or because they failed to read the testing instructions informing them of the ease or difficulty of the test. Consistent with the psychology student population of the Midwestern regional campus where this study was conducted, the majority of participants were female (n = 45) and White/Caucasian (n = 51). Participants ranged in age from18 to 50 years (mean age = 26.74 yrs., SD = 9.04 yrs). Forty-four participants ranked as freshmen or sophomores.
Test anxiety was measured with a variety of other test-taking constructs using selected subscales from the Test Attitude Survey (TAS) (Arvey, Strickland, Drauden, & Martin, 1990). TAS items were rated on a 1-7 likert scale. The selected subscales used in this study are listed below with example items from each subscale.
Test anxiety: I usually get very anxious about taking tests.
Motivation: I wanted to do well on this test.
Test preparation: I prepared a lot for this test.
Need for achievement: I try to do well in everything I undertake.
Lack of concentration: It was hard to keep my mind on this test.
Test ease: This test was too easy for me.
External attribution: I have not been feeling well lately and this affected my performance on the test.
In addition to the TAS, participants completed selected subscales from the Motivated Strategies for Learning Questionnaire (MSLQ) (Pintrich, Smith, Garcia, & McKeachie, 1993). However, items from the MSLQ are not reported here because they were not significantly related to study time, test anxiety, or test score (ps > .05).
We used a procedure similar to the one used by Weber and Bizer (2006). Participants were randomly assigned to one of the test instruction conditions, easy or hard. In the easy condition, participants were instructed that they would be taking a multiple-choice test modeled after the easiest questions of the GRE subject test in Psychology. In the hard condition, participants were instructed that they would be taking a multiple-choice test modeled after the hardest questions of the GRE subject test in Psychology. Unlike Weber and Bizer (2006), all participants in this study were given the option to study for up to 30 minutes and were provided with Introductory Psychology textbooks.
Participants’ study time ended when they felt ready to take the test, or when 30 minutes had passed, whichever came first. When the study time was over, participants were asked to rate their confidence level regarding the test using a 1-7 likert scale rating, where 1 = not at all confident and 7 = extremely confident. Participants then began the test. The same test was used in both the easy and hard conditions. It consisted of a 20 question multiple-choice test covering basic perspectives, methods, and history of psychology.
Participants completed the procedure in one of two testing contexts: a solitary (individual) testing context or a simulated classroom (group) test context. This was done to evaluate the effect of peer presence on test taking. The classroom context is a close simulation to a real college testing situation but it also creates a distraction (peers) that is not present in a solitary testing context. Therefore, we speculated that the testing context might influence either study time or test score, or both. Previous research has not addressed this issue.
Study Time: A 2 (context: individual context, group context) X 2 (condition: easy, hard) between subjects analysis of variance on study time revealed a main effect for context, F(1, 53) = 51.86, p < .0001 and a context X condition interaction, F(1, 56) = 7.15, p < .01. Students were more likely to study when placed in a group context as compared to an individual context. In an individual context, students did not differ on study time regardless of whether they were told that the test was easy or hard.
However, in a group context, students were more likely to use a longer period of time for study when told that the test was going to be hard as compared to being told that the test was going to be easy. See figure 1 for means and standard errors.
Study time (in minutes) for test-takers in individual and group contexts
Test Score: A 2 (context: individual context, group context) X 2 (condition: easy, hard) between subjects analysis of variance on test score revealed a main effect for context, F(1, 53) = 5.11, p < .05. When taking a test in an individual context, students performed better than when taking a test in a group context. Mean for individual context = 12.00 (SD = 2.94); mean for group context = 10.21 (SD = 2.51).
Test Anxiety. We first used Pearson correlation coefficients to identify relationships between test anxiety scores on the TAS and other measures. Anxiety was significantly negatively related to ratings of test ease, r = -.34, p < .01 and self-confidence, r = -48, p < .001. Thus, participants who scored high on test anxiety also felt less confident before starting the test and felt as if the test was harder, regardless of whether they were told that the test would be easy or hard. In fact, an independent samples t-test on anxiety scores revealed no significant difference for easy vs. hard condition (p = .9).
Next, test anxiety scores from the TAS were separated using a median split (median = 40; range = 12-64) into high (n = 28) and low (n = 29) anxiety categories. An independent samples t-test revealed that students with high levels of anxiety performed significantly worse than students with low levels of anxiety, t (55) = 3.76, p < .0001. Mean test score for high anxiety = 10.17 (SD = 2.52); Mean test score for low anxiety = 12.78 (SD = 2.72). Might this difference explain the lower test scores in the group vs. individual context? An independent t-test on anxiety scores revealed no significant difference between the group and individual context (p = .8), suggesting that anxiety was not responsible for that effect.
In this study we found that participants who were in a classroom simulated group context studied longer than participants in a solitary testing group, especially when they anticipated that the test would be hard. However, participants who were in the classroom simulated group performed worse on the test than students who were in a solitary testing group. Whereas participants who reported high levels of anxiety, as measured by the TAS, performed worse on the test than students who reported lower levels of anxiety, it does not appear that anxiety was the mediating factor in the poorer performance of those in the group context. That is, there were no differences in anxiety scores for participants in the group vs. individual context.
At first, it seems paradoxical that students in the group context would study longer but perform worse on the test. However, this is not entirely surprising. The presence of peers in the context of the classroom likely created enough social influence that participants felt compelled to utilize whatever study time was made available to them. It seems participants in this study engaged in a form of public conformity (Brehm, Kassin & Fein, 2005) in which they modeled their behavior after what they observed their classmates doing, but they may not have possessed the internal motivation to study in a quality fashion.
The lower test scores obtained by the participants in the group context might be explained by poor quality study tactics combined with the arousal caused by having others present in the room during the testing. This effect, known as social facilitation is well-established in social psychological literature (Zajonc, 1965). If participants in the group context were novices in psychology, then social facilitation would predict that their performance would be hindered by the presence of others. It is reasonable to assume that they were novices in psychology because they were recruited from General Psychology (PSY 101) during the first week of the term.
Unlike previous research, this study did not replicate previous studies that have found anticipated difficulty results in higher test scores (e.g., Foos, 1992; Weber & Bizer, 2006). However, this study did replicate findings with test anxiety. Namely, higher anxiety results in lower test scores (Hong, 1999). In this study, test-takers did not show higher levels of anxiety as a result of the context (group vs. individual) or as a result of being told that the test was easy or hard. Given that anxiety levels did not depend on specific test-related factors, it may be reasonable to speculate that the type of anxiety measured here was trait anxiety, as opposed to state anxiety. In a testing context, trait anxiety is a general concern with all examination situations whereas state anxiety is a level of worry directed toward a specific exam or exam context (Hong & Karstensson, 2002). As measured by the TAS, test anxiety reflects a more general construct consistent with trait anxiety. This is consistent with Weber and Bizer’s (2006) research in that they found effects with trait anxiety, but not state anxiety.
Future research should address some of the limitations of this study. First, a larger and more diverse sample would have been desirable and would be useful in order to draw more concrete conclusions. Also, future research should measure the quality of studying, rather than simply the quantity. Recent research has highlighted the fact that there are some study techniques that are more effective than others (Gurung, 2005). Furthermore, this research has shown that the most effective study techniques are usually the ones least utilized. It would have been beneficial to know what type of strategies our participants were using in this study. That could have enabled a more comprehensive explanation as to why students who studied nearly the entire allowed time did not perform as well on the exam as their peers who studied for shorter periods of time.
In sum, the current research compliments and extends existing literature on testing and test anxiety. This is a pertinent area of research given the increasing pressure on students to perform well on tests. Graduation from high school, extending clear up to entrance into an undergraduate program, and entrance to graduate school can all be determined by a student’s performance on tests. This study and others like it can help students to implement, and professors to advise, better test-taking techniques. Professors should tell their students that quality of study time is more important than quantity, and that obtaining some level of expertise is important before coming into the classroom context for the exam. Furthermore, professors could advise students on available strategies for dealing with test anxiety (Sapp, 1996)
Arvey, R. D., Strickland, W., Drauden, G., & Martin, C. (1990). Motivational components of test-taking. Personnel Psychology, 43, 695-716.
Brehm, S. S., Kassin, S., & Fein, S. (2005). Social Psychology, 6 th ed. Houghton Mifflin: New York.
Foos, P. W. (1992). Test performance as a function of expected form and difficulty. Journal of Experimental Education, 60(3), 205-211.
Gurung, R. A. R. (2005). How do students really study (and does it matter). Teaching of Pschology,32(4), 239-241.
Hong, E. (1999). Test anxiety, perceived test difficulty, and test performance: temporal patterns of their effects. Learning and Individual Differences,11(4), 431-448.
Hong, E., & Karstensson, L. (2002). Antecedents of state test anxiety. Contemporary Educational Psychology , 27, 348–367.
Pintrich, P. R., Smith, D. A. F., Garcia, T., & McKeachie, W. J. (1993). Reliability and predictive validity of the motivated strategies for learning questionnaire (MSLQ). Educational and Psychological Measurement, 53, 801-813.
Sapp, M. (1996). Three treatments for reducing the worry and emotionality components of test anxiety with undergraduate and graduate college students: Cognitive-behavioral hypnosis, relaxation therapy, and supportive counseling. Journal of College Student Development, 37, 79-87.
Seipp, B. (1991). Anxiety and academic performance: A meta-analysis of findings. Anxiety Research, 4, 27-41.
Weber, C. J., & Bizer, G. Y. (2006). The effects of immediate forewarning of test difficulty on test performance. Journal of General Psychology, 133(3), 277-285.
Zajonc, R. B. (1965). Social facilitation. Science, 149, 269-274.