Driscoll, R., B. Holt, & L. Hunter (2005). Accelerated Desensitization and Adaptive Attitudes Interventions and Test Gains with Academic Probation Students. Education Resources Information Center, 10, 13pp.  
Click above for published article.

revised July, 2006

Test-Anxiety Reduction Benefits
with Students on Academic Probation

by

Richard Driscoll, Ph.D., Westside Psychology, Knoxville Tn
Bruce Holt, M.S., Maryville College
Lori Hunter, M.S., Maryville College

*The anxiety reduction protocol used here is recorded on CD and is available to interested researchers by request.

**Correspondence should be sent to:

Richard Driscoll, Ph.D.
301 Gallaher View Road, Suite 102, Knoxville TN 37919

Phone 865 690-0962 x104

Fax 865 690-0995
email: drD (capital D) at amtaa dot com

*** The authors are grateful to Tuncay Ergene, Ph.D., for his assistance.

Test-Anxiety Reduction Test Benefits
with Students on Academic Probation

Abstract

While anxiety impairments contribute to poor performance, failure and academic probation would be expected to significantly increase anxiety. Twenty college students, half on academic probation, were assigned to an Intervention or to a minimal treatment Control group. The Intervention was a counterconditioning protocol which included physical exertion and positive adaptive associations to eight learning, review, and testing scenes.

The Intervention group attained a significant test gain over the Controls, with considerably stronger gains among academic probation students and more modest gains among students in good standing. Methodological limitations were discussed which warrant caution in interpreting the findings.

As many as 20% of college students are handicapped by high test anxiety, according to various surveys (Hill & Wigfield, 1984), and an additional 10–15% may be somewhat handicapped by "moderately high" anxiety. Anxieties do appear to be increasing in step with the increased national emphasis on testing (McDonald 2001). High test anxiety reduces working memory and impairs concentration and reasoning, and highly anxious students scores about 12 percentile points below their low anxiety peers (Hembree, 1988; Cassady & Johnson 2001; McDonald 2001). High anxiety also reduces ability to comprehend and retain material (Tobias, 1980). Left untreated, performance anxieties continue into adulthood where they restrict career choices and lower quality of life (Topp, 1989; Krohne & Laux, 1982).

Test anxiety reduction interventions have been found to lower anxiety and to improve test performance by as much as 12 percentile points (Hembree, 1988), although current research outcomes are considerably more modest (cf. Spielberger & Vagg, 1995). In the absence of a recent overview, a reasonable estimate of test score gains is not currently available.

Treating the 20 percent of students who are test anxious would appear to be an overwhelming task. The better interventions take between 3.5 and 8 hours, according to a recent meta-analysis (Ergene, 2003), while those taking an hour or less are much less effective.

Hancock (2001) concludes that students with high test anxiety show significantly less motivation in classrooms perceived as highly evaluative compared to students with low test anxiety. In general, high test anxiety is more closely associated with lowered performance in low-ability students than in their high-ability counterparts (Hembree, 1988, p.65).

Many colleges have academic probation programs which attempt to improve performance for their students with unsatisfactory grades. The above findings are clearly pertinent to such programs. Students on academic probation might be expected to include at least an average proportion of highly test-anxious students, and perhaps more, as some identifiable handicap is apt to be contributing to the poor grades. Furthermore, academic probation is itself extremely threatening, as it portends failing out of college if grades are not sufficiently improved. To a student invested in a higher status job or a professional career, it would be hard to imagine more personally intimidating circumstances.

Further, highly test-anxious probation students are apt to be more concerned about their plight than are their low-anxiety counterparts. Low-anxiety college students with poor performance may be simply uninterested in their schooling and unconcerned, or perhaps inadequately prepared. In contrast, those with high anxiety and poor performance would be expected to be concerned and perhaps overly concerned about school. Remove the impairment, and these students may show both the motivation and the ability to perform significantly better.

Many of us have seen a friend or school mate who has failed a critical qualifying exam or licensure exam, and is then too anxious and too unfocused to prepare well for the repeat exam. Highly anxious students on academic probation may be in something of the same predicament.

So far as the above observations are correct, an effective test-anxiety reduction program should show comparatively stronger benefits for those anxious students on academic probation. The aim of the current investigation is to assess the benefits of a test anxiety reduction protocol on test-anxious academic probation students, in comparison to test-anxious students in good standing.

Method

As a compromise between the college's concern to benefit its high-anxiety students and the research requirement for a credible control group, more students were assigned to the Intervention than to the Control group. Twelve students were in the final Intervention group, while eight were in the Control group. (Three originally assigned to the Intervention did not review the protocol and were dropped from the study.)

Student assignment was not completely random. The director of the learning center assigned several students to Intervention who were seeking assistance but appeared to be not benefitting from it. Students assigned to Intervention had the option to participate, and about half declined, whereas students assigned to Control were asked to review anxiety reduction suggestions but not rated on whether they did so. Thus, the two groups were not completely matched on initial assignment or on completion of the program after assignment.

The final Intervention and Control groups were roughly equivalent in initial anxiety scores, initial grades, and in the numbers of seminar vs. probation students in each group. The initial anxiety scores ranged from 3.0–4.4, where 3.0–3.4 indicates moderately high anxiety and 3.5–5.0 is high to extremely high anxiety. The average initial anxiety for the Intervention and Control groups was 3.88 and 3.60 respectively, indicating a modest match with the Intervention group slightly more anxious.

The Intervention and Control groups had initial mean test grades of 77.1 and 78.3 respectively, indicating a close match between the two groups. The college uses a standard scale with 10 points per letter grade, where 75 is a middle "C."

Students imagine an activity that they find especially interesting, and re-experience the sense of interest. Then, by suggestion, students imagine being interested in a series of eight learning, review, and testing scenes. Interest in school subjects contributes strongly to school performance, so an interested attitude is a plausible antidote to a fearful attitude.

Additional stretch–relax sequences are included between scenes, to curb any anxiety created and to reinstate the calmness and prepare students for the next scenes. The vivid experience of release and safety at the conclusion of a stressful scene is expected to condition participants to perceive the situation as safe and no longer frightening.

Students in the Intervention group were given the protocol on recorded CD, along with instructions, and asked to listen to the CD on their own at least twice. Twelve reviewed the CD, most of them twice, usually in their dorm rooms. Thus, the Intervention consisted of instructions and then listening to an average of two 30 minute recorded sessions.

While various anxious students may do well with one or two intervention components (Hiebert, 2000), the composite approach used here addresses both the cognitive and emotional aspects of anxiety and should be expected to benefit a broader range of individuals (Sapp, 1996).

An early component analysis indicates that exposure, strenuous activity, and positive images can all contribute significantly to overall anxiety-reduction benefits (Driscoll, 1976).

The Controls received a minimal treatment package consisting of information from the popular University of Illinois test anxiety management website (1996/2004). Students were asked to review it twice.

The use of stretch–tense, deep breath, release–relax sequences to curb anxiety, before and after exposure scenes, may provide a promising solution. The stretching and muscle tension are thought to consume the physiological components of arousal, and the deep breaths counter the shallow breathing often associated with high anxiety. Tensed muscles fatigue quickly, and relaxation follows naturally as students release their air and release their muscles, and experience their muscles relaxing.

Anxiety tolerances are believed to be not so critical here, as the stretch–tense segments are thought to counter even strong surges of anxiety and cannot be overwhelmed or inhibited by such surges. Recorded versions of the protocol have been found to produce strong anxiety-reduction results across several samples that compare favorably to professionally-administered protocols (see Driscoll, 2005, for protocol wording and anxiety-reduction findings).

Results

Several of the Control students reported that the information was helpful, but added that they had seen most of it before. The Control package thus seems to have served its purpose as a minimal treatment control, maintaining morale without introducing a competitive intervention.

Figure 1

The treatment effect size was .75 sd, calculated as net gain divided by the standard deviation of pre-treatment grades (sd=8.4). The effect size here compares favorably to the .5 sd average effect size for the test anxiety interventions summarized by Hembree (1988), suggesting that the present protocol holds its own against professionally-administered protocols. The .75 sd gain corresponds to an 18 percentile net gain in class rank for the treated students over the controls.

Table 1

Thus, the protocol produced a major full-grade gain for the lower-performing students, but a more modest quarter-grade gain for the higher-performing students. The Intervention gains for the probation students were higher statistically than those for the regular status students (t = 4.34, df = 10, p<.001). The difference was .89 sd, or 21.5 percentile points.

In a related analysis, initial grades correlated r = –.80 with grade gains in the Intervention group (t = 4.23, df = 10, p < .002). For the Control group, initial grades appear unrelated to grade gains (r = .10). These figures show again that the Intervention improved grades for lower-performing students much more than for higher-performing students.

Given the especially strong results from the struggling students, we would expect clear benefits among those on academic probation. Overall, 78 second semester freshmen were on academic probation (out of 340 freshmen). Of the five in the Intervention group, four regained regular standing and one remained on probation. Of the remaining 73 not treated, 30% regained regular standing at the end of the semester, while 43% continued on probation, 20% were suspended, and the remaining 7% withdrew. While the sample size is small, the 80% of Intervention students who regained regular status compares favorably with the 30% average for the untreated probation students.

Discussion

Because of these methodological limitations, the results must be interpreted with some caution. While Treated students did improve over untreated Controls, the initial groups were not sufficiently matched. Thus, the findings might be considered as suggesting rather than solidly confirming test gains.

We might also expect that high-anxiety students who perform well on tests have found ways to manage and compensate for their high anxieties, yielding rather modest performance impairments. Further, their high anxieties may be causing them to overestimate their actual impairments. So, while anxiety-reduction training makes these students feel more confident in tests, reducing their relatively modest impairments provides only modest test score gains.

Among underperforming students who are highly motivated, in contrast, high test anxiety appears to be a primary and critical impairment. Remove what is perhaps the major obstacle, or reduce it, and the grades improve dramatically.

Compared to the gain among the low-performing students, the gain for the regular students appears relatively modest. And yet, if it withstands replication, a quarter grade improvement among already satisfactory students is respectable and worthwhile.

These findings do suggest that test anxiety reduction is especially beneficial for high-anxiety students performing substantially below their apparent potentials. Screening and intervention among academic probation students could significantly improve academic outcomes.

We can also make recommendations, based on high vs. low anxiety and poor vs adequate school performance (see Table 2).

Table 2

Low-performing students who are highly anxious will gain most from anxiety reduction. High-performing students with high anxiety will be more comfortable after anxiety reduction, and should show modest grade gains.

It is reasonable for colleges to teach not just information and reasoning, but also the ability to think clearly and perform well amid the competitive pressures that follow us throughout our lives. Test anxiety reduction may have an important place in our educational institutions.

References: