Reducing social work students' statistics anxiety.

Abstract

This paper focused on how various teaching methods reduced anxiety in social work students (N=13) taking a statistics course. The methods included SPSS assignments, student presentations, in-class demonstrations, and readings. In general, students believed that their anxiety was reduced. Conclusions drawn focused on how the teaching methods might have reduced anxiety.

Reducing Social Work Students' Statistics Anxiety

The need to have a baseline understanding of statistics is extremely important for the social work professional. Comprehension of statistical methods will give the student the opportunity to understand previous and current research, generate their own research, evaluate practice, and draw informed conclusions about their clients. However, if a group of social work professors was sitting around the coffee table and the discussion turned to teaching statistics, there would probably be universal agreement that students have high levels of anxiety about statistics based courses. Unfortunately, only a small and inconsistent body of research exists that demonstrates the belief that statistics courses provoke anxiety in social work majors, thus making research into the anxiety of the students an important topic of study.

Research has demonstrated that undergraduate social work majors report higher levels of math anxiety than other social science majors (Royse and Rompf, 1992). Similarly, in a sample of 132 undergraduate social work students, Gustavsson and MacEachron (2001) report that even prior to enrolling in a statistics course, a moderate level of mathematics anxiety existed. Davis (2003) found that 38 out of 41 African-American females enrolled in a graduate social work statistics course demonstrated moderate to high levels of anxiety about the subject matter at the beginning of the course.

Given the aforementioned research, and the idea that social work students are prone to anxiety, social work professors are often challenged with ways to deliver the content while reducing the student's anxiety of learning statistical methods. An example comes from Forte (1995), who describes a social science statistics course that was designed to reduce anxiety in the student. Forte used computer assisted instruction in the form of teaching about statistical software packages, homework assignments based on these lessons, assignments where students practiced speaking in statistical lingo, cooperative learning exercises where students worked in small groups to further understand the concepts, and humor about statistics. In a survey of 93 undergraduate students (18% of them social work students) Forte (1995) found that anxiety decreased from the beginning to the end of the semester. However, there were no numerical offerings of the claimed decrease. Another example comes from Davis (2003), who used several instructional methods with master level social work students in hope of producing a decrease in anxiety from the beginning of the semester. In her class she engaged the students in dialogues about their anxiety and fears, had students work in groups to conceptualize, implement, and analyze social work data, and write and present on the data. Davis's course model focused on helping students gain an in-depth knowledge of statistics using both oral and written skills. A statistical analysis of the pre and post-test statistical anxiety data demonstrated that there was a significant statistical decrease in the number of students who reported a high level of anxiety (n=27) at pre-test to 10 students at posttest.

The fact that there is a paucity of research examining the relationship between social work majors warrants further examination. In fact, the literature lacks a discussion on which methods students and instructors felt lowered anxiety. This report explores that topic. During a summer semester at a university located in New England, the author utilized several instructional methods, and data were collected to see if student anxiety about statistics changed.

Methods

A convenience sample consisted of thirteen students (N=I3) enrolled in a quantitative methods course during a summer semester at a university located in New England. All thirteen students were females and were taking the course to make up for an undergraduate deficiency in statistics needed to apply for a masters degree in social work. The mean age of the group was 38. All students took the pretest (N=13), and all but one (N=12) took the posttest.

A course called Quantitative Methods in Social Work was offered as a prerequisite for admittance into a Master's level social work program. The course was offered over 5 weeks during the summer of 2005 and met twice a week for 3 hours and 45 minutes. The objectives focused on knowledge, values, and skills. An example of a knowledge objective was to understand the use of statistics in social work journals. An example of a values objective was to understand the ethical issues involved in the use of measurement based upon the normal distribution and use of standardized tests. Finally, an example of a skill objective was to prepare and present statistical data in clear and concise ways to promote understanding and reasonable, supportable interpretations. The course focused on frequencies, measures of central tendencies and dispersion, chi-square and crosstabs, correlations, t-tests, ANOVAs, and regression. In addition, lectures focused on sampling and interpreting significance as it relates to statistical inquiry. The course also focused on building students skills in the use of statistical analysis via computers, understanding ethical concerns when applying statistical analysis to diverse populations, and presenting statistical findings both orally and in writing.

There were several methods used for instruction. These included hands-on instruction with laptops, student presentations, in-class demonstrations, and textbook readings and handouts. The methods were designed to reinforce learning about statistics, provide an opportunity for students to discuss their understanding of the application of statistics, and a chance to present findings to a larger group. Ultimately, a combination of the methods, as described below, would offer a chance for the student to become more comfortable with statistics, thus hoping in a reducing their anxiety.

SPSS Assignments: Two computer homework assignments were required in the class. The first assignment asked students to demonstrate their ability in creating datasets and calculating measures of central tendencies and dispersions. Students were given a dataset with numerous variables and asked to enter them into SPSS a social science statistical package. For each variable, they were asked to create a variable name, label and provide an appropriate coding scheme for each. In addition, they were asked to generate frequency tables and the mean, mode, median, and standard deviation. The second assignment focused on bivariate analysis. Like the first assignment, students were given a dataset that they were required to enter into SPSS. Students were asked to run t-tests, crosstabs, and correlations where appropriate. Prior to the assignment of the homework, class time was used to provide instruction via laptop computers in SPSS. The use of SPSS was taught in two sessions. One session focused on getting to know SPSS, while the other focused on bivariate analysis.

Student Presentations." Two presentations were required in class. First, pairs of students were required to present on statistical analysis found in assigned journal articles. Journal articles were assigned based on their statistics and where that particular statistic fell in the syllabus. For example, during the week spent on measures of central tendencies, an article by Steven Jay Gould where he discussed the use of the median to help understand his diagnosis of cancer was read. Second, groups were required to present on the data they collected during the research project. The groups used any statistic discussed in class. For example, one group used frequencies to discuss the number of adoptions by county, while another group used a paired sample t-test to discuss improvement for a social service agency's employees. For both types of presentations, students worked together independent of class time.

In-class Demonstrations: The instructor used real data to present the various statistics discussed in class. One example asked the students to name as many United States as possible in a 5-minute period. The class then worked together to develop a frequency table. Another exam took quiz scores from three mutually exclusive groups and demonstrated an ANOVA. A third example used pre and posttest quiz scores to demonstrate a paired sample t-test. A fourth example used correlation to demonstrate the relationship between attitudes about on-line learning and time spent learning content on-line.

Readings: Handouts provided content for reading. For example, several popular media articles were used as class handouts. These articles all involved mention of statistical analysis. One example included a discussion on Hillary Clinton's possible run for the Presidency of the United States. The cnn.com article used frequency percentages to discuss Senator Clinton's popularity and what the American public thought about her chances for becoming the first female President. Another example came from Time Magazine. The piece summarized a study about the dangers of driving while talking on a cell phone. The author cited statistics from the British Medical Journal. The journal article was also located and distributed to the class. The article used regression analysis to examine the relationship between cell phone use and car accidents. A third article used z-scores to discuss children's height and weights in relationship to nutrition studies in Angola. In addition to the readings provided in class, students, each week, were required to read a social work article that used that statistics discussed during the class session. For example, one article used multiple regression to analyze the predictors of burnout in the mental health fields. Another article used pair sampled t-test to examine the success of solution focus brief therapy had on depression.

The Statistical Anxiety Rating Scale (STARS) measured anxiety (Cruise, Cash, and Bolton, 1985). The instructions called for the students to rate the 51 items on the STARS on a 5-point Likert scale with 1 = "strongly disagree"; 2 = "disagree"; 3 = "neutral"; 4 = "agree" and 5 = "strongly agree". The 51 items summed produced a possible spread of 51 to 255 with lower scores indicating less anxiety and higher scores indicating a higher anxiety. The purpose of the scale was designed to measure general anxiety about statistics in the areas of worth, interpretation, test, self-concept, fear of asking for help, and fear of statistics instructors. The survey was administered both at the beginning of the semester and at the end. The pre test had a Cronbach's alpha coefficient of .91, while the posttest's Cronbach's alpha coefficient was .96. The alphas demonstrated strong internal consistency for the measure. Additionally, students rated the hands-on instruction with laptops, student presentations, in class demonstrations, and readings as helpful, not helpful, or no opinion when it came to reducing their anxiety. In addition, the students expanded on what teaching methods helped reduce anxiety. The instructor for the purpose of the investigation into the teaching methods designed the questions.

Results

Collecting data occurred by employing a one-group only pre-post test design. The STARS, at the beginning of the semester, produced a mean score of 159 and a median of 154. The pretest mean and median scores indicated somewhat of a neutral level of anxiety. However, the spread of the scores went from 119 to 192. The posttest data indicated a slightly lower level of anxiety with a mean of 145.25, a median of 140.5 and a spread of 118 to 182. Overall, there was a reduction in the means by 13.75 points. A paired sample (dependent) t-test indicated that the change was significant (t= 2.287, DF=11, p=.04). An effect size measure of .94 was calculated. The effect size indicated that that the change from pre to post test was of great magnitude thus substantial (Kline, 2004).

Students indicated that for the most part, the majority of the teaching methods were helpful in reducing their anxiety. For example, all the students (N=12), found the lessons on excel and bivariate analysis, the in class demonstrations, reading article, the final project and the handouts as helpful in reducing anxiety. All but one (n=11) found that presenting the articles and the final project to the larger group, the textbook, and the SPSS write-up as helpful in reducing anxiety. The open-ended responses indicated that summarizing and explaining journal articles particularly helpful. For example, one student said, "I found the journal article presentation very helpful because I was forced to "self-learn" a bit about the stat-test in order to explain it in my own words to the class". In addition, one student thought that the journal articles increased her confidence in learning the content. "I think the reaction papers really helped my confidence and helped me to better understand all of it". In addition, students commented on the fact that they were required to present on the journal articles and do a group presentation. One student stated that "I think the article and group presentations were very helpful, especially since we were able to synthesize and apply everything that we learned". Another area that was particularly helpful was the SPSS assignment. For example, one student said, "I think the two SPSS assignments really helped me to learn the concepts because I had to understand the output and write it out. I felt like I was learning in layers, because every time I went back to analyze the output I saw something that I had overlook".

Discussion

The data indicated a significant decrease in statistical anxiety. The effect size index indicated this change was meaningful. Thus, a conclusion is drawn that the teaching methods effected student anxiety. The reduction can be caused by the fact that there was a variety of teaching methods used to demonstrate a multitude of statistical analysis. The author reinforced the subject material using a variety of different teaching methods. For example, to learn t-tests, they saw how a t-test is used with pre and posttest data, calculated a paired sample t-test (and interpreted the results), and read a journal article where a paired sampled t-test was used to describe the success of a therapeutic technique. Other statistical analysis received a similar treatment. Students viewed correlation, ANOVAs, and frequencies through real examples, journal articles, and SPSS. A conclusion can also be drawn that giving the students the opportunity to discuss the statistics tests through their journal writing ups, group presentations and SPSS discussions allowed them to immerse themselves further into understanding the statistical tests, thus potentially increasing their comfort with the material and further reducing anxiety. Students' opinions (both quantitative and qualitative) also offered further support of this conclusion. Overall, as the empirical and qualitative data indicated, students appeared to become more comfortable with using statistics. This conformability becomes evident in the students desire to read, interpret, and present findings based on the statistics used in the course.

There were limitations to this study. Given the small sample size, generalizing beyond the class may be difficult. However, the small sample does not violate assumptions of a dependent sample t-test (Montclam and Royse, 2002). Second, the conclusions that the teaching methods reduced anxiety are subjective. The methodology of the pre-post test without a comparison group is susceptible to maturation. The change in anxiety could have been the result of student growth over the semester rather than the teaching methods. Students' knowledge about quantitative methods might have also matured outside of class.

In conclusion, this paper presents the possibility that certain teaching methods might lead to a reduction in social work students' anxiety about quantitative methods. Further research needs to examine correlations between specific teaching methods and decreased anxiety. In addition, sampling of larger classes needs to occur. Finally, the research presented here can lend itself to an experimental design, where the control group gets traditional lectures and the experimental group receives novel teaching methods.

References

Cruise, R. J., Cash, R.W., & Bolton, D. L. (1985). Development and validation of an instrument to measure statistical anxiety. American Statistical Association Proceedings of the Section on Statistical Education.

Davis, S. (2003). Statistical anxiety among female African American graduate-level social work students. Journal of Teaching in Social Work, 23(3/4), 143-158.

Forte, J. (1995). Teaching statistics without sadistics. Journal of Social Work Education, 31 (2), 204-219.

Gustavsson, N. S. & MacEachom, A. E. (2001). Perspectives on research-related anxiety among BSW students: An exploratory study. The Journal of Baccalaureate Social Work, 7(1), 111-119.

Kline, R. B. (2004). Beyond significance testing: Reforming data analysis methods in behavioral research. Washington, DC: American Psychological Association.

Montclam, D. & Royse, D. (2002). Data Analysis for Social Workers. Allyn & Bacon: Boston, MA.

Royse, D. & Rompf, E. L. (1992). Math anxiety: A comparison of social work and non social work students. Journal of Social Work Education, 28(3), 270-278.

Andrew Quinn, Department of Social Work, University of North Dakota 

Quinn, PhD, is an assistant professor whose teaching interests are in research and human behavior