Study design and population
This was an institutional-based cross-sectional study conducted from May to May 2018 at Addis Ababa University (AAU). Of 1650 medical students, 390 of them were randomly selected and completed the required questionnaires. All undergraduate medical students were eligible to participate in the current study.
Sample Size determination and Sampling techniques
The sample size (n) to estimate the prevalence of test anxiety among medical students was determined by a single population proportion formula. 95% confidence level was considered to estimate the sample size and since there are no previous similar studies conducted in Ethiopia, we assumed the prevalence of test anxiety 50% and a precision of 5% between the sample and the parameter was taken. α =0.05 (95%) =1.96 (see Formulas in Supplementary Files)
Additionally, we calculated sample size based on the potential associated factors using STATA 14.2 for sample size calculations, assuming the proportion of test anxiety 50% (no similar study in the area) and taking the odds ratio of 2. In addition, we assumed a significant level of 5% and 80% power. The total sample size was 296 (148 per groups). Adding 10% non-response rate the sample was 320 for each associated factors. However, we used the larger sample size (calculated based on proportion, n=423) for this study.
Concerning the sampling technique, in this study, a total of 390 participants, were randomly selected by using a stratified random sampling technique. As suggested, we used a proportionate stratified random sampling to allocated sample for students in each year (year 1 to 5) and we employed a simple random sampling method to select each participant. (See figure 1).
Measures
Measures for the dependent variable (test anxiety)
Data was collected by a self-administrated questionnaire. Westside Test Anxiety inventory developed by Richard Driscoll (WTAI), is a self-reported questionnaire of 10 statements to which respondents are asked to report how often they experience anxiety symptoms before, during, and after taking tests. Each statement response is scored with a 5-point Likert scale (1–5) yielding a total test anxiety score ranging from 10 to 50 points. Participant were attributed to five different levels of test anxiety according to WTAI score: score of 1.0-1.9 comfortably low-test anxiety, a score between 2.0—2.5 Normal or average test anxiety, 2.5—2.9 High normal test anxiety, 3.0—3.4 Moderately high ,3.5—3.9 High test anxiety ,4.0—5.0 Extremely high anxiety and the cut point for problematic test anxiety is ≥30. The Cronbach alpha was alpha of .78, split-half reliability of .77 in a Nigerian sample(17, 18).
Therefore, In this study, those students who scored below 30 were considered as having no test anxiety and those score 30 and above were classified as having test anxiety(18). In our sample, the internal consistency of WTAI was very high (Cronbach's alpha =0.94).
Measures for the predictor variables
Study skills
Study skills referred to as the ability to effectively use the specific skills (learning and planning study, library use, note-taking, course participation, preparation for exams, motivation, preparation for courses, effective reading, writing, health and nutrition, and listening skills). Time Management (TM) includes six items inventory was used in this study to measure the time management of students it has 6 questions and 5 points Likert scale<20 suggest not effective time management skill (19).
To assess the study skills we used approaches and a study skills inventory for students (ASSIST) (19). The inventory consists of 51 items 5-point Liker scale. It measures the scores on six subscales regarding study habits and skills. These subscales are Textbook Reading, Note-Taking, Memory, Test Preparation, Concentration and Time Management. The domain of Time Management (TM) includes six items of the inventory. The convergent validity of the tool is high and consistency reliability is 0.96 which is
Self-esteem
Rosenberg Self-esteem scale (RSES) was used to measure levels of self-esteem among the study participants (20). Students were scored based on a 4-point Liker scale, scoring of the scale “Strongly Disagree” 1 point, “Disagree”2 points, “Agree” 3 points, and “Strongly Agree” 4 points. The sums of the scores for all ten items and keep them on a continuous scale. Higher scores indicate higher self-esteem. The scale ranges from 0-30. Scores between 15 and 25 are within normal range; scores below 15 suggest low self-esteem. Internal Validity is 0.78-0.84 And reliability is 0.94 which is high (20).
Psychological Distress
Psychological distress was measured using the Kessler Psychological Distress Scale (K-10). It is a simple measure of psychological distress, the K10 scale involves 10 questions about emotional states each with a five-level response scale. According to the scale, a score 10 - 19 Likely to be well, 20 - 24 Likely to have a mild disorder, 25 - 29 Likely to have a moderate disorder, 30 - 50 Likely to have severed distress. The scale has a sensitivity of 70% and specificity 67% (21).
Social support
We also measured social support of the students and individual who were scored 3-8 poor, 9-11 moderate and 12-14 strong social support on Oslo 3-item social support scale (OSSS-3) (22). Evidence suggests that the OSSS-3 is a very brief and economic scale which has satisfactory internal consistency with an α=.640 (23). Although OSSS-3 is not validated in Ethiopia, it is extensively used in previous studies to assess social support in different population groups (24-26).
The sociodemographic and other variables
We collected self-reported data on sociodemographic and other characteristics of the participants using self-reported and structured questioners.
Data collection procedure
In this study to maintain the quality of the data high, before they engage in the actual data collection activities the data collectors received adequate training on data collection procedure and protocol. Data was collected by using an English version of self-administrated questionnaire.it had components which assess a different aspect of the participant. The 1st part assessed a socio-demographic characteristic of participants. The 2nd part assessed the level of test anxiety by using the WTAI scale. The 3rd part was on psychosocial factors (psychological distress, self-esteem, and social support) which affect the level of test anxiety and measured by K-10 scale, RSES and oslo3-social support scale respectively. The 4rth part was on questions assessing behavioral factors using SSI scale to assess time management. The fifth part assessed the academic-related factors that affect test anxiety using yes or no questions. This questionnaire disseminated to 423 eligible medical students.
Data quality assurance
The questionnaire was designed and modified appropriately. The self-administrated English version was disseminated. The training was given for data collectors and supervisor. Pre-test conducted two days before the actual data collection on saint Paulo’s medical school and the Pre-test was conducted among 22 samples (5%) of saint Paulo’s medical students; the result was not included in the main survey.
Based on the finding from the pre-test, the questioner revised and helped to estimate the time needed for data collection. The data collectors were supervised daily and assist the students to fill the questionnaire and checked daily by the supervisors and principal investigator. The solution to problems during data collection was given immediately by discussing with the supervisors and the data collectors.
Data processing and analysis
First, the data were checked for completeness and consistency and then coded and entered in the computer using EPI DATA software for cleaning, storing and recording and then imported to SPSS version 25 for analysis. Descriptive statistic was used to explain the study participants in relation to study variables. A p-value of less than 0.05 was considered statistically significant. Bivariate and multivariate logistic regression analysis was conducted to identify associated factors of test anxiety among medical students and the strength of the association was presented by odds ratio with a 95% confidence interval.
Ethical consideration
This cross-sectional study was reviewed and approved by the Institutional Review Board (IRB) of the University of Gondar and Amanuel Mental Specialized Hospital. Informed consent from each participant was obtained after clearly explaining the objectives as well as the significance of the study for each study participant. We advised the study participants about the right to participate as well as refuse or discontinue participation at any time they want and the chance to ask anything about the study. The participants were also advised that all data collected would remain confidential.