A total of 611 veterinary students of India from 14 states and 27 colleges/universities provided their consent to participate in the study. Table 1 concludes the participants’ demographic breakdown. The mean age of respondents was 23.2 ± 2.8 years old with majority identified as male (63.6%) respondents. A total of 247 (40.4%) students had family income below 2 lakhs. Maximum students (67.10%) were from undergraduate (BVSc and AH) degree year, followed by MVSc students (27.33%) and PhD students (5.57%).
Construct validity
The exploratory factor analysis was used to determine the construct validity of the questionnaire. The Kaiser-Meyer-Olkin test (KMO) that measures adequacy of sampling adequacy was applied and the value was found to be 0.957. KMO is a statistic that indicates the proportion of variance in the variables that might be caused by underlying factors and high value (close to 1.0) generally indicates that a factor analysis may be useful with our data. The KMO value obtained for the present study thus indicated the very high sampling adequacy. The Bartlett’s test of sphericity was found highly significant (chi square = 1222, p=0.000) which again indicated that a factor analysis may be useful with our data. The total numbers of components were extracted using principal component analysis and Promax rotation (oblique rotation). The initial exploratory factor analysis resulted in the extraction of 8 factors with Eigen value > 1 that collectively explained 56.98% of overall variance of the data. As several of the 8 factors comprised only two factors and there was shared variance between factors, items were forced to a 6-factor reduction, with suppression of coefficients below 0.30. The 6-factor reduction explained 51.98% of total variance. The loading of each item to the new factors was based on the factor loading and our construct as in method section. Factor structure along with communalities and internal consistency of each individual items along with factors is detailed in the Table 2. In this study, the scores of the items within each subscale were summed and mean score of each subscale was used to represent stress level of the students in six newer categories: academic related stressors (ARS), Interpersonal, intrapersonal and conflict related stressors (IPCS), Teaching and learning related stressors (TLRS), Examination and evaluation stressors (EES), Social activities related stressors (SARS) and Family responsibilities stressors (FRS). A mean score more than 1.0 was taken as a stress and more than 2.0 was taken as high score. Overall stress was calculated by summing up the mean score values from each subscale and taking the average and then evaluated by same method as used for each subscale.
Reliability
The questionnaire showed high internal consistency or reliability owing to its Cronbach’s alpha value at 0.957 which lies in excellent grade. Moreover, Cronbach’s alpha values of subscales ranged from 0.647 to 0.892, indicating good internal consistency (Table 3). When individual items were removed, the Cronbach’s, alpha ranged from 0.578 to 0.889 showing that all the items contributed to the adequacy of the scale (Table 2). Additionally, an anti-image correlation with individual items also confirmed that the sampling was adequate (p<0.001) for further analysis.
Inter-factor correlations as well as correlations of factors with other variables of questionnaire (gender, income, and degree year) were examined using bivariate variation (Spearman’s rho) as shown in Table 4. Inter-factor correlations ranged from between 0.006 and 0.608, which indicates generally acceptable independence. The highest correlation between ARS and IPES (0.608) indicate some overlap between these two factors. All the factors, except IPCS had a higher discriminant ability and were statistically significant (p<0.05). All the factors were significantly correlated with overall stress except for FRS and TLRS, and these two seems to have lower discriminant ability to measure overall stress than other factors. Correlation between factors and other variables varied from .006 to .280. ARS and overall stress were significantly correlated to gender and income but not to degree year. Inter-item total correlation value was more than 0.3 for 20 items. Rest of the items however, had a lower but statistically significant correlations and there was no considerable change in Cronbach’s alpha with the deletion of any item, indicated acceptable reliability. Therefore, all the 44 items were included in the questionnaire. Cronbach’s alpha value of each stressor group is shown in the Table 3 and Cronbach’s alpha value if scale item deleted value for individual item is shown in the Table 2.
Stress and its relations with other variables
All the students reported some degree of overall stress. Students with mean score of more than 1.0 i.e., those who reported score in moderate, high and severe category were considered under stress. Overall, for 611 students, 94.10% (575) were under stress (mean score more than 1.0), out of which 59.65% (343/575) students had high to severe stress (mean score more than 2.0). A total of 11.45% (70/611) were under severe stress. The most important stressors group among veterinary students was ARS (95.58%) followed by IPCS; TLRS; EES; SARS and FRS as shown in Table 5. In summary, Table 6 shows that female students significantly experienced more overall stress, ARS, and IPCS than male students. The students who had less than 2 lakhs of annual family income significantly experienced high overall stress and FRS as compared to students having higher family income. The students of second year of bachelor’s degree experienced significantly higher SARS as compared to others and students of first year of bachelor’s reported significantly higher stress due to FRS.
Hierarchical regression analyses
A series of hierarchical regression analysis were conducted to investigate which stressors significantly predicted the overall stress among the students. In the regression analysis gender was entered in step 1, as according to chi square test it has significant association with stress. In step 2, degree year was added assuming that it might predict the stress among undergraduate students. In step 3, family income was added as chi square has shown significant association with overall stress. Finally in step 4, the six stressors (EES, ARS, FRS, SARS, IPCS and TLRS) were added. Only single dependent variable (i.e. overall stress) was used.
The results of the hierarchical regression analysis indicated that in step 1, gender positively predicted and accounted for 3% of the variance (p < .001). In step 2, degree year was not significant. In step 3, the family income positively predicted and accounted for 6 % of the variance (p < .001). In step 4, the six stressors significantly accounted for additional 68% of the variance (p < .001), with only ARS, IPCS and SARS positively predicting overall stress.
To summarize, female students had a higher level of overall stress than male students. The students with lower family income have reported high overall stress and the students who reported high level of ARS, IPCS and SARS, were more likely to be included in high overall stress category.