A Study On Construction And Application Of A Medical Statistics Course Satisfaction Index Model For Postgraduate Students

Objective:To construct a medical statistics course satisfaction index model for postgraduate students（MS＿CSI）,and to assess the reliability and validity of the satisfaction index model.To explore the relationships of course satisfaction with main predictive variables and outcome variables.To identify priorities for satisfaction enhancement based on student evaluations of the most important but underperforming aspects in course satisfaction.To provide a valid and reliable measurement tool for medical statistics course quality assessment and references for promoting the course satisfaction.Methods:A random cluster sampling method was used to select postgraduates who were participating in the medical statistics course and were about to finish the course.Based on ACSI model,the theoretical framework,initial latent variables and measurement variables pool were defined with consideration of the characteristics of the course and experts consultation.Discriminant analysis,Cronbach’sαcoefficient method,and correlation coefficient method were used for measurement variables selection,and those retained formed the test version of the MS＿CSI model.PLS-SEM was used to construct the test version of MS＿CSI model.Both the measurement model and the structural model were assessed.The IPA was used to identify factors with high importance but low performance.Results:（1）A total of 691 questionnaires were distributed and 633were recovered（recovery rate was 91.61%）.Among these,618questionnaires were valid（effective rate was 97.63%）.（2）Variable PQ7,PQ8 and SC2 were deleted during the process of observed variables removal.The test version of MS＿CSI model with 15 latent variables and 29 observed variables was constructed.The subject framework of the model contained 6latent variables:student expectations,perceived quality,perceived value,student satisfaction,student complaints and student trust.Perceived quality was a second-order latent variable and contained 9 first-order latent variables.（3）For the measurement model,Cronbach’sαcoefficient value for all latent variables were greater than 0.70,except for support services;and CR value were between 0.86～0.95.Outer loadings for all latent variables exceeded 0.708,and AVE were all higher than 0.50.Outer loadings of any observed variable on its corresponding latent variable were greater than its cross loadings on other latent variables.Except the correlation coefficients between perceived quality and student satisfaction,perceived quality and student trust and perceived quality and perceived value were slightly higher than the square roots of AVE of perceived quality,the square roots of AVE for other constructs were higher than correlation coefficients of any other constructs.The HTMT between perceived value and student satisfaction was slightly greater than 0.90,while the HTMT of the remaining variables,which was less than 0.90,suggesting the model had good reliability and validity.For the structural model,perceived quality had the largest total effect on student satisfaction（b=0.80,P<0.001）.Among the first-order latent variables of perceived quality,the latent variable that had the greatest effect of student satisfaction was teaching content（b=0.24,P<0.001）.The coefficient of determination(R_adj²)was 0.77 for student satisfaction,the predictive relevance（Q²）for all endogenous variables was greater than 0,the effect size（f ²）of perceived quality on student satisfaction was 0.37,all suggesting that the model had strong predictive ability for student satisfaction.（4）The student satisfaction index scores of students who majored in clinical medicine,nursing,pharmacy,and public health were 75.06,72.87,72.71 and 72.02,respectively.（5）According to the IPA results of student trust,the important but underperforming latent variables of student trust were student satisfaction and perceived quality,and the performance of the two variables was75.28 points and 77.22 points,respectively.According to the IPA results of student satisfaction,the most important variable affecting student satisfaction was perceived quality,with high importance（0.80）but low performance（77.22）.According to the IPA results of perceived quality,teaching styles and teaching effects were both highly important factor for perceived quality（0.15 and 0.19,respectively）,and the performance score for teaching styles and teaching effects were relatively low（75.26and 63.64,respectively）.Conclusions:（1）The test version of MS＿CSI model was composed of 15 latent variables and 29 measurement variables.Student expectations,perceived quality and perceived value were explanatory variables of student satisfaction,and student complaints and student trust were outcome variables of student satisfaction.Perceived quality consisted of 9first-order latent variables,which included overall quality,teaching content,teaching arrangement,teaching styles,teaching attitude,examination styles,teaching effects,independent learning quality and support services.This model was reliable,valid and had a good predictive ability.（2）In medical statistics course,improving perceived quality was key to improve student satisfaction and student trust,and teaching styles and teaching effects were the key factors for the enhancement of perceived quality in teaching.