| Object To investigate the feasibility of PBL in the medical postgraduates( hereinafter referred to as postgraduates) which is in the initial stage at present by evaluating the teaching quality of the PBL courses offered for the postgraduates in a medical university. Based on the result of data investigation, to revise and optimize the teaching evaluation system and scale of PBL for postgraduates, providing reliable tools for evaluating the teaching quality of PBL courses. To investigate the major influence factors of teaching effects of PBL for postgraduates, multiple linear regression and structural equation modeling were adopted. Meanwhile, to investigate the application of multiple linear regression and structural equation modeling in the field of education and teaching evaluation and provide methodology support for the relevant work.Method Evaluation system of the teaching quality of PBL for the postgraduates and self-made questionnaire with good reliability and validity were established in the preliminary study, on the basis of which the content of questionnaire was supplemented with learning process, course satisfaction and so on. The questionnaire was conducted to survey 16 PBL courses, the respondents were 536 postgraduates from a medical university who attended PBL courses. Normal distribution quantitative data was described by (?)±s, skewed distribution quantitative data was described by M(Q_R) , and qualitative data was described by rate or constituent ratio.T-test and one-way ANOVA were adopted in univariate analysis for statistical inference of quantitative data, SNK-q test was used in multiple comparison,chi-square test was used by quantitative data. Rank sun test was adopted for statistical inference of data of abnormal distributions, Spearman correlation analysis was applied to correlation test between variables. Influence factors analysis and methed comparision of PBL for postgraduates were conducted by multiple linear regression analysis and structural equation modeling. In addition, Items of the scale were analyzed and screened with critical ratio and items and total correlation analysis.Reliability was tested by intra-class correlation coefficient, Kendall’s concordance coefficient andacoefficient, while validity was evaluated by content validity ratio and factor analysis. The level of a test was α=0.05 bilateral. Data entry and management were accomplished with Epidata3.02, the statistical analysis was conducted with SPSS 19.0 and fitted by structural equation modeling with Amos7.0.Result 528 questionnaires were retrieved from a total of 536 questionnaires with the return rate 98.507%, of which 497 questionnaires were valid so that the effective rate was 94.128%. The average age of the respondents was 24.521±1.743, 31.791%of whom were males and 68.209 % of whom were females. Besides, 93.159% of the respondents were non-directional and self-funded postgraduates, while 6.841% were directional and entrusted postgraduates. In addition, 46.881% of them were from the School of Basic Medicine,while 53.118% were from the School of Clinical Medicine.(1) Result of the quality evaluation of PBL for Postgraduates: The general evaluation score of PBL courses was 85.200(11.951), which presented a skewed negatively distribution. Chi-square values of the first grade index and general evaluation score of16 PBL courses were 67.159, 41.405, 58.287, 63.983 respectively,and the difference was statistically significant(P<0.001). The postgraduates with different gender, major, age and source showed different levels of enthusiasm of attending PBL courses, Chi-square values of which were7.446, 223.248, 350.648,5.292, and the difference was statistically significant(P<0.05). In the evaluation of optimum selection, quality evaluation of PBL courses in the university was distributed reasonably but not reaching the elite level.(2) Result of the analysis of influence factors of teaching effects of PBL for postgraduates: Multiple correlation coefficient of multiple linear regression analysis was R=0.698, correction multiple correlation coefficient was Radj=0.686, analysis of variance of multiple linear regression model was F=58.238, P <0.001, and goodness of fit was fairly good. 8 variables that entered the model in sequence were data preparation, relation to the major, types of the courses, the number of selected courses, participation in the class discussion, tutor or not, cognitive degree, working experience. Overall collocation degree test of structural equation modeling was χ~2=19.400, degree of freedom was v=16, P=0.250 and theoretical model could match with the data collected in the research. RMSEA=0.047,AGFI=0.911,CFI=0.930,χ~2/df=1.210 so the model fitting was good. The potential factors which influenced the teaching effect of PBL for postgraduates were learning process, teacher disposition,students, types of courses, among which learning process had a greater influence on the teaching effect than other factors and the factor-students potentially influenced the teaching effect by learning process. There were 12 influence factors of PBL teaching effect on postgraduates which finally entered the model, including professional title,tutor or not, study abroad experience of the teachers, doctor qualification, working experience, category of student source, the number of selected courses, data preparation, relation to the major, participation in the class discussion, cognitive degree, types of the courses.(3) Result of optimization of the evaluation scale of PBL teaching quality for postgraduates: Both critical ratios of 27 three-grade evaluation indexes in the scale and correlation of items and total scores item of scale were statistically significant(P<0.05). ICC of test-retest reliability was more than 0.75; Kendall’W of scorer reliability was 0.896(χ~2 =12017.193, P<0.001); Totalacoefficient of internal consistency reliability was 0.940, whileacoefficients of 3 first-grade evaluation indexes: ”Teaching plan, Teaching strategy and Teaching effect” were0.780、0.890、0.915 respectively; CVI of each item in content validity was between0.6 and 1.0,while CVI of three-grade evaluation indexes “Difficulty level, Critical thinking” was 0.6, which was relatively low; In the test of construct validity, KMO was 0.957 and Bartlett′s test P<0.001, which indicated that data in the sample was suitable for factor analysis in which 4 common factors were acquired whose accumulated variance contribution rate was 61.774%, which showed that the result of factor analysis was consistent with the structure of questionnaire design and presented that 2 three-grade evaluation indexes- “difficulty level” and “English level” needed to be adjusted. Based on the result, 3 three-grade evaluation indexes“Difficulty level, English level, Critical thinking” and correspondence problem item were revised.Conclusion The general evaluation of PBL courses offered in the university is relatively high but there is still great room to improve the teaching quality. PBL conforms to the practical needs of the development of domestic higher medical education. When it comes to dealing with the problems related to the analysis of influence factors in education, multiple linear regression analysis is an easy method with strong application which is more appropriate for the primary stage of the research and provides theoretical support for the late-stage study. While, the structural equation modeling is more applicable to deal with the relatively complicated relation,especially where potential variables exist and acquires more reliable results. The teaching effect of PBL on postgraduates can be improved by adjusting case design, evaluation mode and so on; inviting excellent PBL teachers to give demonstration lectures periodically; realizing the complementation with each other’s advantages of the postgraduates between groups by dividing the groups rational ly; distributing teacher resources reasonably and making effective training scheme for teachers; promoting the autonomous learning awareness of the postgraduates; trying PBL courses in the preponderant disciplines firstly. In this research, the scale and evaluation system are revised and optimized, which provides a more stable and reliable tool for the teaching evaluation of PBL courses for the postgraduates and theoretical basis for the application and promotion of PBL mode in the teaching of postgraduates in China. |
PDF Full Text Request