Path Analysis On The Attitude Towards Premarital Sexual Behavior Research In Medical University Students

Objective: In recent years, with the development of the economic reforms, people have been influenced much deeper by the western culture, and their thoughts have been liberated. The traditional chastity value and morality concepts were impacted. The age of our country's college students is about 20 years old, which is the best age stage to carry on sexual education. Nowadays, people have paid much more attention to the teenager's gender and health problems. So it is necessary to carry out gender and health education. In order to make clear of the students' attitudes towards premarital sexual behavior in medical university and its influential factors, a survey was conducted in Hebei Medical University. In this survey, some methods such as path analysis were used to investigate the influential factors of premarital sexual behavior. According to this, a more directed and valid educational model of gender health in medical university could be set, and better psychologic consultation could be carried through, by which the teenagers'body and mind could be developed much more.Method1 Material: This study based on professor's material. In this material, we regarded 14 classes of undergraduates in Hebei Medical University as 14 clusters, in which 3 classes(540 undergraduates in total) were randomly sampled by cluster sampling to conduct a survey between September and December in 2003.2 Statistical methods2.1 Evaluate the validity, reliability and the acceptability of the questionnaires on the general condition of the medical university students.2.2 Path analysis2.2.1 Multicolinearity diagnosis for independent variables: the indexes of diagnosis were tolerance, VIF, eigenvalue and condition index.2.2.2 Conditional test of the model: the residual test was used.2.2.3 Screen the independent variables by multiple stepwise regression to build up an initial cause-and-effect network model.2.2.4 Partial correlation2.2.5 Calculate the path coefficients:the direct path coefficients=the standard regression coefficients ( P_ij)the indirect path ceofficients=correlation coefficients( r_ij)×standard regression coefficients( P_ij )the rest path coefficients U = (1-R²)^1/2, R² =∑P_ij -r_ijR²: the square of multiple correlation coefficientsP_ij : the path coefficients of X_i to X_jr_ij: the simple correlation coefficients of X_i to X_j the total path coefficients r_iy = P_iy +∑r_ij×P_jy2.2.6 Calculate coefficients of determination:In the causal variables, the coefficients of determination of X_i to Y is d_i = (P_iy)² if other variables were fixupped. And the associated effect of coefficient of determination of( X_i ,X_j)to Y d_ij = 2r_ij P_iy P_jy.2.2.7 Build the path diagram2.2.8 Evaluate the model: the indexes of diagnosis were GFI and AGFI.Results1 The general conditions of the materials In this survey, 540 questionnaires were effective. The effective rate was 94.571%. There were 249 males(46.111%) and 291 females(53.889%) taking part in the survey.2 The evaluation of the questionnaires2.1 The construct validity of the questionnairesThrough factor analysis, the eigenvalue of the four anterior common factors were more than 1, the cumulative contribution rate was 52.078%, and the factor loadings on every term in the corresponding common factors were more than 0.4. The four common factors were relative to the circumstance in school and family, the treatment of sexual behavior and subjective feelings. The above indicated that the construct validity of the questionnaires was good.2.2 The Cronbach's alpha was 0.838, which indicated that the reliability of the questionnaires was good. 2.3 The questionnaires included 16 items. It was single choice form, which must be finished in 10 minutes. The contents were attractive to medical university students, easy to be finished and accepted.3 Path analysis3.1 Screen the independent variables by multiple stepwise regression to build up an initial cause-and-effect network model.3.1.1 Multicolinearity diagnosis for independent variablesIndependent variables'tolerances were all above 0.1, VIFs were all below 5, eigenvalues were all above 0, and condition indexes were all below 30. So there was no severe multicolinearity among the independent variables.3.1.2 Results of residual analysis: The histogram of residual was standard normal distribution. The scatter plots in normal P-P graph of residual were in linear tendency, and no abnormal plot was found.3.1.3 Results of multiple stepwise regression: Choose the multiple linear regression model by Cp, six independent variables were reserved in the regression equation. They were the chastity value(X₁₄), the literary level of mother(X₄), number of cohabitation before marriage(X₁₂), the attitudes towards study(X₅), the attitudes towards regulations of school about amative behavior(X₁₁), satisfactory extent to study and living(X₇). Take the independent variables in the equation was taken as dependent variables, while the other variables were taken as independent variables to carry on the multiple linear regression.3.2 Results of partial correlation analysis: The factors related to the sexual behavior attitudes were chastity value(X14), literary level of mother(X₄), number of cohabitation before marriage (X₁₂), the attitudes towards regulations of school about amative behavior(X₁₁) and satisfactory extent to study and living(X₇); The factors related to chastity value were sex(X₁), the attitudes towards sexual liberation(X₁₃), the times of amativeness(X₉), the influence of amative behavior to study(X₈), income of family(X₂) and the number of cohabitation before marriage (X₁₂), sex(X₁); The factors related to satisfactory extent to study and living were attitudes towards the study(X₅), the times of amativeness (X₉), the number of courses which failed to pass(X₆) and the restrainment of parents(X₁₀). Draw the path diagram according to the variables selected.3.3 Results of path coefficientAccording to the results of path analysis, the factors could be ranked according to their effects as follows: the number of cohabitation before marriage, the literary level of mother, the chastity value, sex, the attitudes towards regulations of school about amative behavior, the times of amativeness, satisfactory extent to study and living, the attitude towards sexual liberation, attitudes to the study, the influence of amative behavior to study, the number of courses which failed to pass, the restrainment of parents and income of family.3.4 Results of coefficient of determination The coefficient of determination was 41.4%.3.5 Results of EvaluationThe values of GFI and AGFI were 0.896 and 0.921, which indicated the model was good.Conclusions1 In this survey, the questionnaire was good in the validity, reliability and acceptability. It could satisfy the requirement of the study.2 Path analysis could decompose the linear coefficient of regression to direct effects and indirect effects when we researched the different forms of different relations. Then we could get a more specific form of the attitude towards premarital sexual behavior between the relationship of cause-and-effect.3 Path analysis is a kind of mehtod to measure the opposite importance of dependent variables and has already been applied well in this research.4 Sex wasn't taken in the equation when researching the influential factors of sexual behavior attitudes. But it could make indirect effect through X₁₁ and X₁₄. Its influence was 0.313. The number of cohabitation before marriage not only directed the attitude of sexual behavior, but also made indirect effects on the attitude of sexual behavior, enlarging the function strength. Its influence was -0.408, which indicated that the attitude towards sexual behavior had become more and more assentient with the growth of the number of cohabitation before marriage.