1、 Principal factor analysis

The so-called principal factor analysis is to conduct principal factor analysis on various indicators of the survey questionnaire and screen out useful indicators for the paper. For example, here we use teacher satisfaction rating data, which includes a total of 10 teachers and 30 pieces of data. Some of the data is shown in the following figure.

Figure 1 Data Display

Click on "Analysis" - "Dimensionality Reduction" - "Factors" in the top menu bar of SPSS to open the Factor Analysis window. Load all indicators into the variable text box, click the description button, and check the "KMO and Bartlett sphericity test" under the correlation matrix item.

Figure 2 Factor analysis

Click the "Rotate" button on the right side of factor analysis and select either the direct oblique method or the optimal oblique method.

Figure 3 Rotation

Seeing KMO and Bartlett's test, if the KMO sampling suitability scale is less than 0.6, it is not suitable for factor analysis. It can be seen that its value is 0.633, which is greater than 0.6, indicating that factor analysis can be conducted.

Figure 4 KMO and Bartlett's test

The above verification indicates that the data can be subjected to factor analysis. The total variance analysis extracted 4 factors from 9 questions, and the cumulative amount of 4 common factors was 67.697%.

Figure 5 Explanation of Total Variance

2、 Optimal Scale Regression Analysis

If regression analysis is divided according to the continuity of variables, it can be divided into two types: one is the regression analysis of continuous variables, mainly using linear regression and logistic regression. The second is the regression analysis of discontinuous variables, mainly using the best scale regression analysis.

For example, a clothing brand collects data on consumer satisfaction, marital status, gender, age, and monthly income in order to understand consumer satisfaction with the brand. Satisfaction is divided into three levels (1 represents dissatisfaction, 2 represents general satisfaction, and 3 represents satisfaction), marital status (1 represents unmarried, 2 represents married), gender (1 represents male, 2 represents female), age has seven levels, and monthly income has four levels. Some data are shown in the following figure.

Figure 6 Data Display

Click on "Analysis" - "Regression" - "Best Scale" in the top menu bar of SPSS to open the classification regression window. Load satisfaction into the dependent variable text box and define the scale as ordered; Load gender, marital status, age, and monthly income into the independent variable text box, and define the dependent variable as ordered.

Figure 7 Classification Regression

Click on the option button on the right and select 'Multiple systematic suspension points' in the initial configuration item.

Figure 8 Options

Click the save button on the right and select "Save converted variables to active dataset" in the variable module after conversion.

Figure 9 Save

Click the button on the right to load the 4 independent variables into the text box of the conversion graph.

Figure 10

Looking at the "ANOVA" project, it can be seen that a significance value less than 0.01 indicates a value less than 0.05, meaning that at least one independent variable has a significant impact on the satisfaction of the dependent variable.

Figure 11 ANOVA table

Looking at the "coefficient" item, it can be seen that monthly income has a significant impact on the satisfaction of the dependent variable.

Figure 12 Coefficient Table