I have dealt with the questionnaire data recently, and I want to record my feelings about questionnaire data analysis while I have just finished it, mainly to talk about the basic processing methods and methods in the process of questionnaire data analysis.
1. Descriptive statistics on each topic
Now the general questionnaire data is done with the questionnaire star scale tool, and then the descriptive statistics of each topic can be output in the questionnaire star, and the chart of each topic can also be output, such as histogram, line graph, bar graph, etc.
2. The relevance of the two topics
If you want to analyze the correlation between the two topics, you need to discuss it separately.
2.1 The relevance of multiple choice questions and multiple choice questions
The correlation between single-choice questions is best done. You can use chi-square test combined with cross-analysis to determine whether there is a correlation between the two questions and how strong the correlation is. This strength can be determined by the card It can be judged by the column connection number or the V coefficient of the square test. For details, click here to see
2.2 Multiple choice questions (multiple choice questions) and sorting questions
When it comes to sorting questions, it will be a bit difficult, because I did not know what good statistical analysis methods are available, but only simple analysis. There are analytical methods for calculating the scores of various options in the questionnaire star, which can be used for single-choice questions ( Observe the scores of each option of multiple choice questions to draw some conclusions
2.3 Multiple choice questions (multiple choice questions) and multiple choice questions
There are also difficulties when it comes to multiple choice questions. There is no very good statistical analysis method. The most common method is multiple response analysis. This can be searched directly on the Internet. SPSS is also very simple to do. In fact, it is to do a cross analysis. child. In addition, if two questions may have a causal relationship, logistics regression analysis can be considered when analyzing the influence of one quantity on the other.
3. Regression analysis
Third, it is natural to make further analysis using questionnaire data. Usually this is to use regression for in-depth analysis to analyze the influence of some variables on a dependent variable. Due to the characteristics of questionnaire data, the most commonly used is binary logistics Regression or multivariate ordered/disordered logistics regression analysis.
For spss to do multivariate ordered logistics regression, please refer to this example: multivariate ordered logistics regression tutorial
unordered multi-classification can see this tutorial: multivariate disordered logistics regression.
However, for the interpretation of the parameters, my own results are contrary to this tutorial. Yes, I don’t know why, maybe I didn’t understand it well, anyway, I personally suggest to do a cross-analysis after the multiple logit regression, and use the results of the cross-analysis to interpret the meaning of the parameters of the multiple logit. This will ensure correctness.
4. Summary
Since the analysis done recently is relatively simple, so basically only the above three methods are used. Of course, there are actually many other analysis methods, such as canonical correlation analysis, cluster analysis, correspondence analysis and so on. Since I have done very little to analyze the questionnaire data, I have a rather weak understanding of this aspect. Please correct me if there is something wrong. If the boss can give some pointers, I would like to thank you!