Current location - Education and Training Encyclopedia - Graduation thesis - How to analyze the data of undergraduate thesis? What data are needed for correlation analysis, hypothesis testing and regression analysis?
How to analyze the data of undergraduate thesis? What data are needed for correlation analysis, hypothesis testing and regression analysis?
Research methods can usually be divided into three categories, namely, difference relationship, correlation relationship and other relationships.

If the thinking is more biased towards the study of differences, such as the different attitudes of different income groups to online shopping. It is suggested to use more standardized scale questions, because the scale is very standardized and many research methods can be used; If you don't use scale questions, then you can consider chi-square analysis for research. If we use richer research methods, we need to use a variety of question designs, both scale questions and non-scale questions, which are expected to enter the category of difference comparison.

If thinking is more inclined to study the influence relationship, such as the influence of satisfaction on loyalty, it seems that both satisfaction and loyalty can be expressed by scale questions, then there is no problem in designing scale questions, because linear regression analysis can be used for research. In addition, there is another situation that can be considered, that is, logistic regression. Whether satisfaction affects the final purchase and whether it is affected by satisfaction should be analyzed by logistic regression. If you want to use both research methods, then the questions corresponding to satisfaction need to have scale questions and classified data questions such as "Are you willing to buy again".

If the expected data needs statistical analysis, please remember that it must be designed as a scale question, otherwise the reliability analysis cannot be carried out. If there is a classification in the expected thinking, that is, for example, dividing the samples into three groups, then we should consider using more standardized scale data.

To sum up, the matching use of research methods should actually be considered before the questionnaire design. After the questionnaire research design is completed, most of the questionnaire research methods have been determined, so it is necessary to take the questionnaire research methods into account in advance in order to analyze more abundant data. Relatively speaking, the scale questions can match more research methods and are more standardized. It is suggested to use scale questions more.

References:

/p/5