Missing Value Filling Method Based on Scale Data-Correlation Coefficient Method

Scales are a common form in the preparation of questionnaires. Missing data becomes common in scales for a variety of reasons. The traditional methods of dealing with scale data with missing values usually include direct deletion of cases containing missing values, interpolation using single values, multiple interpolation, etc. In recent years, there has been an increasing number of fill methods for various types of missing data, but few scholars have conducted research on fill methods for missing values specifically for scale data. This paper proposes a method to fill in missing values by using the correlation coefficients between the items of the scale, which is called the correlation coefficient method, in view of the fact that the items of the scale are usually correlated to a certain extent. In this paper, we use some of the scale data obtained from the Q City secondary school students' family education project as an example, and use the mean squared error and mean absolute error as evaluation indicators. By comparison with the common median and plurality methods, it was found that filling the scale data using the correlation coefficient method yielded the smallest mean squared error and mean absolute error, with the median method being the second most effective and the plural method the least effective. Therefore, the correlation coefficient method is more effective and suitable for scale data with large correlations.