Data cleaning
From PsychWiki - A Collaborative Psychology Wiki
Tips/Tricks for Data cleaning...
(this page has just been created... can you think of more tips to add...)
- Always first look at distributions of data and outlier analysis.
- Try to be consistent in what outliers method you use when cleaning data.
- S+ and other robust procedures inherently correct for outliers.
- SPSS tukey’s box plot rule and graph’s boxlot allow you to see outliers
- Winzoring is another good idea.
- Both univariate and Multivariate outlier analysis should be conducted.
- For pure data cleaning, such as out of range or missing data, maybe you should spot check R.A data entry to make sure data is being entered correctly
- Try having variable data file that codes for who entered which parts of the data.
- Try to enter the data at least twice because everyone makes mistakes, even you.
◄ Back to Analyzing Data page
