Research in physical medicine and rehabilitation. V. Data entry and early exploratory data analysis.

The process of data entry and initial analysis to locate data errors is described. Basic terms are defined and a simple method of entering data by using word processing software is illustrated. Data checking is done by using visual check of the raw data. Statistical programs are then used to locate possible data errors by finding data points (outliers) that are very different from the average. Special graphic output of statistical programs, scatterplots and box and whisker plots can be used to further locate questionable data. Examples of data entry forms and annotated step by step data cleaning with the use of inexpensive programs for personal computers are presented.