Once the data is collected, it is imperative that it gets a thorough data cleaning (also known as data management) before it is analyzed. All values must be checked and crosschecked. For example, a variable Gender must only have two possibilities, male or female. In some cases when the data is entered, it might be entered with typographical errors. Say that for some of the observations the variable, Gender  has  an  answer of female and for some others it may say F. Now, both are the same and we all know it, but unfortunately the computer doing the calculations will consider these two fields as different. During data cleaning, these two answers can be placed into one category.

Other things to check for are inconsistencies. For example, if you have a variable for the year of birth and another representing age, and the two are markedly different, then a decision needs to be made as to which of the two is correct. In most cases, this would be a data entry error and is easily corrected by checking the original records. However, if this is impossible, other solutions must be considered.

Other facets of data cleaning include recoding the variables so as to make the interpretation of the results easier and creating new variables to facilitate in the data analysis. 

Contact Statistical Consulting Network for help with your Data Cleaning