Management

"Missing values are a persistent issue that most active analysts encounter at some point. It is a part of the larger topic of data cleansing and data wrangling. There are several methods, differing in complexity:

i. Ignore the issue at first and let your stats package take care of it. That will depend on the quantity of missing data, whether it pertains to important variables, and the approaches you are employing. 

ii. The second option is to impute data using a technique called ""hot-deck imputation."" At that point, you impute the missing value using the dataset's final non-missing value. 

iii. Third, you can extrapolate the numbers by applying the mean for the cases with complete data to the missing data. That guarantees that the software won't eliminate any cases for you, but it's a pretty direct strategy, so there won't be much variation in that variable (depending on how many cases need to be imputed). 

iv. As a fourth option, you can also impute values using means for subgroups.

v. Fifth, you can assume values based on a relationship you are fairly certain of.

vi. Sixth, you might use a different technique, such cluster analysis, which groups the dataset according to its ""good"" data and then uses the mean of the cluster to fill in the gaps left by missing data for cluster members.

vii. Seventh, there are specialized data management applications and publications on the topic that are designed for this reason (i.e. cleansing data). "