2 things need to do :
1) Evaluate the amount of missing data - <5%missing> 5% missing
2) Evaluate the pattern of missing data - MCAR or MAR; could use Little's MCAR test
Then, need to think about alternative ways to impute data. Have many methods of data imputation. Some like to use mean substitution. In fact, this is the common method use to impute data but care should be taken as this method will somehow reduce the variance of the variable. As such, if we decide to impute values using this method, we might need to do extra checking whether there is mean difference between the original and adjusted variable. This could be done by applying paired sample t-test. Alternatively, we could use Maximum Likelihood Substitution /EM in the SPSS.