Imputation for missing observations panel data

[Victoria Zorzi]

Hi

I do not have any experience with using imputation to fill out missing observations. After having read about various reasons behind possible missing values, I think that the observations in my dataset are missing at random.

The data I am using is countries in Subsaharan Africa in the period from 2000-2014. The variables with missing observations are CPI (Corruption Perception Index) and PolityIV (The PolityIV democratic index).

Most of the missing observations are in certain countries in the early period of my dataset. Fx Benin is missing CPI observation from 2000-2003.

How do I deal with this using imputation?

. misstable summarize
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
           CPI |       127                 518  |     55         10          65
      PolityIV |        33                 612  |     18         -7          10

Thank you in advance.

[Phil Bromiley]

You will increase your chances of useful answer by following the FAQ on asking questions-provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

There is both the multiple imputations and a maximum likelihood approach to handling missing data. Maximum likelihood approach is explained in the SEM documentation. However, exactly how much these help is a matter of some discussion. If the data are missing at random, then you should still get consistent estimates it with the data ignoring the missing issue.