HLM_Problem with Missing Values

I am using multilevel modeling to analyze data (HLM) for my dissertation. The problem I ran into is with importing the data into HLM. The data set has 26,000 observations. It's students nested within programs that is nested within institutions. I removed all the missing values in STATA before importing it into HLM. After removal of all the missing values, the data set was cut down to 20,000 observations. However, when I tried to create an HLM data set, there was only 2,500 observations imported at level one. I just can't figure out how to solve this issue.

Any idea where I might have done wrong?