art photo



How does one handle weights in a longitudinal analysis with differential sample attrition over time, etc? I am using selected years in the NLSY.


Weights allow one to estimate a nationally representative population. However, with attrition the weight established for the first wave may not be applicable over time. Because of the complexity of the sample, NLS creates a cross-sectional weight for each round. If you are using data from just one survey the weight series provides an accurate adjustment. However, if you are using data from multiple survey years, no single set of survey weights provides an accurate adjustment.

If your research spans multiple survey years, you can create a custom set of survey weights which adjust both for the complex survey design and for using data from multiple years. To do this, use the custom weight generator developed by the NLS:

Annotated Resources:

Direct Links:

Related Question Groups: