Previous analysis of rotation group bias in the Current Population Survey has concluded that if the biases are additive, the ratio and composite estimators of month-to-month change in unemployment are unbiased. This article shows that if the biases contain a multiplicative aspect, both estimators of change are then biased. The article also presents some empirical results that cast doubt on the validity of a purely additive model.
No takes yet. Share an insight, caveat, or question.
Gary Solon (1986) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: