Sunday, February 20, 2011

Non response and Measurement Error in Employment Research – Making Use of Administrative Data

Public Opinion Quarterly, Vol 74, No 5, 2010 pp 880-906


 

-          Household surveys face threat of non response bias if respondents differ systematically from no-respondents

-          In absence of good non response bias estimates

-          Survey researchers try to maintain high response rates as safeguard against bias

-          By increasing efforts to recruit respondents

-          Such efforts are costly

-          May not reduce non response bias.

 

Practical strategies


-          Increase number of calls

-          Change mode of data collection

-          Refusal conversion strategies

 

-          Lack of motivation may lead later / reluctant respondents to satisfice

 

 

No comments: