ASOC 220 Lecture Notes - Lecture 36: Survey Data Collection, Response Bias, Stata
2 Jun 2017
School
Department
Course
Professor
Document Summary
Someone is 32 but has lived for 45 years at their address! Straight-lining when a respondent gives the same response without reading the question. * checkbox questions too many na , other , none of the above . * data cleaning is never a cure for poor study design. * may be needed to maintain integrity of your data/study. * data monitoring detection of error patterns as they occur. * don"t over think or try to interpret response. * don"t change the question after the fact if responses seem to not make sense. * can editorialize in your analysis but take care to not change questions and/or responses it can be a fine line. * describe what you"ve done but there"s no standard on how to report on editing. * field checking and proper supervision of data collection. Impossible values change if you know what correct value should be otherwise, delete the case.
