Comparison of Data Quality Indicator Effectiveness With Unidirectional Versus Bidirectional Scales in Survey Research

• In Event: Data Consideration in Survey Research

Abstract

Survey data collected via online panel aggregators were evaluated for data quality using five low cognitive effort (LCE) data indicators. Using these five data quality variables, two scales were investigated: one with unidirectional item wording and one with bidirectional item wording. Internal consistency reliability was compared for each scale using three sample groups: a complete sample prior to data quality checks, subgroups identified as exhibiting LCE, and subgroups not identified as LCE. Results indicate that random response indicators can be effective for unidirectional and bidirectional item sets, however straightlining and response time indicators are more effective for bidirectional items. Overall, the use of quality data indicators were more effective in identifying aberrant response patterns for bidirectional item sets.

Authors

• Ronna C. Turner, University of Arkansas

• Jennifer Reimers, University of Arkansas at Fayetteville

• Wen-Juo Lo, University of Arkansas

• Kristen Jozkowski, Indiana University

• Brandon Lee Crawford, University of Arkansas