Does Item Sequence Order Impact Local Item Dependence in Surveys?

• In Event: Poster Session 4
  In Poster Session: Survey Research in Education SIG Poster Session

Abstract

Many survey researchers consider it a best practice to group related items together as it makes it easier for participants to complete the survey, gives the appearance of greater cohesiveness, and requires a lesser cognitive load from participants (Bradburn, Sudman, & Wansink, 2004; Dillman, 2000). However, a considerable number of studies have found that item order effects, often called “assimilation effects” or “carry-over effects”, can result in biased participant responses (Heiman, 2002). This study sought to compare local item dependence (LID) by way of a quasi-experimental design using a sample of national respondents. Results indicate item ordering effects may result in false-positives for statistically dependent items.

Authors

• Kenneth Royal, North Carolina State University

• Kent Hecker, University of Calgary