A HOW -TO GUIDE !
+ Ensure another instrument does not exist
+ Select appropriate survey response format
+ Identify limits of the construct
+ Base items on previous instruments or standards found in literature+ Consider multidimensionality of the construct
+ Consider clarity and level of difficulty (Bloom’s Taxonomy; Bandura, 2006)
+ Build a large initial pool of items
ConsultationWith
Experts
Final ItemInclusion List
Document every stage,every item removed, and justify each step.
+ Are items grounded in theory?
+ Is the population of interest represented appropriately? + Transcribe interviews
to identify key issues.
+ Opportunity to use verbal probing to determine comprehension, retrieval, decision, and response processes of participants in relation to each item.
+ Internal Structure
+ Observed Correlation Analyses+ Latent Variable Analyses
Evidence for Test Content Evidence for Response Processes
Evidence for Internal Structure
Abigail M. A. Love (M.S.), Christen M. Dillon (M.M.),
Renae Mantooth (B.A.), & Ellen L. Usher (PhD)
+ Item Removal Criteria
(Bensen & Clark, 1982; DeVellis, 2012)
+ Create items to reflect underlying latent construct(s)
CognitiveInterviews
w/ SamplePopulation
Don’t skip steps!
(AERA, APA, & NCME, 2014)
(AERA, APA, & NCME, 2014) (AERA, APA, & NCME, 2014)
Evidence for Test Content(AERA, APA, & NCME, 2014)
Writing socially desirable items should be avoided.
Items should be age appropriate and differentiated.
Items should not be redundant, ambiguous, vague, or carry multiple meanings.
Relations to Other Variables(AERA, APA, & NCME, 2014)
p20motivationlab.org/research