Construct Validity PUSH CONSTRUCT VALIDITY Factor analysis Bivariate
Construct Validity PUSH
CONSTRUCT VALIDITY Factor analysis: Bivariate correlations Improv. Learn. Recall Confir. Reass. No imp. Frust. too much Frust. not enough Disag. Harmful Improv. 1. 00 0. 17 0. 05 -0. 02 -0. 21 -0. 02 -0. 07 -0. 04 -0. 03 Learn. 0. 17 1. 00 -0. 13 -0. 30 -0. 20 -0. 47 -0. 02 -0. 11 -0. 08 -0. 06 Recall 0. 05 -0. 13 1. 00 0. 09 0. 08 -0. 13 -0. 01 -0. 03 -0. 02 -0. 01 Confir. -0. 02 -0. 30 0. 09 1. 00 0. 21 -0. 02 -0. 07 -0. 04 -0. 03 Reass. 0. 02 -0. 20 0. 08 0. 21 1. 00 -0. 18 -0. 02 -0. 06 -0. 03 No imp. -0. 21 -0. 47 -0. 13 -0. 21 -0. 18 1. 00 -0. 02 -0. 07 -0. 04 -0. 03 Fru. too much -0. 02 -0. 01 -0. 02 1. 00 0. 03 0. 00 0. 01 Fru. not enough -0. 07 -0. 11 -0. 03 -0. 07 -0. 06 -0. 07 0. 03 1. 00 0. 04 0. 06 Disag. -0. 04 -0. 08 -0. 02 -0. 04 -0. 03 -0. 04 0. 00 0. 04 1. 00 0. 16 Harmful -0. 03 -0. 06 -0. 01 -0. 03 0. 01 0. 06 0. 16 1. 00
CONSTRUCT VALIDITY Factor analysis (Principal component analysis) Best solution = 8 factors 1. Practice Improvement 2. Learning* 3. Reassurance, Confirmation 4. Recall 5. Frustration not enough info 6. Frustration too much info 7. Disagree 8. Harmful Total variance explained = 89. 8% *No impact
CONSTRUCT VALIDITY Multilevel Factor Analysis • Accounting for dependence due to multiple ratings from each participant • Accounting for dependence due to multiple ratings for each Info. POEM Best solution = 8 factors 1. Practice Improvement 2. Learning 3. Reassurance, Confirmation 4. Recall 5. Frustration not enough info 6. Frustration too much info 7. Disagree 8. Harmful 9. Harmful Total variance explained = 90. 0% Best solution = 9 factors Total variance explained = 97. 5%
Results - Summary • Use of the method is good • The ten most frequent impact patterns account for 89% of all reports of ‘cognitive impact’ • An eight factor solution explains 90% of total variance, and is a good fit for the data
- Slides: 5