Table 3.
Psychometric Properties for the Attitudes Toward Medical Statistics Scale
Property | Measure Used | STAT-Interest | STAT-Confidence |
---|---|---|---|
Item nonresponse | Proportion not answering question Mean (range) | 0.08% (0% to 0.4%) | 0.4% (0% to 1.3%) |
Score distribution | Mean score (standard deviation), median and range | 68 (SD=17) Median 70 Range 15 to 100 | 65 (SD=19) Median 65 Range 11 to 100 |
Test–retest repeatability | Correlation of scores at test and retest 2 wk later (n=220) | Pearson r=.60 | Pearson r=.62 |
Internal consistency reliability | Cronbach's α (goal is 0.7 to 0.8; lower value suggests lack of coherence as a scale, and higher value suggests that items are redundant) | α=0.70 | α=0.78 |
Content validity | Initial questionnaires revised based on feedback from experts in education, statistics, and cognitive psychology | ||
Construct validity (interest scale only) | “Next time I have to make a decision about my health, I plan to…” (%yes) | ||
Interest in statistics score | Ask my doctor for statistics (%)* | Track down statistics (%)* | |
Lowest quartile (15 to 54) | 46 | 19 | |
2nd quartile (55 to 69) | 81 | 49 | |
3rd quartile (70 to 79) | 96 | 68 | |
Highest quartile (80 to 100) | 97 | 80 |
P<.001 (χ2).
r, correlation coefficient.