www.mangold-international.com - Tuesday, 6. January 2009
Printversion from "Reliability"
(URL: http://www.mangold-international.com/23.html)
Measuring inter rater reliability
INTERACT offers some very useful and easy to use operations, such as the calculation of inter rater reliability, using Cohen's Kappa. Select any observation to be compared with any other observation (e.g. from a different coder) and let INTERACT calculate the congruence between the analysis sessions.
Really fabulous in INTERACT is the fact, that you can tune the method of inter rater reliability calculation to your own needs:
- An observation can be compared with any other!
- If you are not interested in what was coded prior or after a certain point in time, you just select the time frame in which the data shall be compared. This allows you to focus the measurement on certain behavioral events.
- You are only interested in the comparison of certain behavioral codes? No problem, simply select the codes to be taken into calculation.
- You are working with mutual excluding codes and at the same time with not mutual excluding ones? No problem, just define, how INTERACT shall treat them.
- Best of all: Are you interested in "How is the inter rater reliability, compared in 1 second intervals?" and "How is it down to single video pictures or in 5 minute intervals?". This is really easy: Just select the appropriate time accuracy and press OK. That's it!
INTERACT will calculate every behavioral category in one single pass within a few seconds. All included in the standard INTERACT version.
- Some of the parameters that can be set:
- Snapshot of the final report:
Mangold-International 2006-2009