Questions and lessons can't be printed or saved as a PDF.
Whose agreement is being measured, by which statistic?
More, if you want it
What is the patient's score compared with?Tap a term: its definition, what you can do with it, and how often the exam asks it.
Definition: chance-corrected agreement between raters on categorical ratings.
You can: Tell kappa from the intra-class correlation coefficient by whether the ratings are categories or continuous scores.
Definition: agreement and reliability between raters on continuous scores.
You can: Tell the intra-class correlation coefficient from kappa by whether raters give continuous scores.
Definition: a measure of how strongly the items of a scale correlate with each other.
You can: Tell Cronbach's alpha from split-half reliability by whether all items or two halves are correlated.
Definition: the correlation between scores on two halves of one test.
You can: Tell split-half reliability from inter-rater reliability by whether items or raters are compared.
Definition: agreement of a measure with a gold standard or a later outcome.
You can: Tell validity from inter-rater reliability by whether a gold standard or later outcome is used.
Each card is one term. The grey line is what the patient says, which is what the question stem gives.
Summary
Inter-rater reliability is present when two or more observers independently give similar ratings of the same patient, scan or trainee.
is present when two or more observers independently give similar ratings of the same patient, scan or trainee. Reliability means a measure gives consistent results.
Inter-rater agreement is a standard way to assess reliability. Internal consistency and test-retest are others. A scale can show excellent reliability between raters while its internal consistency stays limited. Candidates read raters agreeing as proof that a scale is valid. Rater agreement only shows that the raters are consistent.
Philippa Okafor is a consultant on a College working group testing a new trainee rating scale. Two consultants rate one trainee's video separately. Philippa says, "Both consultants gave nearly the same ratings."
Different raters or respondents agreeing on one subject show inter-rater reliability.
5 more sections on Reliability statistics: inter-rater reliability, kappa, intra-class correlation, Cronbach's alpha and more
Log in free to read the rest: the full mental model, exam recognition, and the look-alikes that trip people up. Your daily session is free, always.
mrcpsych.co.uk · An independent study tool · not affiliated with the Royal College of Psychiatrists