home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!KEAN.UCS.MUN.CA!PGILL
- Message-ID: <0095E338.2ECF50A0.21242@Leif.ucs.mun.ca>
- Newsgroups: bit.listserv.stat-l
- Date: Mon, 27 Jul 1992 11:53:16 -0400
- Reply-To: pgill@kean.ucs.mun.ca
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- From: pgill@KEAN.UCS.MUN.CA
- Subject: Is it good agreement ?
- Lines: 55
-
- A researcher has developed a questionaire containing 25 questions.
- Each question refers to an action in a given situation. The
- respondent would answer whether the action is appropriate (1)
- or not (0). Before sending the questionaire to the field, the
- researcher wanted to test the questionaire on a panel of 5 experts
- including herself. The following are the responses of the experts,
- expert 1 being our researcher :
-
- Expert \ R E S P O N S E S TO Q U E S T I O N S
- _____________________________________________________________________
-
- 1 0 0 1 0 1 1 0 1 1 0 1 1 0 0 1 1 0 1 0 0 0 1 1 0 1
-
- 2 0 0 1 0 1 1 1 1 1 0 0 1 0 0 1 1 0 1 0 0 0 1 1 0 1
-
- 3 0 0 1 1 1 1 0 1 1 0 1 1 0 0 0 1 1 1 0 0 0 1 1 0 1
-
- 4 0 0 1 0 1 1 0 0 1 0 1 1 0 0 1 1 0 1 0 0 0 1 1 0 1
-
- 5 0 0 1 0 1 1 0 1 1 0 1 1 0 0 1 1 0 1 0 0 0 1 1 0 1
- _____________________________________________________________________
-
- The following table the number of times (out of max 25) a pair of
- experts agree on the appropriateness of the action :
-
- EXPERT 1 2 3 4 5
- --------------------------------------------------------
- 1 - 23 22 24 25
- 2 - 20 21 23
- 3 - 21 22
- 4 - 24
- --------------------------------------------------------
-
- It is reasonable to assume
- Prob{ two given experts agree on a given question by chance} = 0.5.
- The Binomial probablity of more than or equal to 20 (pure chance)
- agreements between two given experts is (approximatley) 0.0026.
-
- QUESTION 1. Does there seem to be a good agreement among the five
- experts ? Moreover, all five experts agree 19 out of 25 times.
- The Binomial probablity that all five experts agree by chance on
- a given question is (.5)**5 = 0.03125. Therefore, the probablity
- that all five experts agree (by pure chance) 19 or more times out
- of 25 times is almost 0.
-
- QUESTION 2. The researcher tried to use Cohen's kappa as an
- indicator of agreement among the experts and found it to be
- negative. Is Cohen's kappa approprite for binary agreement ?
-
- Any comments are welcome.
-
- P. S. Gill
- Dept of Maths & Stats
- Memorial Univ of Newfoundland
- St. John's CANADA
-