Hi, i calculate some calculations for a specific target variable. People should rate for 3 things some independends and one dependend variable. People should not difference in importances as possible between the 1.,2.and the 3.Chose. Now i filter all people which rate the same thing (i.e. a Product/Company) and calculate the different correlations dependend from the thing is rated as 1./2./3. or all togehter!
Not a surprise for me are the different values in correlation, but make the use of correlations for this doubtful. What is more a cause for this results'S ? (1) The small N's (2) The social-psychological phenomen that people want prefer something and not really indifferent when they ask for 3 rating's relative to a product/company. Result: So people who rate a product/company as 1. and others as 2. or 3. ( what is not a ranked order but it works neverthless in consciousness ) must be different analyze, because their exist different attitudes ? r N r N r N r N 1.Chose 2.Chose 3.Chose Cummulate 1. to 3.Choose 0.53 28 0.66 23 0.19 20 0.43 71 0.13 28 0.28 22 0.41 22 0.25 72 0.37 23 0.34 24 0.17 22 0.28 69 0.50 20 0.02 17 0.48 17 0.35 54 0.55 27 0.54 22 0.44 20 0.52 69 0.14 21 0.20 19 0.15 16 0.21 56 0.13 27 0.10 22 0.26 20 0.17 69 0.75 24 0.38 21 0.28 18 0.51 63 0.51 23 0.36 18 0.57 18 0.50 59 0.83 26 0.46 22 0.29 20 0.57 68 Thanks for any suggestion, alternatives and regards, Christian ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =================================================================