Reflecting a Variable in SPSS Anne administered a questionnaire which includes items that comprise a scale intended to measure contingent self esteem. Some of the items need to be reflected (reverse scored). Item value 5 needs to be changed to item value 1, 4 to 2, 3 unchanged, 2 to 4, and 1 to 5. The cumbersome way to do this in SPSS is to use the recode transformation. I’ll show you the easy way using the compute function. All you need to do is replace the item score (IS) with (k + 1)-IS, where k is the largest possible item value. In this case, you wish to transform IS to 6-IS. I illustrate using variable “cse1.” Look at the initial distribution of responses on cse1: cse1 Valid 1 Not at all like me 2 Somewhat unlike me 3 Neutral 4 Somewhat like me 5 Very much like me Total Frequency 6 32 131 171 53 393 Percent 1.5 8.1 33.3 43.5 13.5 100.0 Valid Percent 1.5 8.1 33.3 43.5 13.5 100.0 Cumulative Percent 1.5 9.7 43.0 86.5 100.0 In the data editor, click Transform, Compute. In the Compute Variable window, enter ‘cse’ as the target variable. As numeric expression, enter “6-cse1.” Click OK. When SPSS asks “Change existing variable?” indicate OK. Now look at the transformed distribution of cse1: cse1 Valid 1 Not at all like me 2 Somewhat unlike me 3 Neutral 4 Somewhat like me 5 Very much like me Total Frequency 53 171 131 32 6 393 Percent 13.5 43.5 33.3 8.1 1.5 100.0 Valid Percent 13.5 43.5 33.3 8.1 1.5 100.0 Cumulative Percent 13.5 57.0 90.3 98.5 100.0 And now the cumbersome way. From the data editor, Transform, Recode, Into Same Variables. Scoot ‘cse1’ into Numeric Variables and click Old and New Values: Enter Old Value Value 1, New Value Value 5, click Add. Old 2, New 4, Add, Old 4, New 2, Add, Old 5, New 1, Add. Continue, OK. SPSS Lessons Karl L. Wuensch, Dept. of Psychology, East Carolina Univ., Greenville, NC 27858 USA January, 2006