Matching Familiar Figures Test Norms Based on IQ

Abstract

The Matching Familiar Figures Test (1) has been the most commonly used instrument to assess the dimension of reflection-impulsivity in children. Subjects are typically labeled by use of a median split for time to first response (latency) and for total number of errors. Subjects labeled as reflective score above the median on response time and below the median on errors, while those who score in the opposite direction are labeled impulsive. A problem with this method is that a child's classification could change as the comparison group changed. Norms have been provided based on thousands of subjects which, if used, would obviate this problem ( 3 ) . The present study compared responses of a single sample to the established norms and added IQ for classification purposes. Salkind's norms do not take IQ into consideration, although there is evidence that matching scores are influenced by IQ ( 2 ) . Thc subjects were 104 8-yr.-old children (42 boys and 62 girls) from a lower middle-class group. They were administered the marching task and the Otis-Lennon Mental Ability Test. Analysis showed IQ significantly correlated by Pearson's method with both latency ( r = .22, p < .05) and errors ( r = .34, p < .01). The latency-error r was -.59 ( p < .01) which is comparable to Salkind's correlation of -.54. Mean number of errors was 13.36 and mean latency was 8.79 sec., which contrasted with Salkind's mean errors for 8-yr.-olds of 12.39 and mean latency of 13.55. A number of children would have been classified differently in this study without referral to the norms. Most interesting, however, was the classification according to IQ. Group I (IQ 85) had mean latencies and errors of 9.13 and 15.00, respectively. This group might be called slow-inaccurate. Group I1 ( IQ 8 6 1 0 0 ) had mean latencies and errors of 7.44 and 15.65 and might be referred to as fast-inaccurate or impulsive. Group I11 (IQ 101-115) had mean latencies and errors of 8.96 and 11.72 and might be referred to as fast-accurate. Group IV ( IQ I 116) had mean latencies and errors of 9.64 and 11.09 and might be referred to as slow-accurate or reflective. It is imponant to refer to norms based on a large sample when classifying children along this dimension, but it is proposed that exist~ng norms be modified to take IQ into consideration.