A longitudinal study of the utility of social preference and social impact sociometric classification schemes

The stability of standard score and probability method sociometric group assignments was examined over a 2-year period with an initial group of 334 preadolescents. The popular, neglected, and controversial sociometric groups evidenced low stability of group membership over intervals of approximately 1, 6, 12, 18, and 24 months; the rejected group evidenced slightly higher short-term stability. These findings of limited stability were attributed to measurement error and to the failure of both classification systems to identify groups with homogeneous social reputation profiles. Social role scores contributed to the prediction of stable group membership in the rejected and controversial classification, although these scores added little to the prediction of stable popular and neglected group membership. Stability over short intervals could be used to enhance the prediction of stability over longer periods; however, this procedure resulted in the classification of numerous false positives and false negatives. The instability of sociometric group assignments completed with the standard score and probability methods indicates that researchers should be cautious about the use of classifications based on only one data collection and that the selection of children for both clinical intervention and further nomothetic research may require alternative assessment procedures.