Interest in and use of the System of Multicultural Pluralistic Assessment's (SOMPA's) Health History Inventories (HHI) has been intensive despite the unavailability of psychometric data estimating its stability. This paper reports the results of a longitudinal study designed to provide data on the stability of the HHI over a four year period. The 44 elementary school children studied were drawn from middle and lower class Anglo, Black, and Mexican-American families. Variance in scores is observed between the two administrations of HHI. Means change somewhat, test-retest correlations generally are moderate and lines of best fit for the two sets of data suggest change in scores. Although some variance may be attributable to health incidents which occurred between the two administrations, some variance also is attributable to respondent inconsistency as noted on questions that should receive the same answer across administrations. Despite this variance, "at risk" and "not at risk" classifications made from each Health History Inventory generally are stable. (Author)