Curriculum-based measures of computational skills: A comparison of group performance in novel, reward, and neutral conditions

Research has established that curriculum-based measurement (CBM) procedures yield outcomes that are sensitive to a variety of educational conditions including the level of skill acquisition, context of assessment, duration of assessment, format of probe construction, and relative motivation of the target student. This study was designed to compare performance on multiple-skill mathematics CBM (M-CBM) across three assessments conditions: novel, reward, and neutral. A novel assessment condition was established when the examinee had not been exposed to the assessment conditions within temporal proximity of an administration. A reward condition was established when the examinee received contingent access to desirable stimuli or events dependent on performance. Finally, the neutral condition was established when the conditions were familiar (i.e., not novel) and there was no contingent access to rewards (i.e., not reward). Ninety elementary students across four grades (second grade, n = 24; third grade, n = 25; fourth grade, n = 14; fifth grade, n = 28) were assessed on three consecutive days. Results of repeated measures analysis of variance and follow-up planned analysis of main effects suggested that the median level of M-CBM performance on the first day of assessment (novel condition) was reduced by a statistically significant amount in comparison to either reward or neutral conditions. The difference approximated 3 digits correct per minute. There were no significant differences in performance between neutral and reward conditions. These outcomes suggest that M-CBM is an indicator of optimal performance after the initial day of assessment. Implications and limitations are discussed.