An Investigation of How Variables Impact Accuracy of Data Collection

Abstract

Assessment of accuracy should be conducted ongoing throughout research and intervention. This study aimed to identify how the number and frequency of behaviors effected measurement accuracy. Participants were students attending a four-year institution in the Mid-West. Two participants were undergraduate students with no prior data collection training and one participant was a graduate student with three years of experience as a Registered Behavior Technician. Participants were trained to identify and collect data on nine target behaviors. Participants watched and recorded target behaviors for six, 10-min. videos. The predetermined behavior and frequency occurrence assigned to each of the six, videos were: 1) three behaviors occurring two times each, 2) three behaviors occurring five times each, 3) six behaviors occurring two times each, 4) six behaviors occurring five times each, 5) nine behaviors occurring two times each, and 6) nine behaviors occurring five times each. Accuracy was identified by comparing participant data to true value observer data agreed upon by two experienced observers. Results indicated the number of behaviors being recorded had no effects on accuracy of data collection. During the initial and maintenance assessments, two participants measured behavior with lower accuracy when measuring conditions with five occurrences than when measuring two occurrences per behavior. One participant measured with lower accuracy when measuring two occurrences during the initial assessment, but measured less accurately when measuring five occurrences during the maintenance assessment. This indicated the frequency of behavioral occurrences may have had an effect measurement accuracy.