Type of Document	Dissertation
Author	Colón, Marilin
Author's Email Address	mcolon@bmcsoutheast.com
URN	etd-04112006-171605
Title	Improving the Reliability and Validity of Visual Inspection of Data by Behavior Analysts: An Empirical Comparison of Two Training Methods to Improve Visual Inspection and Interpretation, The Job Aid and the Conservative Dual-Criteria
Degree	Doctor of Philosophy
Department	Psychology, Department of
Advisory Committee	Advisor Name Title Jon S. Bailey Committee Chair Bruce Thyer Committee Member Bryan Loney Committee Member Frank Johnson Committee Member Robert Contreras Committee Member
Keywords	Graphing Visual Analysis Job Aid
Date of Defense	2006-03-27
Availability	unrestricted
Abstract The present study compared the effects of two job aids on the behavior of visual data analysts. Twelve participants were randomly assigned to a CDC Lines Group, an Active Job Aid group, or a Control group and viewed a series of 56 within-subject behavioral graphs. They were asked to determine whether the graphs depicted a change that was attributable to the intervention. Results indicated that participant answers in the CDC Lines group unanimously agreed with the answers determined by a panel of experts of the graphs for 92% of the graphs during the intervention condition. For the Active Job Aid group, participants unanimously agreed with the experts for 73% of the graphs during the intervention condition. Unanimous agreement was 50% for the Control group. For dichotomous agreement with the experts, the CDC Lines group averaged 94% and the Active Job Aid group averaged 81% agreement with the experts during the intervention condition. Dichotomous agreement was 72% for the Control group. The CDC Lines Job Aid decreased both Type I and Type II errors during the intervention condition, as compared to baseline and maintenance conditions. The Active Job Aid decreased Type I errors, but increased Type II errors, as compared to baseline and maintenance conditions. The CDC Lines Job Aid improved agreement with the experts for all participants. The Active Job Aid improved agreement with the experts for only two out of the four participants. Participants in groups were asked to identify the graph characteristics that most influenced their decision about intervention effect for each graph. The job aids did not increase the ability of the participants to correctly identify these characteristics. It is argued that the results obtained have important implications for the training and credentialing.
Files	Filename       Size       Approximate Download Time (Hours:Minutes:Seconds)     28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access    dissertation2b.pdf 7.48 Mb 00:34:38 00:17:49 00:15:35 00:07:47 00:00:39