Published as: Crowder, Todd A., and Stephen E. Erfle. Focal Behavior and Performance By Male U.S. Service Academy Cadets: A Preliminary Analysis. In Goal Setting and Personal Development: Teachers' Perspectives, Behavioral Strategies and Impact On Performance , edited by Bernice Higgins, 201-20. New York: Nova Science Publishers, 2015. This author pre-print is made available on Dickinson Scholar with the permission of the publisher. For more information on the published version, visit Nova Science Publisher's Website.

Background . (Erfle & Gelbaugh, 2013) and (Erfle, 2014) examined a regular irregularity in physical activity performance histograms for curl-ups and push-ups from a sample of more than 9,000 middle-school students. These histograms showed periodic spikes at 5 and 10 unit intervals. They showed that students who used focal counting on one event were more likely to do so on another event, or on the same event at a later assessment. They also found that students who ended at these focal endings outperformed their non-focal peers on these fitness tests. They found that males were more likely to be focal than females. Methods. This chapter examines focal proclivity and performance using two Male U.S. Service Academy Cadet convenience sample datasets. One dataset is of 520 Cadets who did at least one pull-up, push-up, and sit-up; the other is of 301 Cadets who did at least one cadence pull-up, (CPU), and one 175-pound bench press repetition, (BP rep). Results. Neither dataset exhibits the pronounced regular spikes seen in the middle-school data. We therefore expand our analysis to counting by bases other that 5 and 10. The BP-CPU dataset exhibit focal counting by 2, 4, 6 and 8 for BP reps and by 5s for CPU. There is limited evidence of focal counting in the Pull-Push-Sit datset with sit-ups being the only event with significant focal counting (by base 5). Pull-up performance by those who count by 4s have higher pull-up performance than those who do not count by 4s (ΔM = 0.98, p = .01). Those who count BP reps by 2s and 4s significantly outperform those who do not on both BP reps and CPUs. Regression models suggest the couting BP reps by 6s leads to 2.91 more BP reps, p < .001, all else held constant.