09.12.2012 Views

2003 IMTA Proceedings - International Military Testing Association

2003 IMTA Proceedings - International Military Testing Association

2003 IMTA Proceedings - International Military Testing Association

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

310<br />

USING DECISION TREE METHODOLOGY TO PREDICT ATTRITION<br />

WITH THE AIM<br />

Wayne C. Lee<br />

Department of Psychology<br />

University of Illinois at Urbana-Champaign<br />

603 East Daniel Street<br />

Champaign, IL 61820<br />

wlee@s.psych.uiuc.edu<br />

Dr. Fritz Drasgow<br />

Department of Psychology<br />

University of Illinois at Urbana-Champaign<br />

603 East Daniel Street<br />

Champaign, IL 61820<br />

fdrsagow@s.psych.uiuc.edu<br />

This paper describes the Assessment of Individual Motivation (AIM) and efforts to use it<br />

for predicting first-term attrition in the United States Army. This description provides a context<br />

for the three papers presented in this symposium, including this one where the results of one<br />

investigation are presented. In this first investigation, a non-linear, “configural” approach to<br />

prediction is applied to examine whether we can improve on linear methods used to determine<br />

the predictive validity of the AIM with respect to attrition in a 12-month time interval.<br />

The AIM<br />

Attrition is among the most studied of the organizationally relevant outcomes among<br />

personnel researchers. One estimate as early as 1980 put the number of articles and book<br />

chapters devoted to attrition between 1500 and 2000 (Muchinsky & Morrow, 1980). Certainly,<br />

the popularity of this topic is due in part to the high cost associated with attrition in<br />

organizations. Earlier research conducted by the U.S. Army Research Institute for the<br />

Behavioral and Social Sciences (ARI) with the Assessment of Background and Life Experiences<br />

(ABLE) suggested that temperament measures might indeed be good predictors of attrition in the<br />

U.S. Army. Unfortunately, as with many temperament measures, concerns regarding the<br />

potential effects of faking and coaching restricted the implementation of the ABLE in new<br />

recruit selection (White, Young, & Rumsey, 2001).<br />

Beginning in the mid-1990’s, this line of research was continued formally by ARI with<br />

the development of the AIM. This new measure was designed specifically to target first-term<br />

attrition while also being less susceptible to faking and coaching. The AIM is comprised of 27<br />

items in a forced-choice format measuring 6 constructs. While “27 items” may seem like a small<br />

number of items for any measure of multiple constructs, it is important to understand that each<br />

item is comprised of four descriptive stems. Each of these stems –108 in all—could very easily<br />

be presented as a single-item in Likert-type format. For each item in the AIM, two stems are<br />

worded negatively and two are worded positively –representing low and high levels of a<br />

particular construct, respectively, if a respondent endorses them.<br />

45 th Annual Conference of the <strong>International</strong> <strong>Military</strong> <strong>Testing</strong> <strong>Association</strong><br />

Pensacola, Florida, 3-6 November <strong>2003</strong>

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!