An Explanation of Statistical Tools from DocumentingExcellence.com A consulting practice focusing on working with colleges', organizations', and individuals' utilization of quantitative and qualitative assessment tools to analyze and document their quality outcomes through providing staff development, research design and analysis, and psychometric evaluations.
Analysis Procedures Conceptual Issues Competency Learning Domains Item Construction Item Assessment Scale Construction Measurment Issues Evaluating - Single variable Descriptive statistics Validity, Reliabiltiy, Cause & Effect Reporting Modeling Correlation Regression & Correlation Statistical Control Unemployment & Crime Model Reduction Outcomes Assessment FAQ Consulting Services

x

Exogenous Variables

There may be a number of variables that might be exogenous (outside, external) to the model of relationships being examined, which we would want to control for.  In the figure below the variables of sex, age, motivation and credits earned at other colleges have been introduced.

The cumulative effect of introducing these four variables is minimal, only 1% more variance is explained.  If explanation were all we were interested in, then these variable might not be included in a model.  However, each of them is included because they are important to building a comprehensive model of math skill acquisition for the point of view of theory and/or prior findings.

A few other changes of interest are the following:

The path from math courses to skill has increased slightly from .24 to .26.

The path from pre-test to skill has decreased from .61 to .55.  These two changes indicate that most of the effect of sex, age, motivation, and other college credits was included in the pre-test score. The sex variable is coded as 0 for male and 1 for females.  This the negative path from sex to skill indicates that females tend to have slightly lower math skills.  The age variable indicates that older individuals tend to have lower skills. The combined motivation score is included because motivation in taking a standardized test is a strong predictor of performance, and thus control for that motivation is helpful whenever possible. The other college courses variable indicates that the native college students tended to have lower math scores than those transferring in.

This completes the introduction of variables used to control for exogenous or external characteristics.

	A next step is to refine the math skill model
	Or examine an example where an exogenous variable explains it all

Send mail to with questions or comments about this web site.
Copyright © 2010 by Peter T. Klassen, Ph.D. Principal, www.DocumentingExcellence.com
9 January, 2010