CPSC 689/602--Spring 1997
Computer-Supported Cooperative Work

Paper Report

Title: : Methods for the Study of Groups (Excerpt)
Authors: J. McGrath

Citation: From Groups: Interaction and Performance, pp. 28-37, 1984.
Reprinted in: Groupware and Computer-Supported Cooperative Work, pp.
200-204, Edited by R.M. Baecker.


Learning that occurs in science is based on some combination of empirical and theoretical methods. All of the different methods used to gather and analyze information have inherent advantages and disadvantages. McGrath suggests that researcher can utilize carefully selected multiple methods to offset the inherent weaknesses thereby adding strength to the resulting evidence if they show consistent outcomes across divergent methods.

Research in the social and behavioral science always involves somebody doing something in some setting. When you design a research study you are always trying to maximize three things:

For the comparison relational question to be valid you are manipulating, controlling, or matching on a certain limited set of variables. In order to assure that all the other variables do not affect the outcome you must use a process of random assignment of cases to conditions (randomization).

A study should have a high validity in regard to four different types of validity questions.

An important aspect of the different forms of validity is that they an increase in one measure does not necessary result in a decrease in other measures. Procedures which increase internal validity will have a positive impact on statistical validity but will probably decrease external validity. In conducting research studies one should always be aware of the strengths and limitations of the method(s) used in interrupting results.

Report prepared by: Gary Baker Email: sadiki@tamu.edu
Discussion date: 2/3/97            Report date: 2/5/97