Designing and implementing a qualitative evaluation protocol for non-credit lifelong learning programs

Dennis L McElhoe, Purdue University

Abstract

The purpose of my study was to determine whether an evaluation model employing multiple methods of data collection and analysis might yield more useful information for improving non-credit lifelong learning courses than previous models, most of which have used variable analytic research designs and quantitative methods almost exclusively. Based on Tashakkori and Teddlie's (1998) Dominant-Less Dominant Mixed Method Design, the study's dominant component was qualitative in nature and involved collecting interview, observational, and archival data. These data were analyzed using both the constant comparative method of grounded theory analysis (e.g., Strauss & Glaser, 1967) and discourse analysis (e.g., Gee, 1999). The less dominant component was quantitative in nature. Data from students' course surveys were collected and analyzed using both descriptive and inferential statistics. In all, 30 adults enrolled in two different 10-week computer software-training courses at a community college participated in the study. The study's two major findings were: (1) the most satisfying way adults learn appears to be dependent on the instructional environment that they may be most comfortable with. Older students were provided with a formal, subject-oriented environment, while their younger counterparts were provided with a more self-directed, task-oriented environment; (2) although students from both courses reported that they had achieved their learning objectives, the confidence they gained in using computers, rather than the actual skills they learned, was the greatest benefit they derived from their participation. This study was one small attempt to develop and implement an evaluation model using a mixed-methods design that might be more useful than more typical quantitative designs for generating information that is useful for improving non-credit lifelong learning courses. Although the study suggests that such a model holds promise for accomplishing this goal, more research is needed to make this claim with confidence. Particularly important would be studies with much larger samples, as well as studies within a more diverse range of learning environments and learning activities.

Degree

Ph.D.

Advisors

Peters, Purdue University.

Subject Area

Adult education|Continuing education

COinS