Salt Lake City, Utah
June 20, 2004
June 20, 2004
June 23, 2004
9.1088.1 - 9.1088.6
Selecting an Appropriate Statistical Test for Research Conducted in Engineering/Graphics Education: A Process
Alice Y. Scales, Julie H. Petlick North Carolina State University
Abstract Individuals in institutions of higher education who are involved in research on teaching engineering graphics, and other projects, are frequently confounded by the process of selecting the appropriate statistical test to analyze the data they collect. Research studies are usually only a portion of faculty member's work, and they generally only have taken one or two required courses in statistics during their graduate work. For these reasons, they either have to consult with a statistician or may select a test that is inappropriate for the research they are conducting.
This paper will layout a flow chart for selecting appropriate statistical tests based on the nature of the data being collected for a study. Examples used in the paper will use situations that might be encountered by individuals conducting research in teaching engineering/technical graphics.
Introduction At the heart of statistics is answering questions with data. According to Finnney, statistics is “concerned with finding out about the real world by collecting, and then making sense of data” (p. 164).1 However, statistical findings are only meaningful if the data are analyzed by using an appropriate statistic test. When the test is wrong, then the data analysis may be meaningless. Individuals in our field involved in research have frequently had a problem when it comes to the selection of a statistical test. There are several reasons for this difficulty.
Issues Underlying Test Selection The first issue that creates problems in selecting an appropriate test is there is more to it than meets the eye. The form of the data, sample size, sample distribution, test power, and test robustness are all part of the equation of test selection. The second issue is that most researchers usually only have taken the prerequisite one or two required graduate statistics courses, which are designed to make them low-level practitioners of statistics, not experts in the field. These courses primarily teach the technical aspects of using the statistical formulas and often have developed a reputation for being difficult, mechanical, and boring. Where they fail, even in the eyes of many teachers of statistics, is in the larger area of research design. According to Wild, “The process of investigation as a whole should be the heart of any statistics program.”1 Most graduate level courses are about calculations rather than role of statistics in an investigation.1,2 In the courses attempt to cover the technical side of statistics, a clear pattern of how and why “Proceedings of the 2004 American Society for Engineering Education Annual Conference & Exposition Copyright © 2004, American Society for Engineering Education"
Petlick, J., & Scales, A. (2004, June), Selecting An Appropriate Statistical Test For Research In Engineering/Graphics Education: A Useful Procedure Paper presented at 2004 Annual Conference, Salt Lake City, Utah. 10.18260/1-2--13372
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2004 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015