Resources for Designing Studies, Selecting Appropriate Statistics and Presenting Results in Science Fair Projects |

OREGON CHAPTER
The American Statistical Association |

If you are a K-12 science fair student or faculty mentor, selecting appropriate statistics for your data may be cause for concern. Even though statistics and probability are part of our (Oregon) Curriculum Goals, this knowledge would be useful but not sufficient to be used with confidence. Some high schools offer a statistics course but, even then, application in a real world situation could be a problem. It may be comforting to know that graduate students and even professional researchers can find statistical treatment of their data a daunting task. But you do have an investigation, you are collecting data and, although you may not have had the [statistics] course(s), you need to make justifiable sense of that data. We have three suggestions. 1. Get a statistician. 2. Design the study well. 3. The WEB has resources on study design, selecting a statistical test, and data presentation. Get a statistician. Suggestion number one is to get a statistician. Unless there is a school faculty member who is facile with statistics it would be very helpful to get outside help. A volunteer might be found among parents of students or from other community resources such as nearby colleges or businesses who employ statisticians. Also, there may be statistically oriented personnel in the central offices of a very few school districts. The American Statistical Association (ASA) Center for Statistical Education sponsors a program called Adopt-A-School in which statisticians volunteer to become involved with a school working with teachers as they integrate statistics into the curriculum. These activities could include helping students with science fair projects deal with data and study design issues. ASA has a growing portfolio of resources supporting statistics in K-12 education. Their WEB site is www.amstat.org. The Oregon Chapter of ASA is supportive of the Adopt-A-School program and some members are available for direct support of K-12 education activities. Please contact one of the chapter officials for further information. Statisticians might be useful advising individual students or concentrating on helping teachers. They can be most useful early in the research planning when design issues are being discussed. Statisticians are useful but less so when brought in to help select a statistic after the data has been collected and design errors are already made. Design the study well. The study design is the structure or pattern that relates the research questions to the data collected and determines the inferences that can be made. The design is critical to the validity of any statistical test. Design decisions include questions like who or what to sample, where to get control or comparison data, whether to perform an experimental manipulation or observe an existing situation and so on. Design sets up the logical inferences that can be made from the data whereas inferential statistics tell us how much we can trust the data pattern we see. Descriptive statistics and graphs make patterns of results more apparent to researchers and their audiences. In addition to generic presentations of research methods (many of which are social science oriented) one might look for texts on research methods written from the perspective of the disciplinary area of your study. WEB Resources for Research Design, Selecting and Presenting Statistics A WEB search will produce copious numbers of resources for understanding research design issues and selecting statistics. The statistical selection sites usually have a tree structure set of questions for investigators to answer about their data which leads to a suggested appropriate statistic. The research design sites are more like textbooks. There are also WEB resources for graphing and charting. You are encouraged to use your browser to round up some of these resources. For starters you might try "selecting statistical tests" and "research design." The following are several sites that you may find useful. 1) WEB sites re: selecting statistical tests. http://www.graphpad.com/www/book/Choose.htm 2) WEB sites re: research design. http://spsp.clarion.edu/mm/RDE3/start/RDE3start.html 3) WEB site re: graphing and charting. http://www.bettycjung.net/Graphing.htm 4) WEB sites re: general statistical resources. The Cleveland site has science fair resources. American Statistical Association Education Department http://www.amstat.org/education/ Oregon Chapter of the American statistical Association |