SCIENCE FAIRS AND STATISTICS

Joe Ward
Health Careers High School

Newsletter for the Section on Statistical Education
Volume 4, Number 1 (Winter 1998)

How might statisticians help high school science fair participants? It is suggested that statisticians should: (a) assist high school students with statistical design and data analysis techniques that will enhance the quality of their research projects; and (b) provide students with information about the relationship between winning statistics special awards and winning specific science-category awards. These two activities are essential for "selling" science fair students on the value of using quality statistics in their research projects.

Each year Amstat News includes descriptions of the statistics special award winners at the International Science and Engineering Fair (ISEF).

These projects use statistical design and analysis techniques that are beyond the scope of most first high school non-calculus-based statistics courses. Topics mentioned include two-factor ANOVA, multiple comparisons, Newman-Keuls pairwise comparisons, factor analysis, response surface models, and many uses of multiple regression. At the 1997 ISEF the statistics special awards winners used two-factor ANOVA, Scheffe Multiple Comparisons, response surface models, and multiple regression models.

Those who have judged in local and state fairs can observe that those projects that use the combined power of statistics and computer software frequently receive high recognition. Statisticians can use these examples to motivate science fair students to learn to use advanced analysis techniques to increase their chances of recognition for a statistics special award.

Casual observation of the ISEF statistics winners' names and their recognition in specific science-fair categories, indicates possible association between statistics "winners" and other-category "winners". Statisticians, above all, should use their own data analysis competence to study the association between the use of "quality" statistics in projects and receipt of specific science-category awards. This should be done for not only the ISEF but for each local and state science fair for which there are statistics special awards. While "causation" may not be possible it is appropriate to study "association". Consider two extreme scenarios for a science fair:

(a) Each statistics special awards winner wins FIRST PRIZE in their own specific science category. In this situation it might be argued that "being a statistics winner is 'HIGHLY associated' with winning in a specific science category". (b) Each statistics special awards winner receives no recognition in their own specific science category. In this situation it might be argued that "being a statistics winner has a NEGATIVE relationship with winning in a specific science category".

The actual situation is somewhere between these two extremes and might even indicate NO relationship. Statisticians should be able to present actual science fair results in some form, graphically or numerically, that exhibits the extent of association.

In the 1997 International Science and Engineering Fair at Louisville, KY, the FIRST AWARD winner in Statistics, appearing for her fourth straight year, received many awards in non-statistics categories. The project used a well-designed two-factor ANOVA. Her project received the FIRST Grand Award and Best of Category in the Behavioral and Social Science category, a special award from the American Psychology Association, an INTEL Achievement award, the Weissman Institute award and several offers of college scholarships. The direct cash awards totaled over \$10,000. While it may not be possible to claim that the use of statistics "caused" the other recognitions, it seems reasonable to say that her experimental design, analysis and presentation of the statistical results "enhanced" the quality of her project.

The SECOND award winner in Statistics received FOURTH Grand Award in the Behavioral and Social Science category. The project used response surface models. The THIRD winner in statistics, which involved the use of multiple regression models, received Fourth Grand Award in the Mathematics category.

At the Alamo Regional Science Fair in San Antonio TWO of the top four statistics projects were among the SIX overall projects that qualified to attend the International Fair in Louisville.

It is important for ASA chapters to continue their efforts toward the Adopt-a-School program, QL activities, and AP-Statistics. However, those statisticians who serve as science fair mentors and judges should direct the highly-talented science-fair students toward statistical design and analysis procedures that can take their research projects to a higher level.

Joe Ward may be contacted at:

Health Careers High School
4646 Hamilton Wolfe
San Antonio, TX 78229
e-mail: joeward@tenet.edu