Author
 |
PAYTON, MARK - OKLAHOMA STATE UNIVERSITY |
|
Greenstone, Matthew |
|
Submitted to: Journal of Insect Science
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 9/3/2003 Publication Date: N/A Citation: N/A Interpretive Summary: Entomologists use statistical methods in order to determine whether the differences between experimental treatments, for example the effectiveness of different insecticides, are real or simply the effect of random variation. Sometimes an investigator will use a shorthand approach, asking whether measures of variation in the data, referred to as standard error intervals or confidence intervals, overlap. The problem with this approach is that there is no accepted procedure for determining how large these intervals should be so that whether or not they overlap will indicate that differences are significant according to the accepted statistical standard "differs from chance with a probability of 95%." To solve this problem, we used a computer to simulate experiments in which these intervals were of different sizes. We found that when 84% intervals are used, they give the desired 95% outcome. These findings will be useful to scientists in interpreting experimental results. One application of particular importance to entomologists is the determination of effective dosages for insecticides or effective concentrations of microbes applied for biological control of insect pests. Technical Abstract: We investigate the use of confidence intervals and standard error intervals to draw conclusions regarding tests of hypotheses about normal population means. Mathematical expressions and algebraic manipulations are given, and computer simulations are performed to assess the usefulness of confidence and standard error intervals in this manner. We make recommendations for their use in situations in which standard tests of hypotheses do not exist. An example is given that tests this methodology for comparing effective dose levels in independent probit regressions. |
