Location: Commodity Utilization Research
Title: A non-iterative approximation for critical values for the Dunnett’s testAuthor
|
Submitted to: USDA-ARS Research Notes
Publication Type: Research Notes Publication Acceptance Date: 2/2/2023 Publication Date: 2/24/2023 Citation: Klasson, K.T. 2023. A non-iterative approximation for critical values for the Dunnett’s test [Technical Report]. New Orleans, LA: USDA, Agricultural Research Service, Southern Regional Research Center. 9 p. https://www.ars.usda.gov/ARSUserFiles/60540520/CriticalValuesForDunnett.pdf Interpretive Summary: The Dunnett’s statistical test is sometimes used to find out if means of treatment groups are different than a control group. The method relies on the availability of statistical tables or sophisticates and expensive computer programs, but is not available on calculators or in spreadsheet programs. Thus, method was developed that can be implemented in spreadsheets, which are often used for data storage and evaluation. The method was implemented as a Microsoft Excel. The method provides values, similar to those in tables, with low error rates compared with exact methods. Technical Abstract: The Dunnett’s test is sometimes used following an ANOVA to discern if means of treatment groups are different than a control group. The method relies on the availability of quantiles from a distribution that is not available on calculators or in spreadsheet programs. Thus, a non-iterative method was developed that can be implemented in spreadsheets, which are often used for data storage and evaluation. The method was implemented as a Microsoft Excel user-defined function which requires Microsoft Visual Basic for Applications (VBA). In addition, the method was implemented as simple spreadsheet formulas that does not require VBA. The method provides values of the quantiles with low error rates compared with exact iterative methods. |
