These programs are used for statistical analysis of data from bioassay studies, such as determining the effects of insecticides on insect mortality. The programs are written in the Mathematica language. You must have the Mathematica software on your computer to run these programs. Note: to determine whether you have free access to Mathematica, visit the Mathematica web site (www.wolfram.com) and the web site will automatically notify you if you have a site license. If you do not have a site license or do not have the Mathematica software package on your computer, you will not be able to run the programs that I have written.

Note: The probit programs were updated in February 2006 to simplify comparisons of slopes and intercepts and calculating relative potency by including slopes, intercepts, and their standard errors and covariances in the output file for cutting and pasting into other programs. WordPerfect files describing the numerical methods used were converted to pdf files. Also, the programs were updated in October 2007 to address changes in Mathematica version 6.0. When using older versions of the programs in Mathematica 6.0, the graphs do not print correctly and several error messages are generated, but numerical output is correct. Two versions of the programs are now available - for Mathematica version 6.0 and higher, and for previous Mathematica versions.

1. Create a folder/directory for the programs.
2. Click here to download the .zip files that you want.
3. Extract the .zip file to the newly created directory, and then delete the .zip file (A Windows zip utility can be found at www.winzip.com).
4. The programs are ready to edit and then run in Mathematica.

The PROBIT program is used to analyze bioassay data when multiple observations over time were made on the same groups of organisms at one dose of a stimulus. If you made multiple observations over time at more than one dose in a single study, don't use this program to analyze those data. Instead, use the program by Preisler and Robertson (1989. Journal of Economic Entomology 82: 1534-1542). Our program gives you the option of using any of six possible transformations of the data (probit, logit, CLL [complementary log-log], log-probit, log-logit, log-CLL). All information for complete reporting of probit analyses is provided by the program, including: the slopes and intercepts, with their variances and covariance; the chi-square for goodness-of-fit of the regression line; and lethal time values, with confidence limits. The program PROBIT2 does the same thing, but will calculate all six possible transformations at one time.

Backtransformation programs

The backtransformation programs use output from the PROBIT program, or from any other probit analysis program. The BACKTRAN program can be used to transform probit-, logit-, or CLL-transformed data back to the original units (proportion organisms responding to the stimulus) to help assess goodness of fit. The program will also calculate residuals and standardized residuals of proportion organisms responding to the stimulus. The program outputs time or dose, the observed and predicted proportion organisms responding at each time or dose, and the residual and standardized residual corresponding to each time or dose. The program also outputs the observed and predicted probit-transformed (or logit- or CLL-transformed) data corresponding to each time or dose. These data can be used to plot observed vs. predicted proportion organisms responding to the stimulus, or the corresponding probits, to assess goodness-of-fit. These graphs are also automatically generated by the BACKTRAN program. BACKTRAN should be used only for data that are correlated. I.e., you looked at the same insects over several time periods. If your data consist of independent observations - i.e., a different batch of organisms for every observation time or for each dose - use BACKTRN2.

Accessory programs

The accessory programs use output from the PROBIT program, or from any other probit analysis program. The SLOPE program is used to calculate whether slopes and intercepts from two regression lines differ. The RELPOT program is used to calculate relative potency of two stimuli, including confidence limits on relative potency.

Complete instructions are included at the beginning of each program and can be viewed using an editor (such as Notepad) or word processor. To use the programs:

1) Edit the programs in an editor or word processor to include your data
2) Save the modified files in ASCII format (not in word processor format)
3) Run the programs in a Mathematica session. Output is written to a user-named file.

PDF files describing the numerical techniques used in the programs are also available for downloading.

Dr. James E. Throne
USDA, ARS, GMPRC
1515 College Avenue
Manhattan, KS 66502
(785) 776-2796
james.throne@ars.usda.gov