Updated on Aug 25, 1999
You may be able to save these programs as text from your Browser, or Copy and Paste them into a text processor. They should work with most BASIC compilers, such as GWBASIC for PC and MS QuickBasic for Mac.
Except as noted below, all are copyright Michael H. Birnbaum, all rights reserved. They may be used for educational and scholarly purposes without charge. No warranty is expressed or implied. No guarantees can be made for the software.
ADDFIT is a BASIC program to fit an additive model to a two-factor experiment. Seminar Students: You can use this program to check your work on Homework No. 5A (Mystery Table) and 5B (Additive Resistance). You can also use this program to fit a subtractive model, but you will have to reflect (i.e. reverse the signs) of the subtrahend (difference = minuend - subtrahend).
MULTFIT is a BASIC program to fit a multiplicative model to a two-factor experiment. Seminar Students: You can use this program to check your work on Homework No. 5A (Mystery Table 1, Birnbaum, 1973), Pizza Problem, and Salary problem (Birnbaum, 1983). You can also use this program to fit a ratio model, but you will have to invert (i.e., 1/s) scales for the divisor (ratio = dividend/divisor). This program uses an iterative method to get a least-squares solution.
RDFIT.FOR is a FORTRAN program to fit two arrays of data simultaneously. The data are judgments of "ratios" and "differences" of magnitude in the same factorial designs. The program can fit one operation theory and several versions of two operation theory. This program is described in Birnbaum (1980).
RDFIT.DAT is a file of data showing setup for RDFIT. Included are mean judgments from Birnbaum's (1978) Cognitive Theory chapter and mean judgments from Birnbaum & Sotoodeh (1991).
RECIPE.FOR is a FORTRAN program that gives least-squares estimates of weights and scale values in a relative, constant weight averaging model to judgments collected in the recipe design. In recipe, a factorial design of three informational factors, ABC, is intermixed with two-way designs of AB, AC, and BC, and single factor designs of A, B, and C, each created by leaving out one or two pieces of information.
STEPIT.FOR is a fortran subroutine that can be used to minimize or maximize a function of several parameters. It was written by Chandler (references included in the file). This subroutine must be appended to recipe to complete the program.
RECIPE.DAT is a sample data setup showing how to arrange data for recipe. The data are hypothetical.
RTFIT.FOR is a FORTRAN program to fit
three arrays of data simultaneously. The data are "LESS" choice
Response Times, "MORE choice response times, and judgments of
"differences" of magnitude in the same (subset of) a factorial.
The program fits the theory of Birnbaum & Jou (1990), which
accounts for the distance effect, serial position effect, and
semantic congruity effect as well as "difference" judgments using
the same scale of subjective magnitude.
RTFIT.DAT is a file of data showing setup for RTFIT.
Included are means from Experiment 1 of Birnbaum & Jou (1990), COGNITIVE PSYCHOLOGY,
MONANOVA.FOR is a FORTRAN program written by Kruskal and Carmone
to transform data to additivity. References are included in the file. This version
has a small modification to allow the program to create a file of the transformed data,
which is useful for further analysis of the transformed scores.
MONANOVA.RTF PC VERSION Save as TEXT
MONANOVA.DAT is a file of data showing the setup
for MONANOVA. Included are means from Birnbaum(1974; 1982) and from Birnbaum (1978).
TAXFIT and CPTFIT
FORTRAN programs that fit choices between gambles, according
to Birnbaum and Chavez (1997) configural weight TAX model and Tversky and
Kahneman's model of cumulative prospect theory (CPT).
Data may be forced choice, or they may also include
judgments of strengths of preference between gambles.
Programs allow compromise between log likelihood criterion
that only depends on choice and sum of squared deviations
that depends on both choice and judgment of strength of
preference. Variants of these programs were used by
Birnbaum and Chavez (1997). Sample data from Birnbaum,
Patton, & Lott (1998) are included here,
TAXFIT.DAT and here:
to illustrate how to set up the data. You will also want to download
the following RTF file, which can be read in most
word processing programs.
Handbook for TAXFIT and CPTFIT
surveyWiz and factorWiz
Web pages that make Web pages that run simple surveys and experiments. SurveyWiz allows the
user to enter up to 90 personality or agree-disagree survey items or items requiring either
numeric or text response from the subject. FactorWiz creates the HTML Web page for a factorial
design. You specify the number of rows and columns, and the values of the factors, and
this Web page does the rest. Instructions for use are contained in the Web pages themselves.
taxcalculator and cwtcalculator
work only with Netscape, and they calculate predictions for the TAX model, the RAM model,
and the CPT model with parameters specified by the user. Taxcalculator calculates under
the TAX model of Birnbaum and Chavez (1997), and cwtcalculator calculates under the
RAM model of Birnbaum and McIntosh (1996). Taxcalculator uses the 2-parameter weighting
function of Lattimore, et al and Tversky and Fox. cwtcalculator uses the 1-parameter
weighting function for CPT given in Tversky & Kahneman (1992).
Return to Birnbaum's Home Page
This material is based upon work supported by the National Science Foundation under Grant No. SBR-9410572. Any opinions,
findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily
reflect the views of the National Science Foundation.
Jan 14, 1999