BRM article: DIY bootstrapping

[Bruce Weaver]

The following article in the latest issue of Behavior Research Methods may be of interest to some of you.

Cheung, S.F., Pesigan, I.J.A. & Vong, W.N. DIY bootstrapping: Getting the nonparametric bootstrap confidence interval in SPSS for any statistics or function of statistics (when this bootstrapping is appropriate). Behav Res 55, 474–490 (2023). https://doi.org/10.3758/s13428-022-01808-5
https://link.springer.com/article/10.3758/s13428-022-01808-5

Here is the abstract.

Researchers can generate bootstrap confidence intervals for some statistics in SPSS using the BOOTSTRAP command. However, this command can only be applied to selected procedures, and only to selected statistics in these procedures. We developed an extension command and prepared some sample syntax files based on existing approaches from the Internet to illustrate how researchers can (a) generate a large number of nonparametric bootstrap samples, (b) do desired analysis on all these samples, and (c) form the bootstrap confidence intervals for selected statistics using the OMS commands. We developed these tools to help researchers apply nonparametric bootstrapping to any statistics for which this method is appropriate, including statistics derived from other statistics, such as standardized effect size measures computed from the t test results. We also discussed how researchers can extend the tools for other statistics and scenarios they encounter.

Supplementary materials are available here:  https://osf.io/2twf5/

Cheers,
Bruce

PS- Cross-posted to SPSSX-L.

[Art Kendall]

Thank you.

I'll try to take a close look at the article and extension.

In 1971, the SIFT (Statistics by Interaction with Files from a Terminal) package had a module called Mother which could generate correlation matrices that had a particular factor structure. IIRC it could also generate practice datasets for descriptive stats.

I have long advocated using SPSS to generate examples simulating  the confidence intervals from random
data with known parameters.

INPUT PROGRAM is useful for (a)simulating many things: cards, dice, etc.  (b)for developing better understanding of obtained probabilities, (c) for handling unusual distributions.

For example, wrap the syntax below in an outer loop look at the kind of IQ scores that could come from sets of 40 cases.
new file.
input program.
   loop id = 1 to 40.
      compute IQ = rnd(rv.normal(100,15)).
      end case.
   end loop.
   end file.
end input program.
execute.

* for all such simulations.
*paste your syntax here to save the results in a data set .
* then see where your obtained mean falls in the set of simulated results.
* how does that compare to the results from the usual procedures

I do not recall when it was introduced, but parallel analysis is and extremely useful tool in approximating a stopping rule in FACTOR.  With today's computers is trivial to generate a million sets of eigenvalues from randomized data with the same number of cases and variables.  One can then see how the obtained eigenvalues compare to the distribution of eigenvalues from the simulation.