Chronological Listing of Posted Works


Critical values for the Stouffer-max probability-pooler (3/99)

Binomial tables. Each table shows the probability p of s or more successes in k independent trials, where P is the probability of success on each trial. For k up to 1000, tables show all values of p between .1 and .0001. (3/99)
Table for P = .01..... Table for P = .05..... Table for P = .1..... Table for P = .5

Compact binomial tables. Let P denote the probability of success on each trial, s the number of successes, and p the probability of s or more successes in k independent trials. For s up to 1000 and for four values of P, these tables allow you to tell whether p falls below .05, .01, or .001. These tables are particularly convenient for "file-drawer" analyses, but are useful for other purposes as well. These are the same tables referenced in the meta-analysis section. (2/98)
Table for P = .01..... Table for P = .05..... Table for P = .1..... Table for P = .5


Critical values of R(k,s) for k up to 30 and s up to 20, for alpha values of .1, .05, .025, .01, .005, .0025, .001 (6/98)

The SWEEP algorithm for multiple regression (4/98)

ALLSETS: a simple algorithm for all-subsets regression (4/98)

Using ALLSETS to identify collinear sets (4/98)

Some crude tables for assessing power and establishing confidence bands for the Wilcoxon signed-ranks test (3/98)

Typos and other errors in Regression and Linear Models (3/98)

A new diagnostic measure di for regression (3/98)

The Electoral College: bulwark against fraud (3/98)

Tables for calculating "fail-safe N" (2/98)

New tables for the p-product meta-analytic method (2/98)


REPTOC--An adaptive alternative to the Wilcoxon signed-ranks test (1/97)

Transforming a variable to a normal distribution or other specified shape (1/97)

Factor analysis (1/97)

Correcting the standard errors of regression slopes for heteroscedasticity (1/97)

Combining human judgment and multiple regression (1/97)

Multivariate analysis (1/97)

Regression methods in statistical process control (1/97)

Are we measuring "life expectancy" the best way? (1/97)


The Wilcoxon signed-ranks test (9/96)

A normal-scores alternative to the Wilcoxon test (9/96)

A meta-analytic "p-pooler" with three advantages (9/96)

Measures of association in crosstab tables (9/96)

Some new 2 x 2 tests (9/96)

The sign test (with table) (9/96)

Computation-based estimates of true shrunken multiple correlations (9/96)

A regression approach to time-series analysis (9/96)

Were the conclusions of Finlay and Darlington (1995) distorted by their failure to control for taxonomic grouping, body size, and part-whole correlations? (9/96)

Integrating the models of Robinson and Dreher (1990) and Finlay and Darlington (1995) on the timing of mammalian neurological development (9/96)

Correction to Table 1 of the June 1995 Science article "Linked regularities..." by Barbara Finlay and Richard Darlington (9/96)

The use of infant looking times in statistical analysis (9/96)

"The Bell Curve"--solid center or abnormal deviate? (8/96)

A permutation test for testing the difference between averages of correlation types in a large correlation matrix (8/96)

Supplementing random assignment with statistical control (8/96)

How many covariates to use in randomized experiments? (8/96)