Premium
Serially correlated errors in some single‐subject designs
Author(s) -
Phillips J. P. N.
Publication year - 1983
Publication title -
british journal of mathematical and statistical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.157
H-Index - 51
eISSN - 2044-8317
pISSN - 0007-1102
DOI - 10.1111/j.2044-8317.1983.tb01132.x
Subject(s) - autocorrelation , spurious relationship , statistics , statistic , mathematics , monte carlo method , test statistic , statistical hypothesis testing , econometrics , computer science
This note considers the problem of autocorrelated errors in systematic and magic figure designs. Their effects on the distribution of F ratios were investigated by Monte Carlo methods for a number of such designs and found to be severe, but specific to the particular design and F ratio within it. To allow experimenters to test for the presence of autocorrelated errors, critical points of Durbin & Watson's statistic r are tabulated for each of the designs investigated. What to do if the test indicates the presence of autocorrelation such that a significant F ratio may be spurious is still an unsolved (and apparently difficult) problem, but this ought not to discourage the use of such designs, since in a large number of cases it will not occur, and when it does the experimenter can always choose to err on the side of caution.