Loftus's (1993bLoftus's ( , 1995 recommendation that investigators routinely include plots of appropriate means along with 95%confidence intervals or some other indication of variability has considerable merit, yet we agree with Morrison and Weaver (1995) that such plots can supplement but not supplant the usual reporting of analysis of variance results. Providing them may be easier than Loftus and Masson (1994) indicated, especially when error bars are understood as supplemental descriptive devices. We suggest a general, unified approach that applies to the explication of both between-and within-subjects effects. Variability is estimated separately for each group of scores identified as different by analysis because this serves description better. Raw scores are used for between-subjects effects, scores adjusted for between-subjects variability for within-subjects effects. All computations and figures are easily effected using common spreadsheet programs.In a recent series ofarticles in this journal and elsewhere, Loftus (1991Loftus ( , 1993aLoftus ( , 1993b argued for replacing traditional analysis of variance (ANaYA) results with pictorial representations of those results based on error bars, which he referred to as a plot-plus-error-bar (PPE) approach. He argued that traditional hypothesis testing emphasizes a binary yes-it-is or no-it-isn 't decision (see also Cohen, 1990;Rosnow & Rosenthal, 1989) that obscures two critical aspects ofany study, the strength of the associations between variables (i.e., the magnitude of the effects) and the relations among parameters, whereas error bars render these important aspects visible.In response, Morrison and Weaver (1995) argued that presentation oferror bars may supplement, but cannot supplant, standard hypothesis-testing procedures. Although agreeing that error bars are often useful and desirable, Morrison and Weaver noted that when designs include repeated measures or within-subjects factors, the definition and computation of the appropriate standard error is unclear. Loftus (1995) agreed that whatever one does technically, insight into research questions remains paramount and, drawing on another recent paper (Loftus & Masson, 1994), proposed definitional and computational solutions to the technical issues Morrison and Weaver raised concerning within-subjects error bars.Error bars are of many kinds. Discussing Loftus's PPE approach in general, Morrison and Weaver (1995) asked, why the standard error? If the intent is to provide a descriptive measure of variability, why not simply use the standard deviation or some other descriptive measure?

On the basis of recent work by W. Gardner (1995), implications of fallible observers for observational research are discussed. Analysis shows that for identically fallible observers, values for kappa are lower when codes are few and their simple probabilities variable than when codes are many and roughly equiprobable; thus no

No abstract

When determining whether a particular transition is more characteristic of one group than of another, two things are required: an index associated with the transition of interest and a statistical test that can determine whether group membership systematically affects values for that index. Here the familiar parametric t test is compared with a test based on sampled permutations. Indices considered are the odds and log odds ratio, Yule's Q, Wampold's (1989) transformed kappa, and phi. The odds and log odds ratio are monotonically increasing functions of Yule'sQ and so give similar results. Yule's Q and phi are essentially rank order invariant and usually give similar results. Transformed kappa, however, rank orders subjects somewhat differently than the others; moreover, it appears somewhat biased. With respect to the tests, when subjects are 20 or more it does not matter much whether sampled permutation or parametric t tests are used; both yield essentially the same result. However, when subjects are fewer than 20, or whenever there is any other reason to think that parametric assumptions may not be met, permutation tests are recommended. A computer program that effects such tests is described.

For preverbal children, episodes of joint attention are contexts for communication with responsive adults. This study describes the joint attention of 3- to 5-year-old children, 15 with autistic disorder (AD) and 15 with developmental language disorder (DLD), during play sessions with unfamiliar adults. Adults used fewer conventional than literal bids for joint attention with AD children and vice versa with DLD children. Children with AD were less likely to engage in joint attention than children with DLD. In the allocation of attention, AD children monitored the channel of communication with the adult 37% less often than DLD children. We discuss how perturbations in reciprocal interactions permeate the sharing situation and the implications of this problem for the mastery of cultural conventions.