Academic journal article Canadian Journal of Experimental Psychology

Using the Analysis of Covariance to Increase the Power of Priming Experiments

Academic journal article Canadian Journal of Experimental Psychology

Using the Analysis of Covariance to Increase the Power of Priming Experiments

Article excerpt

Abstract Although priming paradigms are widely used in cognitive psychology, the statistical analyses typically applied to priming data may not be optimal. Conceiving of priming paradigms as change-from-baseline designs suggests that the analysis of covariance (ANCOVA), using baseline performance as the covariate, is a more efficient (i.e., powerful) analysis. Specifically, ANCOVA provides more powerful tests of 1) the presence of priming and 2) between-group differences in priming. In addition, for within-subject designs with multiple baseline conditions, ANCOVA may increase the power of within-subjects effects. Efficiency gains are demonstrated with a re-analysis of priming datasets from implicit memory research. It is suggested that similar gains may be realized in other areas of priming research. Important assumptions of this procedure, which must be evaluated for the appropriate application of ANCOVA, are discussed.

Priming paradigms are heavily used in many areas of cognitive psychology. Examples include repetition priming paradigms in the study of implicit memory, semantic priming paradigms in the study of semantic memory and attention, negative priming paradigms in the study of attention and inhibition, and phonological and morphological priming paradigms in the study of language processes. Despite a heavy reliance on data from these paradigms, priming data may not be analyzed in an optimal fashion. In the present paper, we describe the conditions under which the typical analyses are less than optimal and propose the use of the analysis of covariance (ANCOVA) to render a more efficient (i.e., more powerful) analysis. Specifically, ANCOVA renders more powerful tests of 1) differences in the amount of priming across groups, and 2) the presence of nonzero priming within groups. In addition, for certain within-subject designs, ANCOVA may yield more powerful tests of within-subjects effects. In the final section of the paper, we consider some important assumptions for the proper application of ANCOVA. We preview one of these assumptions at the outset because of its critical nature: In the case of between groups analyses, it is assumed that the expected value of baseline does not vary across populations. If this assumption is violated, the ANCOVA approach is not recommended; if the assumption is valid, the ANCOVA approach produces an appropriate and more powerful analysis (see final section for details).

In a typical priming experiment, a researcher assesses the amount of facilitation (or inhibition) in the processing of one or more classes of stimuli relative to a control or baseline class. Additionally, it is often of interest to determine if the level of facilitation (or inhibition) differs across groups of subjects, where the groups may be experimentally created or pre-existing. In sum, a priming experiment consists of a within-subject manipulation of stimulus class (a baseline condition and one or more experimental conditions) and may also include one or more between-subjects variables.

To illustrate, consider a repetition priming experiment, a design frequently used to investigate implicit memory. In such experiments, subjects are presented with a stimulus set (often a list of individual words) during the study portion of the experiment. Some time later, subjects are presented with a disguised memory test, designed to measure implicit retention of the study stimuli. An example of such a memory test is the word-fragment completion task, in which subjects are asked to complete word fragments (e.g., _ 1 _ p _ a _ t) with appropriate English words (e.g., elephant). Unbeknownst to the subject, some of the fragments correspond to studied (or old) items and some correspond to nonstudied (or new) items. The measures of interest are the completion rates of the fragments. Retention of the studied information is inferred by enhanced completion of fragments corresponding to old relative to new items, with the completion rate for the latter items serving as the baseline completion rate. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.