Abstract # 4402 Event # 160:

Scheduled for Saturday, June 23, 2012 10:15 AM-12:30 PM: Session 25 (3rd Floor All Space) Workshop


APPROACHES TO THE ANALYSIS OF COUNT AND CONTINUOUS DATA WITH POSSIBLY MANY ZEROS

S. Blozis
Department of Psychology, 134 Young Hall, Davis, CA 95616, USA
line
     Count data as well as continuous data can present analytic challenges because inference from standard statistical procedures, such as ANOVA and regression, is based on an assumption of normality. Thus, alternative methods are required. This talk reviews methods for the analysis of count data that may also include a large frequency of zeros, such as data that represent measures of a behavior that is not often observed. Also discussed are methods for continuous data that are also not normally distributed, including cases in which there may be many zeros, such as measures of time use in which a sizable portion of the sample do not engage in a particular behavior. Examples are provided.