Statistical Models for Behavioral Observations

Abstract

Statistical models for the Behavior Stream (renewal processes) are used to explore properties of behavioral observations. For continuous observation, statistical and psychometric properties are obtained for four types of behavioral measures: (i) empirical rates of behavior; (ii) empirical proportions, or relative frequencies, of a type of behavior; (iii) empirical prevalence (proportion of time the behavior occurs); and (iv) empirical event duration. Also, time-sampling alternatives to continuous observation are evaluated. Our formulation includes representations for three sources of unreliability: (a) finite observation time, (b) recorder errors, and (c) heterogeneity (instability) over occasions of observation. Traditional psychometric methods carried over from the analysis of responses to test items (including generalizability theory) are shown not to be applicable to behavioral observations. Our results provide a guide for design and a framework for statistical analysis in behavioral observation research.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 47 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.