Time Sampling: A Methodological Critique

Abstract

Numerous developmental studies of social behavior employ an observational technique called time sampling. This technique is used to estimate the frequencies and proportions of time that actual behaviors occur and to measure group or individual differences. By contrasting time sampling to continuous sampling, we demonstrate that (a) time sampling is inaccurate for estimating actual durations or frequencies of behavior, and (b) resulting individual and/or group differences can change depending on whether time sampling or continuous sampling is used. Error rates are particularly high when the total bout lengths of the actual behaviors are short (as in mother-infant interactions) and/or when interval length is long. Because the majority of mother and infant behaviors are short in bout length (less than 60 sec), when percent error is the greatest, we suggest that alternative sampling methods that have proven validity be used in future studies (e.g., scan, point, or continuous sampling).

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 — 78 citations exist upstream.

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