Effects May Both Increase or Decrease in the Long Term: A Statistical Illustration

Area of Science:

• Psychology
• Statistics
• Research Methodology

Background:

• Small effect sizes are sometimes dismissed, with the argument that they can accumulate over time (Abelson, 1985).
• This accumulation of small effects into significant long-term impacts is a debated concept in psychological research.

Purpose of the Study:

• To statistically evaluate the assumptions underlying the accumulation of small effect sizes.
• To determine the likelihood of Abelson's argument holding true in typical psychological research contexts.
• To explore the potential for long-term effect sizes to be smaller than initially observed.

Main Methods:

• Utilized a simple statistical illustration to demonstrate the conditions for effect size accumulation.
• Analyzed the restrictiveness of Abelson's assumptions regarding effect size aggregation.
• Considered scenarios without long-term counteracting mechanisms like habituation.

Main Results:

• Abelson's assumptions for effect size accumulation are shown to be highly restrictive and unlikely to be met in psychological research.
• Even minor deviations from assumptions can significantly alter long-term effect size predictions.
• Long-term effect sizes may be smaller than those observed at a single time point, even without habituation.

Conclusions:

• The argument for the automatic accumulation of small effect sizes is based on unrealistic assumptions.
• Psychological research should empirically examine long-term effects rather than relying on speculative accumulation.
• Short-term observations may not reliably predict long-term psychological impact.