Are all data created equal?--Exploring some boundary conditions for a lazy intuitive statistician

Area of Science:

• Cognitive Psychology
• Decision Making
• Memory Research

Background:

• The naïve sampling model posits independence between data's temporal encoding and memory retrieval.
• This model implies random sampling from long-term memory for judgments.
• Investigating presentation order effects on statistical judgments is crucial for understanding memory retrieval.

Purpose of the Study:

• To examine how the order of numeric information presentation influences retrospective judgments of descriptive statistics.
• To test the assumptions of the naïve sampling model regarding memory retrieval and order independence.
• To assess the impact of repeated judgments on accuracy and potential anchoring effects.

Main Methods:

• Experiment 1: Participants judged descriptive statistics (mean, variability, shape) of number sequences with varying orders.
• Experiment 2: Employed a repeated judgment procedure with a mid-sequence change in data distribution.
• Utilized Arabic numerals and focused on retrospective subjective judgments.

Main Results:

• Neither experiment revealed significant effects of presentation order on judgments of mean, variability, or distribution shape.
• Experiment 2 confirmed that explicit judgments did not impair subsequent accuracy, refuting anchoring and insufficient adjustment.
• Participants demonstrated high responsiveness to data properties, yet were largely unaffected by presentation order.

Conclusions:

• Results align with naïve sampling models, suggesting data are stored as exemplars and randomly sampled from memory.
• Order effects in retrospective statistical judgments appear minimal, even with explicit judgment tasks.
• Cognitive processes involved in statistical judgment are robust to presentation sequence variations.