The Effects of Sample Size on Measures of Subjective Correlation

Abstract

The ability to accurately infer relationships in our environment is critical to our survival.One factor influencing the detection and evaluation of population correlations is the size of the sample used in their assessment.There are three primary categories of tasks used to measure perceived correlation: rating, prediction, and frequency estimation.Several studies have investigated factors related to participants' subjective correlations, but few have included manipulations of sample size.The results of the present experiment demonstrated that the subjective correlations of participants who performed the frequency estimation or prediction task increased with information quantity.However, the opposite trend was found when participants judged moderately correlated populations in the rating task.In addition, simulations of artificial decision-makers demonstrated the possibility of extracting rules capable of accurately inferring population correlations which attenuates for the skewed nature of the sampling distribution of Pearson's r for small samples.

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 — 0 citations exist upstream.

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