Beyond inference by eye: Statistical and graphing practices in JEAB, 1992-2017

Area of Science:

• Behavioral science
• Psychology
• Scientific publishing

Background:

• Debates on p-values and data analysis have led to new statistical inference guidelines by the American Psychological Association (APA).
• Changes in reporting results are evident in many scientific journals.
• The impact of these shifts on the Journal of the Experimental Analysis of Behavior (JEAB) remains unevaluated.

Purpose of the Study:

• To analyze statistical and graphing practices in JEAB empirical articles from 1992 to 2017.
• To determine if and how these practices changed over the evaluated period.
• To compare JEAB trends with those in mainstream psychology journals.

Main Methods:

• Content analysis of empirical articles published in JEAB between 1992 and 2017.
• Tracking the frequency of null hypothesis significance testing, confidence intervals, and error bars in figures.
• Assessing changes in graph features, specifically the proportion of single-subject data depictions.

Main Results:

• The likelihood of reporting null hypothesis significance tests, confidence intervals, and error bars in figures has increased over time in JEAB.
• Graph characteristics, such as the depiction of single-subject data, showed no systematic changes.
• JEAB's statistical and graphing trends largely paralleled those in other psychology journals.

Conclusions:

• JEAB's statistical and graphing practices have evolved, mirroring broader trends in scientific psychology.
• There is no direct evidence that American Psychological Association (APA) style changes impacted JEAB.
• Authors, reviewers, and editors must ensure continued evolution of statistical and graphing practices in JEAB while preserving its unique characteristics.