Sample size in quantitative instrument-based studies published in Scopus up to 2022: An artificial intelligence aided systematic review

Abstract

Despite their popularity, quantitative instruments like Likert scales struggle with a practical issue for research projects - how many participants have to fill out the instrument? This study started with the data for 31,271 articles downloaded from Scopus and, after exclusions, reviewed the sample size used in 21,506 studies. Scimago highest quartile data was brought in for 1999-2021 and linked to the exported articles. Anthropic's Claude and Claude-Instant AI tools were used to analyze the journal article abstracts and extract the sample size information. Frequency distribution of sample size used are presented. Descriptive statistics such as maximum, average and trimmed averaged sample size values are presented by quartile of the journal where the article was published, range of years when the article was published, 3 common analyses which tend to increase sample size, and two population groups which have been shown to impact sample size. The study concludes with ranges for commonly used sample sizes based on a number of criteria. This is one of the first studies to use AI tools to assist in the analysis for a systematic review study.

Ähnliche Arbeiten