Open Access

Table 2

Comparison of item scores among artificial intelligence chatbot and web search (categorical variable).

Artificial Intelligence Chatbot Web Search


Normal level 6th Grade Level



ChatGPT-3.5 Copilot Gemini ChatGPT-3.5 Copilot Gemini Google







n % AR n % AR n % AR n % AR n % AR n % AR n % AR
PEMAT (> 70)
 Understandability 12 40.0 −0.9 17 53.1 0.6 17 58.6 1.2 19 65.5 1.9 22 71.0 2.7 26 92.9 4.9 58 32.8 −5.7
 Actionability 0 0.0 −0.3 0 0.0 −0.3 0 0.0 −0.3 0 0.0 −0.3 0 0.0 −0.3 0 0.0 −0.3 1 0.6 1.0
jReadability difficulty level 10 8.0 −7.0 32 23.9 −3.1 69 55.2 4.9 42 34.7 −0.3 74 55.6 5.2 73 56.2 5.2 38 21.5 −4.4
 Very readable 0 0.0 −0.4 0 0.0 −0.4 1 0.8 2.6 0 0.0 −0.4 0 0.0 −0.4 0 0.0 −0.4 0 0.0 −0.5
 Readable 0 0.0 −2.3 2 1.5 −1.4 9 7.2 2.3 0 0.0 −2.3 14 10.5 4.6 8 6.2 1.7 1 0.6 −2.4
 Neutral 0 0.0 −3.7 16 11.9 1.4 29 23.2 6.2 1 0.8 −3.3 13 9.8 0.5 20 15.4 2.9 3 1.7 −3.7
 Somewhat difficult 10 8.0 −4.4 14 10.4 −3.8 30 24.0 0.2 41 33.9 2.9 47 35.3 3.5 45 34.6 3.3 34 19.2 −1.5
 Difficult 66 52.8 1.3 73 54.5 1.8 48 38.4 −2.2 73 60.3 3.0 55 41.4 −1.5 54 41.5 −1.4 79 44.6 −0.8
 Very difficult 47 37.6 8.6 25 18.7 2.0 8 6.4 −2.4 6 5.0 −2.9 0 0.0 −4.9 3 2.3 −4.0 36 20.3 3.1
 Not measurable (too difficult) 2 1.6 −1.3 4 3.0 −0.4 0 0.0 −2.3 0 0.0 −2.3 4 3.0 −0.4 0 0.0 −2.4 24 13.6 7.9

∗Very readable to Somewhat difficult total is listed.

χ2 test was performed to calculate adjusted residuals (AR); AR = adjusted residuals; significant difference if adjusted residuals are greater than or less than 1.96 (p < 0.05).

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.