Table 1.

Comparative analysis between GPT-3.5 and GPT-4 for certain domains in biostatistics

Domain	Version of ChatGPT	Correct in the first attempt	Correct in the second attempt	Correct in the third attempt	Incorrect after more than 3 attempts
95% Confidence interval	V 3.5			Y
	V 4	Y
Binomial distribution	V 3.5			Y
	V 4		Y
Categorical data	V 3.5	Y
	V 4	Y
Cross-sectional study	V 3.5	Y
	V 4	Y
The chi-square test	V 3.5				Y
	V 4		Y
Measuring reliability	V 3.5	Y
	V 4	Y
One-way analysis of variance	V 3.5				Y
	V 4			Y
Probability: properties	V 3.5	Y
	V 4	Y
Sample size calculation	V 3.5				Y
	V 4			Y
T-test for paired data	V 3.5	Y
	V 4	Y

Y, yes.