Table 4 Distribution (%) of evaluators’ scores on responses by extension agents (EAs) with and without extension materials and chatbot to 32 questions.

From: Evaluating responses by ChatGPT to farmers’ questions on irrigated lowland rice cultivation in Nigeria

 

Very good (Score 5)

Good (Score 4)

Acceptable (Score 3)

Poor (Score 2)

Very poor (Score 1)

EAs without extension materials

2 (6)b

8 (26)a

16 (51)a

5 (15)a

1 (2)a

EAs with extension materials

3 (8)b

11 (34)a

14 (46)a

3 (9)a

1 (3)a

ChatGPT

13 (40)a

8 (24)a

5 (18)b

5 (15)a

1 (3)a

Pearson chi-squared (p-value)

< 0.001***

0.314ns

< 0.001***

0.327ns

0.67ns

  1. Evaluators judged “the quality of information provided” with scores as very poor, poor, acceptable, good, or very good. Within a column, different letters indicate statistically significant differences (P ≤ 0.05).
  2. As the scores recorded vary across evaluators, we made an average for each score. The percent can be different depending on the average; for example, if the average number of responses with a score of 3 is 7.75 and 8.25, the percent is 24 and 26, respectively. *** statistical significance at and 0.1% (P < 0.001) level; ns, not significant. Values in the brackets denote the value in percentage.
Back to article page