What is the difference between "Two-sample test (normal approximation)" and "One-way test (chi-square approximation)" displayed in the Wilcoxon test report?
The "Two-sample test (normal approximation)" gives results with continuity correction, whereas the "One-way test (chi-square approximation)" gives results without continuity correction.
The "(mean - mean 0) / standard deviation 0" displayed in the report is calculated using the continuous correction value as follows:
| (Rank Sum - Expected Rank Sum + Successive Correction) / Standard Deviation of Rank Sum |
Here, if "Rank Sum - Expected Value of Rank Sum > 0", the consecutive adjustment value is -0.5, and if "Rank Sum - Expected Value of Rank Sum < 0", the consecutive adjustment value is 0.5.
This post originally written in Japanese and has been translated for your convenience. When you reply, it will also be translated back to Japanese.