Two-sample t uses the conservative df vs Welch's df.
Compared to Welch's df, the conservative df produces:
- A
a narrower CI
- B
the same CI exactly
- C
a smaller p-value
- Dcheck_circle
a wider CI / larger p-value (more conservative)
Explanation
Smaller df gives a larger t* and a larger upper-tail probability, hence a wider CI/larger p.