1.    Should you use significance tests of skew and kurtosis in large samples?

1. No, because they are likely to be non-significant even when skew and kurtosis are significantly different from normal.
2. No, because they are likely to be significant even when skew and kurtosis are not too different from normal.
3. Yes, because large samples produce more accurate results.
4. Yes, because large samples add power to the test.