When two sample means are significantly different

1. The difference is of practical importance
2. The pooled variance estimate should not be used
3. The difference is too great to attribute to chance
4. The corresponding population means are significantly different