In which of the following cases might a small effect still be important?
1) when external validity is high
2) when it is aggregated over many situations
3) when the sample is very large