Statistical result and economically meaningful result

A result is more likely to be statistically significant if the sample size is large. The increase in sample size leads to a decrease in the standard error. A decrease in standard error will lead to a higher value of test statistic, and the null hypothesis is more likely to be rejected because of that. But the result may or may not be economically significant if the null hypothesis has been rejected due to small departure from the critical value. It can occur due to a larger value of sample size.

The presence of taxes, transaction costs, and risk can further make the statistically significant result as an economically insignificant result. The investment gains due to the strategy can easily be wiped out by taxes and transaction costs.

Previous LOS: Type I and Type II errors, power of a test, decision rule, relationship between confidence interval and hypothesis tests

Next LOS: p-value of a hypothesis test