Question

It is important to keep the probability of making Type I equal to α. With a t-test, how do we keep the probability of Type I error in check?

Answer 1

A Type I error is rejecting the null hypothesis when it is in fact true.
Generally, people decide beforehand doing a hypothesis test that what will be the maximum p-value for which they will reject the null hypothesis. This value is denoted & is called the significance level.
Follow the following points to keep a check and deciding what significance level to use:

• choosing should be done prior to analyzing the data.
• Choosing the significance level should depend only on the consequences of  Type I and Type II errors.
• If the outcome of a type I error is serious or expensive, then a small significance level is appropriate( less than 0.05).
• If the reactions of a Type I error are not very serious and importantly if a Type II error has serious outcome), then a larger significance level is right(0.05 or more).
• 0.05 is preferred in general practice