Question

1. Provide a short example where you might prefer to calculate a confidence interval instead of conducting a hypothesis test. Provide a short example where you might prefer to conduct a hypothesis test instead of calculating a confidence interval. In both examples, provide a brief justification for why you prefer one method over the other (for example, what was the advantage of one method over the over).

Answer 1

Example where we might prefer to calculate a confidence interval instead of conducting a hypothesis test : Let us assume that a survey on cars is carried out and it's corresponding results are obtained. However, there is also a scope of error the way in which the survey is conducted and thus, statisticians or reseachers would never quote there results with a 100% confidence level. They would provide a result with (say) 95% or 99% confidence level i.e. a margin of error is always included in the results obtained. Therefore in such case of (say) mean number of people with a car, it would be better to make use of confidence intervals as compared to the hypothesis test as there may be the cases where people possesing a car, might lie in the survey about not having a one or a vic-a-versa case.

Example where we might prefer to conduct a hypothesis test instead of calculating a confidence interval : Let us assume a case of determining whether the melting point of a metal is same as that of the pre-determined value in the earlier studies. In this case, an exact value in degrees is required to be tested and thus, hypothesis test is the most appropriate test, since it is related to a test where only a single conclusion of either accepting or rejecting the pre-determined value.