Suppose you want to estimate the proportion of polydactyl cats (cats with extra toes). You sample 419 randomly selected cats, and find 56 polydactyls. You have no prior assumption to test - you only want to estimate the proportion. Should you carry out a hypothesis test using the sample proportion 56/419 as your null hypothesis? Select the best answer. Yes. Your null should be: The proportion of cats who are polydactyl is 56419, or 13.4%. 0 A. No. You should have a null hypothesis that the proportion of cats who are polydactyl is 0, since it's a genetic defect. O B. ° C. No. 419 isn't 10 times larger than 56. O D. No. There's no way to get a truly random sample. No. When all you wantto do isestimate a population parameter, you should construct a confidence interval. E. O F. I'm not sure.

Question

Mathematics

Dahlia Hudson

11 April, 19:50

Suppose you want to estimate the proportion of polydactyl cats (cats with extra toes). You sample 419 randomly selected cats, and find 56 polydactyls. You have no prior assumption to test - you only want to estimate the proportion. Should you carry out a hypothesis test using the sample proportion 56/419 as your null hypothesis? Select the best answer. Yes. Your null should be: The proportion of cats who are polydactyl is 56419, or 13.4%. 0 A. No. You should have a null hypothesis that the proportion of cats who are polydactyl is 0, since it's a genetic defect. O B. ° C. No. 419 isn't 10 times larger than 56. O D. No. There's no way to get a truly random sample. No. When all you wantto do isestimate a population parameter, you should construct a confidence interval. E. O F. I'm not sure.

+3

Answers (1)

Know the Answer?

Alexis Krause · Answer 1

No. When all you want to do is estimate a population parameter, you should construct a confidence interval.

Step-by-step explanation:

In this case, there is no other prior estimation about the population to test (a hypothesis to nullify). The only thing you can do is construct a confidence interval of the proportion, where the standard deviation can be calculated in function of the proportion and the sample size.

The right answer is E: "No. When all you want to do is estimate a population parameter, you should construct a confidence interval."