Hey I'm doing some homework and I got stuck.
The question I'm facing is: A poll was conducted to determine whether or not readers of a newspaper trusted sources used in articles. 552 responded yes and 531 responded no. You believe that the number of people who trust the information is greater than 50%. Test your belief.
I'm not exactly sure what the null(Ho) and alternative(Ha) Hypothesis is...
is it Ho: <= 0.5 and Ha > 0.5?
Then I'd guess you'd want to reject the hypothesis that less than 50% trust the info.
It's kind of an aside, but you ought to look into Symbolic Logic. It's used to determine if a conclusion follows logically from a premise, independent of the truth value of the premise. http://books.google.com/books?id=tq2...page&q&f=false
It's great for wading through multiple negations.
Your research question ("belief") is what you're testing for and that is always going to be your alternative hypothesis. Therefore, your setting up this test is correct. One way to look at why we do this comes down to the possible outcomes. Either we reject the null or fail to reject ('accept') it, depending on what the evidence of the test reveals. Our aim is to find sufficient evidence to accept our belief. However, we never truly find evidence for the null hypothesis. We're testing for a reason to reject it and accept the alternative. So if our belief were the null hypothesis, we'd be trying to find evidence against it, which tacitly says you're accepting it by default. If you understand that, you should be able to understand that whatever it is you're testing for should be the alternative. Sometimes we also think of the null hypothesis as the "status quo" because we take it to be a default position we're testing our belief against. More likely, we're just making an assumption about the data and testing our alternative against it (e.g., that the data is normal and centered on such and such a mean to test if there is actually a statistically significant departure). Make sense?
Posting Permissions
Reply With Quote