Hypothesis testing: the null hypothesis?

Question

A CEO claims that at least 80% of its million customers are very satisfied. 100 customers are surveyed using simple random sampling. The result: 73 percent are very satisfied. What is the null hypothesis here?
Is it P >= 0.80? That's what it says here:
http://stattrek.com/Lesson5/Proportion.aspx?Tutorial=AP

However, I've got another question from a textbook:
A manager claims that over 60% of a mall's visitors shop at the store. Let p be the proportion.
Then the null hypothesis is that p=0.6.

So in the first case, the claim is the NULL hypothesis, while in the second the claim is the ALTERNATIVE hypothesis. Am I missing something here? Thanks

Fred · Accepted Answer

The null hypothesis is what we assume the value of the population parameter really is.  In your case the pop parameter is P so we assume it is 80%.  So  H0: P=80%.  If you want to determine if customer satisfaction is not 80% the alternative hypothesis is HA:  P (not=) 80%.  If you want to determine if customer satisfaction is lower than what is assumed we would test the alternative hypothesis HA:  P<80%.  Finally if we wanted to test if customer satisfaction is greater than what is assumed we would test HA:  P>80%.  

The null hypothesis is what we assume is true for the population parameter.  The alternative hypothesis is determined by the context of the problem.

Anonymous · Answer

The null selection is the two kinds are an identical So the mean could be at 5% point the mean is 5.2 the alternative hypothesis could be that the mean are distinctive so mean<5.2 or mean>5.2 could be selection. yet no you are able to think of , that the breeder isn't unterested via the undesirable effect so in the subject you soak up account in simple terms the 2nd hypothesis

Trending News

Hypothesis testing: the null hypothesis?

2 Answers