A purchaser of electronic components wants to test the hypothesis that they last less than 100 h. To do this she takes a random sample of 16 components and finds that, on average they last for 96 h, with a standard deviation of 8 h. If the purchaser knows that they lifetime of the components is normally distributed, should she accept the hypothesis that they last less than 100 h at the 95% confidence level?
I want to do a two tailed test
H_0: mu_1=mu_1 H_1: not H_o
root 100=10
t= (96-100)/(8/10)
= -4/0.80
= -0.5
t 0.025,99=1.9843
therefore, we can not reject the null
I am worried about the critical t value, when I calculate it for a two tailed test I have to divide the alpha value by 2 right? So we had alpha=0.05 I would divide it by 2 to get 0.025 and use that alpha value to look up the t critical right?
Thanks!
I want to do a two tailed test
H_0: mu_1=mu_1 H_1: not H_o
root 100=10
t= (96-100)/(8/10)
= -4/0.80
= -0.5
t 0.025,99=1.9843
therefore, we can not reject the null
I am worried about the critical t value, when I calculate it for a two tailed test I have to divide the alpha value by 2 right? So we had alpha=0.05 I would divide it by 2 to get 0.025 and use that alpha value to look up the t critical right?
Thanks!