ECMT Quiz 3 (1 Viewer)

[s.m.i.t.h]

ok, here's where i get confused.
am i doing the p-test right?
How the hell do you calculate the progaybility of a type 2 error?

why in part d) arent u using null hypothesis: >5.80 and alternative < 5.80??

 

[stazi]

i dunno, i kinda started getting confused after laine. i also started looking through text examples, tutor examples etc and it seems that
H0:u=5.80
H1 " >5.80

as hes testing whether its at least 5.80

well actually this confuses me further. i have no idea.

 

[absolution*]

H0 is the status quo and what you expect
H1 is the alternate hypothesis that you are trying to prove

therefore id assume

H0: mean (more than or equal to) 5.80
H1: mean < 5.80

 

[absolution*]

btw stas, you need to work on your definitions of why H0 is not rejected. you cant just say it is so.

Ie. "Based of the evidence provided by the sample, one is not able to reject the null hypothesis which states: etc etc etc"

thats what my tute said anyway.

Also, with type I and II errors, i dont think you have to calculate them. basically if there is a large difference in the variation between the first mean and the second mean than there is a small probability of type II error. If they are close together than there is a high probability of type II error.

 

[WaffenssLAH1985]

help desperate man

Does any one tell me how to calculate significance F, P-value & t-Stat of bo and the value of bo in question 2? plz!!!!!!!!!!!!!!!!!!!!!!!!

 

[absolution*]

will have question 2 answers tomorrow hopefully

 

[stazi]

btw stas, you need to work on your definitions of why H0 is not rejected. you cant just say it is so.

Ie. "Based of the evidence provided by the sample, one is not able to reject the null hypothesis which states: etc etc etc"

thats what my tute said anyway.

Also, with type I and II errors, i dont think you have to calculate them. basically if there is a large difference in the variation between the first mean and the second mean than there is a small probability of type II error. If they are close together than there is a high probability of type II error.

thanks man. yeah, i just couldn't be bothered writing that down for working out on bos . My tutor said that as well. Also kept stressing that we can't simply say "accept h0"
There is a formula for working out the type 2 error probability though. A bit too busy right now to look it up (doing marketing presentation stuff).

Also, can someone just double-check the p-value things.

Furthermore, i'll likely put up how to get the A-K thingos tonight/tomorrow.
I finally understand most of regression

 

[myg0t]

Absolution, is your signature photo taken from Redfern - perhaps near that one that was painted over today about the syringes and how everyone turns their back on it?

 

[ressul]

P-value for about a)

I just started the quiz. For the first question. If we get z=1.11(according to the answer provided), so what's the p-value for it? I calculated the area below Z=1.11, and the p-value I got was 0.8 sth(). Does it conflict with the answer we got? Because I think the 0.8 p-value means that there is a 80% probability for the mean to be below the average we predict, in other words, the H null should be reject.

Is my thinking right?

 

[sarevok]

I agree with the hypotheses Lainee posted, in which case the p-value would be the area above 1.11, rather than area below it.

 

[ressul]

Why?

Why is it above? Is the p-value always measured for the above area? What does that mean? I don;t understand it.

 

[sarevok]

Because, using Lainee's hypotheses, the rejection region is entirely in the upper tail; thus, the p-value is the area above the test statistic. Refer to p. 308 of textbook.

 

[ressul]

http://www.boredofstudies.org/community/attachment.php?attachmentid=9205

No, it says that the rejection region is the low tail.
http://www.boredofstudies.org/community/attachment.php?attachmentid=9205

 

[stazi]

i may have made a mistake.
if p-val > l.o.s. then we do not reject from memory.

 

[ressul]

What is LOS? Do you mean the significant level? So you agree with the p-val should be 0.8 for this question? Or.......

 

[stazi]

What is LOS? Do you mean the significant level? So you agree with the p-val should be 0.8 for this question? Or.......

level of signifacance.
I'm not sure about the pvalue stuff. i really don't know. oh and the question 2 sheet is nearly done. I'm having some probs though.

 

[stazi]

awesome. question 2 (the heart disease one) done.
theres a few im unsure off and theyre highlighted.
Help would be great!

 

[Lainee]

question 1b

For this question, I would have taken a completely different approach.

(Not a complete solution, but just highlighting some differences in approach in my solution and yours.)

Ho: u=90
H1: u<90

At 0.05 level of significance, reject Ho if t<-1.6604

Calculating test statistic using t-statistic:
t=-2

Since t<t(critical) reject Ho at 5% level of significance and conclude that average amount spent per household per week is significantly less than 90.

 

[Lainee]

ok, here's where i get confused.
am i doing the p-test right?
How the hell do you calculate the progaybility of a type 2 error?

Q1(d)(i)

Again, check your substitution of values into the t-statistic; n=100 not 99. I think uou would've gotten the test statistic as t=1 if this was done correctly.

I think your diagram is not completely accurate either.

(ii)

We had t=1, keeping in mind 0.05 level of significance (which is our a) and 99 degrees of freedom.

We will reject Ho if P(test-statistic)<P(critical). See Slide(37) of Week 9 lectures for a similar diagram of what we have here.

Looking at your t-table, t=1 is in between 0.25 and 0.10 (we cannot pinpoint the p-value without using Excel).

(p value is between 0.25 and 0.10) >= (a=0.05)
Therefore, do not reject Ho... etc.


...does that make any sense?

 

[ressul]

Why?

Can anyone tell me why we set hypothesis like Ho: u=90 H1: u<90 for b)?