The Student Room Group

S2 critical points

When finding the critical values in hypothesis testing, do you ALWAYS half the significance value?

Posted from TSR Mobile
Original post by cooldudeman
When finding the critical values in hypothesis testing, do you ALWAYS half the significance value?

Posted from TSR Mobile


If I understand you correctly:

Only if it's a two-tailed test, when each of the tails with have half of the area.
Reply 2
Original post by ghostwalker
If I understand you correctly:

Only if it's a two-tailed test, when each of the tails with have half of the area.


Huh area???
Original post by cooldudeman
Huh area???


The significant level refers to the area in the tails of the pdf for a continuous random variable.

If it's discrete then it's the sum of the probabilities of the values that the rv takes in the critical regions.
Reply 4
Original post by ghostwalker
The significant level refers to the area in the tails of the pdf for a continuous random variable.

If it's discrete then it's the sum of the probabilities of the values that the rv takes in the critical regions.


In part c I assumed it to be a 2 tail test so I worked it out like this. O halved the significance level. On the MS it stated the alternative hypothesis as p <0.45 (p not equal to 0.45). But they did the critical region version and never halved the significance level to get the lower bound. Why would the write the two tail thing in brackets and assume its one tail...
Original post by cooldudeman
In part c I assumed it to be a 2 tail test so I worked it out like this. O halved the significance level. On the MS it stated the alternative hypothesis as p <0.45 (p not equal to 0.45). But they did the critical region version and never halved the significance level to get the lower bound. Why would the write the two tail thing in brackets and assume its one tail...


Markscheme?


Whether they use a one or two tailed test, the critical region at the lower end is the same.

Their alternative hypothesis allows a two-tailed test.

There is nothing in the markscheme to contradict a two-tailed test.
Reply 8
Original post by ghostwalker
Whether they use a one or two tailed test, the critical region at the lower end is the same.

Their alternative hypothesis allows a two-tailed test.

There is nothing in the markscheme to contradict a two-tailed test.


So im guessing my working out would be correct?

Posted from TSR Mobile
Original post by cooldudeman
So im guessing my working out would be correct?

Posted from TSR Mobile


Yes, but you really need to improve your layout.

At one point you have P(X<=5)<=0.025 which is wrong as it stands, but I know what you mean.

Get into the habit of writing what you mean, and not just abbreviating it - it would also help your understanding.
Reply 10
Original post by ghostwalker
Yes, but you really need to improve your layout.

At one point you have P(X<=5)<=0.025 which is wrong as it stands, but I know what you mean.

Get into the habit of writing what you mean, and not just abbreviating it - it would also help your understanding.


Would i have to write, reject null hypothesis if (X<=5)<=0.025
Is that better?
Posted from TSR Mobile
Original post by cooldudeman
Would i have to write, reject null hypothesis if (X<=5)<=0.025
Is that better?
Posted from TSR Mobile


That certainly makes more sense. I'm sure there must be examples in your textbook, with layout of working. I don't want to get dogmatic about what to write, but just make sure it's meaningful. If you had to read it again a couple of days later, would its meaning be obvious, or would you have to think about it?
Reply 12
Original post by ghostwalker
That certainly makes more sense. I'm sure there must be examples in your textbook, with layout of working. I don't want to get dogmatic about what to write, but just make sure it's meaningful. If you had to read it again a couple of days later, would its meaning be obvious, or would you have to think about it?


Ok thanks I get what you're saying

Posted from TSR Mobile

Quick Reply

Latest