# P VALUES ARE THE DEATH OF ME

For the love of God, can someone please help me find a way to help me remember the rule about p values? I know it’s easy and I also know I’m an f’ing retard. But I can’t keep it straight. If P value is below signifigance, then you can reject? Or not reject? Oh crap help me.

P-value is the LOWEST significance level you can reject the null at…It is the POWER of the test. The lower the signficiance level you can reject the null at the more POWERFUL the test. So lower values are better. If the p-value is lower than the signficiance level you are testing than you can reject the null.

We always use three standard rejection levels 10%, 5% and 1%. If your p-value is .00001 you could reject the null at .001% level.

For some reason my brain rejects that. I read what you write, and I understand it. Then I close my eyes and try to repeat it, and my mind goes numb. Anyone have a trick I can remember?

I don’t know if this is going to help, but the way I remember it, is that P value is probablility that T-stat will be such that null hypothesis would be true. well, we don’t want that, we want low probability that null hypothesis is true. Sorry if this confuses you more.

If p is 0.04 (or 4%) you can reject Ho at any level of 0.04 or higher… So if your P is lower than your Sig.Level, you could reject… If your penis(=code for P value) is lower, reject it…significance

Think of p-value as a score on a horrible looking girl. YOU WANT TO REJECT THAT ONE. REJECT IT AND YOU’RE HAPPY. p value is the score your friends give the girl. You want them to give it a low score so you can reject her on that basis. Otherwise you cannot reject her and she’ll want happy times. The critical value (10%, 5%, 1%) is the ‘okay’ score. If she scores below it, you can reject her and look for a better one. If your friends say her score is above that value (say, they give her a p-value of 12%) you can NOT reject. If you cannot reject you’re screwed. You don’t want that. Low p-values is what you want to see.

I look at it in comparison to the T-stat to. I really really high T-stat will have a really really low P-value. A low P-value shows you which variables are stat. different from 0 at a level of significance. If we are comparing to a 1% significance and we have a P value of .0002 that would show that the .0002 is significant at that level.

Sorry no tricks. For me it helps if i think it through like this: If you can reject something at the 1% significance you can certainly reject it at the 5% level. So what you need is the lowest level of significance you can reject it at which is the p value. Given that the p value is the lowest level you can reject it at , the rest follows i.e reject if significance is higher than p-value.

That’s great mcpass!!! we need more analogies like this for other concepts - really DOES make it easier to understand!!!

mcpass Wrote: ------------------------------------------------------- > Think of p-value as a score on a horrible looking > girl. YOU WANT TO REJECT THAT ONE. REJECT IT AND > YOU’RE HAPPY. > > p value is the score your friends give the girl. > You want them to give it a low score so you can > reject her on that basis. Otherwise you cannot > reject her and she’ll want happy times. > > The critical value (10%, 5%, 1%) is the ‘okay’ > score. If she scores below it, you can reject her > and look for a better one. If your friends say her > score is above that value (say, they give her a > p-value of 12%) you can NOT reject. > > If you cannot reject you’re screwed. You don’t > want that. Low p-values is what you want to see. Mcpass, that did it. It’s locked in. Can’t thank you enough

I wish my mind worked as brilliantly as McPass’s

I can promise you… you don’t. Thing is my gf went away for three weeks with her parents. Been 5 days now and studying gets me horny as hell. I actually think the actual exam will be one of the most erotic events in my life now that I’m linking every single subject to these kinds of things. Luckily for me she’s coming back on June 7th… Poor thing she doesn’t even see it coming.

Get all heteroskatastic on her

Someone had a nice one the other day here in one of the threads… in fact I’m going from a holding effect right now to a flow effect when she gets back.

" Been 5 days now and studying gets me horny as hell." You don’t say? You won’t be studying for anything during our Chicago trip, right?

You suckers have me rolling on the floor over here. Maybe I am glad that I’m taking the boring approach to the cfa exam. hahahahah

CFAHouston Wrote: ------------------------------------------------------- > " Been 5 days now and studying gets me horny as > hell." > > You don’t say? You won’t be studying for anything > during our Chicago trip, right? Houston… you’d have a problem!

The way I think of it is (given the value): Just how likely is it to be significant enough to make the null wrong. If the number is low then happy days.

mcpass Wrote: ------------------------------------------------------- > Someone had a nice one the other day here in one > of the threads… in fact I’m going from a > holding effect right now to a flow effect when she > gets back. mcpass, I love it when people quote my own jokes back to me.