because there is no evidence.

oh noes...

There's quite a sizable difference between emergency response times and the existence of a higher power.

The response time has to be something, and 3 minutes might be quite a reasonable estimate. There's also information about response time available from residents who have made emergency calls, and from those responders. They should be able to give a reasonable (but probably not accurate) estimate of the response time.

My cliff notes:

- RainmanTrail claims that a higher power is an accepted and obvious truth.
- Despite a request, RainmanTrail does not explain why this is.

This is incorrect - you may have done the mathematics behind statistics but this is a serious flaw in understanding. (I don't think expertise matters here as it is an elementary error, however in case it matters to you - I've also done the maths behind statistics and I have both worked as an econometrician and run several university courses in "statistics for dummies" and hypothesis testing in particular). The choice of null hypothesis has got nothing to do with "whatever the current acceptance is" - it's based on what claim you're seeking evidence for. This isn't a subtle technicality it's basic.

In your first example the null hypothesis is chosen to be 'response time less than or equal to three minutes' because as stated they are specifically trying to prove that the response time is greater than three minutes. (The police department could just as easily conduct a study with the null hypothesis of 'response time greater than three minutes' to counter the claim that they were failing in their goal).

In your second example, the police department are trying to establish a claim that aftertaking remedial action the response time is now less than three minutes - therefore that becomes the alternate hypothsesis (no matter what is 'currently accepted'). Whatever you are seeking evidence for is your alternate hypothesis - it is irrelevant how accepted the claim or its negation is. This is obvious since, if you were correct in your formulation, it would be impossible to test anything without some general pre-existing consensus as to what was true.

In summary - what is currently accepted is irrelevant when choosing a null and an alternate hypothesis. What matters is what you are trying to find evidence for (you may be trying to confirm something universally accepted, you may be trying to establish something radically new - it doesnt matter. What you want to find evidence FOR becomes your alternate hypothesis).

I agree with you that hypothesis testing is not really relevant to your discussion (though in my mind it could be used to argue for agnosticism), however it's got nothing to do with 'current acceptance'. And I can find no charitable way to interpret this earlier statement:
It seems you are tying a high alpha to an unlikely hypothesis - I can only presume you had something else in mind as any such linkage is a gross misunderstanding of significance levels (even for someone with technical training but no practical experience as I've assumed).

I'm a statistician. I disagree. Many of my friends & co-workers are also statisticians. They also disagree (at least the two I got a hold of).

I already stated that you can formulate your study whichever way you want to. I'm just stating that the generally accepted method is to begin with the accepted ideals as the null.

I don't really care if you agree with me or not. It won't change my life one bit. I'll still be a statistician tomorrow.

My "two pages" are up ladies and gentlemen, maybe I'll troll these threads another day, but I doubt it.

If I find time, I will make a new thread regarding my "Big Picture" and perhaps another regarding why I accept the Resurrection account in the Bible to be true.

Until then... take care everyone and thank you for keeping things civil (not always the case in religious discussions).

Rain

Ah, the internet equivalent of sticking one's fingers in his ears and pouting.

lmao later chump

Thanks for responding RainmanTrail. Oh.

Being exceptionally arrogant and never explaining your claims is not a good way of convincing anybody. It just makes you look silly, while everybody else continues to disbelieve.

Given your apparent confidence I was expecting quite the discussion.

A: "I've found a new drug, it might cure cancer!"
B: "Cool! Any evidence?"
A: "Unfortunately not. I can't perform any statistical test until there is a widely accepted view of it's efficacy."
B: "Bummer."

Do you work as a statistician or have you done a degree which includes statistics?

And what do you mean by this:

"Also, one can certainly "choose" to perform a scientific experiment using the unlikely hypothesis as the null, but that is generally accepted as incorrect. Have you ever seen an experiment with the (100-alpha) level of significance at something like 5%?"

It seems you are suggesting that alpha is related to how likely it is that the null hypothesis is true - which is clearly incorrect. I can't think of any other meaningful interpretation of the above.

I noticed this quote in your post Bunny and would like to comment on it.

"Also, one can certainly "choose" to perform a scientific experiment using the unlikely hypothesis as the null"

Just because God is commonly accepted as truth does not mean it's likely to be true.

Of course. That quote was me citing Rainman Trail - I was asking him what he meant by it as it seemed he was linking the significance level of a statistical test with whether the null is likely to be true. This is an elementary error and I doubt he meant it as I've understood it, but I couldnt see any other interpretation.