Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: ABX (Read 2030 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

ABX

I´ve seen that the ABX plugin still uses the unfortunate - because it is incorrect - statement "probability that you´re guessing" which leads to quite frequent misunderstandings as it seems to support a common fallacy about the p - value for this kind of test.

It would be better to exchange this statement by another one, expressing the true meaning of the p - value like "probability to reach your number (or even more) of correct answers by random guessing" .


Re: ABX

Reply #1
Are you seriously debating this?

Are you really that offended? Objective scientific fact is not meant to be sugarcoating things. If you are randomly pressing buttons as to get a expected proscribed response, yes, you are guessing.

Re: ABX

Reply #2
You mentioned "objective scientific fact" which was my concern.
The p-value (*) does _not_ represent the probability that the listener was randomly guessing, that is the _objective_ _scientific_ _fact_in this case.
Hence the statement is incorrect, misleading and represents a common fallacy, therefore it should be replaced by a correct one.

So, yes, if someone is randomly pressing the buttons then he is guessing, but you don´t know from the p-value; it´s as simple as that.

(*) the p-value is the result of a exact binomial test using the number of correct responses as test statistic. This way the observed data is analyzed under the assumption that the null hypothesis (p = 0.5) is true, it can´t tell you _if_ the null hypothesis is true.
Iow, p(observed data l H0) is not the same as p(H0 l observed data).


Re: ABX

Reply #3
We’ve been down this road already. The wording isn’t accurate.

It should read as the likelihood that you would do the same or better by flipping a coin.

But it doesn’t and the world will explode as a result. Hyperbolic enough?

Re: ABX

Reply #4
Not sure, what i should do with the "hyperbolic" remark.
Does it mean, spreading incorrect information is ok as long as the world does not explode?
I´d hope that is shouldn´t have that meaning; there exists already a lot of misunderstanding around statistics in general and particularly around this kind of tests (see the numerous common fallacies wrt conclusions from results).

If already all agree that it is incorrect what´s the problem to exchange the message by a correct one?
No special coding effort is required, just using the correct wording.


Re: ABX

Reply #6
Not sure, what i should do with the "hyperbolic" remark.

Easy, translate: bugger off.

Ah, thanks for the kind help.  :)

Basically Mudlord´s "sugarcoat comment" illustrates the impact of the incorrect statement and obviously the "we´ve been down this road already" event did not help while a corrected statement would help (at least for the future).

 

Re: ABX

Reply #7
Take a hint buddy. Get lost.

Re: ABX

Reply #8
I believe Peter acknowledged the need to change the wording. And try to remember jazzthieve doesn't speak for all of us.

 


Re: ABX

Reply #11
The component was updated with fixes and wording change.