The declining significance of p-values

[Warren Weckesser]

Probably of interest to users and developers of statsmodels:
http://www.scientificamerican.com/article/scientists-perturbed-by-loss-of-stat-tool-to-sift-research-fudge-from-fact/

Warren

[josef...@gmail.com]

Interesting, Thanks for the link.

following up on it
p-values is just the headline, they are banning all Null Hypothesis
Testing including, it sounds like, confidence intervals
http://www.tandfonline.com/doi/abs/10.1080/01973533.2015.1012991#.VTFU0o7F80M

my opinion
They are doomed, or turn the journal into a collection of Bayesian or
storytelling articles.

The editors also say that they want larger sample sizes, but without
an inferential procedure in mind, I don't see how you can say how
large is large, how many observations should we get.

Also, the Scientific American article, but also one sentence in the
editorial seems to mix up frequentist sampling distributions, with
statements that the p-value is calculated **under the assumption that
the Null holds**.

I don't worry that statsmodels will loose many users of null
hypothesis testing, but someone should include more of the effectsize
measures.

Josef


> Warren
>

[josef...@gmail.com]

It looks like there is an interesting sequence of articles on pacifier
use in that journal

"The main effects of sex and length of pacifier use were not
significant (ps > .18). There was, however, a significant interaction
between sex and length of pacifier use, F(1, 84) = 6.07, p = .02, η2 =
.07. Post hoc analyses revealed that there was an effect of length of
pacifier use for boys (β = −.39), F(1, 84) = 7.23, p = .009, η2 = .08,
but not for girls (β = .10), F(1,84) = 0.44, p = .51 (see Figure 1).
There was no three-way interaction between sex, length of pacifier
use, and type of expression, F(1, 84) = 0.00, p = .98. The effect of
mother's education was significant, F(1, 84) = 6.57, p = .01, η2 =
.07, as was the effect of thumb sucking, F(1, 84) = 4.29, p = .04, η2
= .05, such that higher levels of both were related to higher levels
of mimicry."
(2012)

It looks like extended pacifier use by boys during the day has a
negative effect.

In another study (2014), there is a significant difference in the
response of adults to facial expression of the child depending on
pacifier use.

Are you smiling? Or starting to cry?

The second study uses SPSS and R.

Josef
:)

>
> Josef
>
>
>> Warren
>>