Talk:Testing hypotheses suggested by the data

This article should make reference to the concept of (and cross-reference the article for) double dipping. — Preceding unsigned comment added by 129.67.14.209 (talk) 19:55, 29 January 2014 (UTC)

How does "Testing hypotheses suggested by the data" differ from data dredging? Should the articles be merged? Fedos (talk) 17:44, 9 May 2016 (UTC)

Sourcing

Is this entire article taken from the sole source listed? Angryapathy (talk) 19:31, 25 September 2009 (UTC)

I don't know, but the tone made me suspect it was directly taken from elsewhere because sections shouldn't be informally titled "How to do it wrongly" and "How to do it right", as per WP:NOT#HOWTO. The title of the article itself is an unlikely search term, so while the subject may be notable (though it could be a section of a statistical hypothesis article), it needs some restructuring or rewriting to be up to standard. The intro sentence gives insufficient context for example. Actually, after verification, statistical hypothesis testing also has tone and "jargon" issues.--74.56.234.186 (talk) 02:29, 28 September 2009 (UTC)

Article repeatedly implies we accept the alternative

This article several times talks about accepting the hypothesis but the hypothesis in question is the alternative. We don't do that; our decision is whether to reject the null. That is, with regard to hypothesis tests we either *reject the null* or we *fail to reject it*. Anything else is in the province of conclusions or choice of actions on the basis of that decision. Glenbarnett (talk) 01:37, 11 June 2017 (UTC)

The article is not useful and the content is problematic

"Testing hypotheses suggested by the data" is not a notable phrase in itself, and as a concept is not sufficiently distinct from post hoc analysis (which already has a wikipedia article) to merit its own article. Moreover, the description in the "general problem" section appears to describe data dredging---which also already has a wikipedia article. And the example in the "example" section doesn't even involve hypotheses suggested by the data; rather it is an example of multiple studies testing the same A PRIORI hypothesis. The section then adds to confusion by conflating the topic with publication bias. The "correct procedures" section is misleading and gives odd weight to the Scheffé procedure, which is essentially obsolete and was only ever applicable in a specific type of design anyway. It is also notable that, notwithstanding the Scheffé paper, the article is completely unsourced. In short, the article appears to be redundant (to other articles) in its scope and highly problematic in its content.

My recent proposal for deletion was immediately removed on the basis that a previous proposal resulted in a decision to keep (https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletion/Testing_hypotheses_suggested_by_the_data). However, that discussion---which apparently involved only two editors---did not conclude that the article should be kept per se. Rather, the discussion concluded that the appropriate solution was to redirect to the data dredging article. That would be preferable to keeping this problematic and unnecessary article. 23.242.198.189 (talk) 06:25, 13 November 2019 (UTC)