View Single Post
Old 01-05-2019, 04:13 AM   #377
DeluxeMoustache
 
DeluxeMoustache's Avatar
 
Join Date: Mar 2010
Exp:
Default

Quote:
Originally Posted by Textcritic View Post
I think this is your first mistake. "common sense" is always a pretty poor exploratory method with the litany of other much more effective analytical tools at ones disposal. We retreat to common sense in instances in which we do not have adequate information upon which to make good decisions. This is one of those instances
Paging ‘This Post is Terrible’.

Brutal oversimplification.

Common sense is not necessarily something to which one retreats when ‘adequate information’ is not available. It should be considered and included or disqualified, with reasons, in the assumptions and scope of work when doing any analysis. For people who are involved in research, summary dismissal of potentially relevant information is definitely poor practice.

It is particularly a problem with people who are involved in data analytics in a professional capacity, that they can selectively include and dismiss data that suits their purposes. Academia is often competitive and agenda driven as well.

Before I go further, maybe you need to define ‘common sense’ and give examples of it and its failures in the context of this discourse. And also define ‘effective’. There is a lot to unpack in this paragraph of what looks like a lot of hooey.

Look at these guys who are trying to build careers on analytics. Many are doing the test they can with the data available, and failing, but selling themselves like snake oil salesmen. There are limitations which are material, they are doing the best they can with the info they have available, and they are often glossing over limitations of their methods. I have read Vollman’s book and while I am adequately impressed with the accomplishments, I am equally disappointed with the importance they attribute to results that are statistically relatively insignificant.

You know who embraced methods believed to be effective over common sense? The Oilers* with coach Dallas Eakins and analytics guru Tyler Dellow. It was a disaster.

A bloody huge part of data analysis is understanding relevance and context.

*Enjoyable for Flames fans so I am not complaining.
DeluxeMoustache is offline   Reply With Quote
The Following User Says Thank You to DeluxeMoustache For This Useful Post: