Page 1 of 1

Published Studies - What do I do with them?

Posted: Thu May 31, 2018 10:45 am
by josher89
I don't remember where I got this from, so sorry if it was you and I didn't give you credit. This provides some decent food for thought when reading research articles. There seems to be a lot of research coming out regarding our discipline and this will give you some guidance on how to masticate on them before claiming them as the gospel or outright discrediting them immediately based on the author.

Saini_Tips for making sense of published studies_Final 120911.pdf

Once you've had a chance to read it, are there any research articles that you have read that are one-sided or promote a particular agenda in a "partisan" way?

Re: Published Studies - What do I do with them?

Posted: Wed Jun 06, 2018 8:21 am
by Boyd Baumgartner
Thanks for posting this Josh, I tend to think of garden variety Examiners like myself as the ultimate consumer of science in the discipline. As such I think it's healthy to view any publication through the lens of Richard Feynman's famous quote:
Science is the belief in the ignorance of experts
By ignorance, I believe he means fallibility. This is especially healthy in our line of work where the dogma in the literature and culture was that fingerprint identification was infallible and the expert was unassailable. A look at many of the critiques still reference this historical claim of infallibility and it's offspring, overstating conclusions.

That being said, I think you need to characterize categories of published works. We have five main categories in my estimation. Physical Science publications, Psychological Science publications, Technological Publications , Policy Publications and Academic Publications.

Physical Science publications are the most straightforward and would include chemical validation studies like the effectiveness of an HFE7100 formulation of Indanedione. These studies might present themselves in the discipline specific publications like the JFI or other similar peer reviewed publications.

Psychological Science publications are a bit more thorny but can also present themselves as performance studies (black box, white box, etc) and/or expertise and will also be found in peer reviewed publications like PLOS One and Elseiver (these will contain many types of publications including physical science and technological publications as well)

Technological publications include using technology to advance the discipline in some capacity, either through image processing technology, orAFIS style detecting, feature extraction and matching technology. These types of papers are typically found in trade specific journals under the banner of discipline specific associations like the IEEE

Policy publications include the bigger aggregate analyses that want to make sense of trends and gaps within the discipline usually to inform policy makers and influence policy decisions. This includes the NAS Report, the Human Factors Report and the PCAST Report. These publications make use of peer reviewed publications listed above and will include recommendations based upon how these studies are interpreted.

Academic Publications include your text books like those from CRC Press or academic collaborations with government funding like The Fingerprint Sourcebook. These will be a mix of evangelism, academics and scientific translation of physical science, psychological science, technology and policy papers.

That being said, I think you need to keep in mind that there are some prevailing pernicious problems that permeate the perspectives (hello alliteration). Chief among them is the publish or perish mentality of the publication world which biases research towards the novel and towards affirmative findings over replication and null findings. This has lead directly to what is known as the 'Replication Crisis' and is going to be the most problematic in areas like bias studies, sufficiency studies, performance studies and other psychometric studies which form the basis for error rates according to the PCAST reports. I would say Dror's original bias paper is the poster child for such a phenomenon. Is it methodologically sound? Has it been replicated? Is it biased itself towards finding an effect?

It would seem too that we have somewhat of a problem with pseudo-fame as well, in which seemingly good actors want to rescue the discipline from itself. Just look at who are the prominent speakers year after year in the conferences and their 'one size fits all' solutions. As the article which helped launch the replication crisis movement states:
Scientific claims should not gain credence because of the status or authority of their originator but by the replicability of their supporting evidence
Anyway, what you posted is definitely of value and that's how I go about supplementing those techniques with the broader aim of putting findings in context to the discipline.