Racial Bias by Police?

Do you think there's widespread bias and violence by police officers against blacks and other minorities? What you suspect and what you can prove are two different things.

In the midst of much lamented acts of police violence and equally lamentable reprisal comes a paper from the National Bureau of Economic Research titled An Empirical Analysis of Racial Differences in Police Use of Force. Among the media and related punditry are some claiming this study as ‘proof’ of the absence of racism in policing.

Yet the paper itself shows why that conclusion may be erroneous, starting with a frank notice on page one that it’s a working paper that hasn’t been peer reviewed. Other caveats start on page 7. (Page 7 of the PDF actually displays the number 5 on it’s bottom.) Caveats include problems with their source data and a concern that their methodology is inadequate. Snopes.com catalogs many other problems noted by critics and others.

I know what my gut's telling me about the veracity of the NERB paper. If this were an opinion piece I'd share that opinion. A New York Time story on the paper opens with the unequivocal “a new study confirms”. If the Times actually believes that then they need to either read more carefully or acquire a better understanding of how scientific research works. If they understand how science works, they're editorializing. in which case their whole article should on the opinion section, not posted as news.

Refusing to Research

A few years ago, a Facebook acquaintance posted an article stating that President Obama had failed to give a Memorial Day speech at Arlington National Cemetery three years running, and that he was the first president in modern history to do so. The acquaintance stated he didn't need to "research" it (quotes his) because he knew it was true.

One thing struck as wrong, immediately. After observing politicians for a lifetime, including six as a super-volunteer where I could get to know a few and observe them at close range, it struck me as unlikely that any politician, no matter what his character, would risk alienating every veteran in America by skipping a memorial day appearance at Arlington. Whatever one's character, one does not get within spitting distance of the White House by being politically stupid.

But how to confirm or refute? I quickly found that Snopes.com had an article on this claim from 2010. That told me the claim's been floating around a while, but Snopes is not a primary source. An even better place to look would be whitehouse.gov, where they publish not only the President's public schedule, but the text of every public statement. Under May 28, 2012 I found the text of one speech given by President Barack Hussein Obama at Arlington National Cemetery. I guess you could claim he posted a speech he didn't actually make. Now, you're just being silly. Republicans would certainly have made hay of such a thing. As I said, one does not get within spitting distance of the White House by being politically stupid.

You Can't Get There from Here, Part II

It takes me a while to write one of these posts. If I published what came off the top of my head, it would be easy. I want to make sure I can back up anything I say and that takes time.  That path was my choice, so I'm not complaining.

There is a bramble in the path that I will complain about: access to research. Here's a case in point. The American Journal of Public Health released a study on Friday reporting a link between areas of high gun ownership and police deaths. Even before it's out, it's a political football. I want to evaluate it for myself. I don't want to read any of the million opinion pieces saying why it's a good study or why it's a bad study. I want to see the study.

In particular, I want to see the data. One site has helpfully provided a graph of the data. But I don't know anything about this site. Statistics aren't necessarily lies as the popular meme started by Benjamin Disraeli would have it. They're forms of communication. They can be lies or they can be the truth. This news site is new to me. I don't know if they've reproduced this graph straight out of the study or if they've shaded it in some way. If I could see the data, I could reproduce the graph, including the regression line. It's easy to do in Google Spreadsheets or Microsoft Outlook. Reproducing it would allow me to figure out what if anything they had done to shade it. 

Can I get to the study or any of its data? If I follow one of the many links to the study, I get this, a request for money (top image):

This is actually an improvement over what I would have gotten five or ten years ago. News sites wouldn't link to studies. Often they wouldn't even provide the name of the paper. If I wanted to find it, I'd need to search through a dozen articles, hoping someone would provide the study's name. Or I'd need to search on the study's subject along with the organization that published it. 

This particular site wants $22 just for the article. That's a bargain. I've seen sites that want north of $30 per paper.

So, I can view the opinions of thousands, even tens of thousands of blowhards for free, but I have to pay to see the work of people who do real research? "But!", I can hear you asking, "you're not a an expert in this field. You're not qualified to judge the veracity of the research." But I know people who are. I could ask them to fact check my posts. Better yet, I could ask them to write a piece for my blog. 

We live in an era that is increasingly anti-science. While bloviators proliferate, the way research is published hamstrings the voices of knowledgeable people. What if all qualified professionals could view this study for free? We'd potentially have an army of qualified debaters for every comment thread, every family gathering, every city council meeting. That's not to say we would have this in every case. But it would be better than what we've got.

You Can't Get There From Here

A few days after posting History On Line more items showed up in my news reader describing online archives of various flavors. I started thinking about problems I've had with other online archives. Rather than explain it, I'm going to ask you to do it.

Let's start with the Dead Sea Scrolls archive, a web site containing images of the scrolls put up by he Israeli Antiquities Authority.  Go to the archive and search on the "The Damascus Document".  Let's pretend you can actually read the text on the photos of these ancient documents. Why? Because the relevant portion of this discussion isn't nearly as cool as Dead Sea Scrolls.  I"ll get to that in a minute.

Now go to Google and search for "Dead Sea Scrolls."  Notice that the archive we just looked at is the second result, right under the Wikipedia article on the same.

Finally, do a Google search for "The Damascus Document."  Notice what you didn't find.  The Israeli archive is nowhere to be seen.  I clicked through five pages of search results and never found the Israeli archive.  I don't know if Google's results are the same at all times and all places, so your results may vary. I think my results demonstrate the point.

I put a premium on source materials. I hope I don't need to explain why. Even in the age of Google, you may not find source materials unless you can guess what web site might have it. This problem isn't confined to databases of ancient history.  More often than I'd like, finding the right answer to something means finding the right database.

Let me give you some every day examples. Who owns the derelict property down the street? Turns out, it's owned by a paper corporation.  Who really owns the paper corporation? Can I do a background check on someone without paying for it? Maybe. Do you have time to find all the databases you need to check in all fifty states, then check all those databases individually? (On reflection, this last example might be a good thing. It keeps casual snoopers from poking around my records.)

The phenomenon I demonstrated with the Dead Sea Scrolls archive is one I first observed while searching property records and doing background checks. To find what I was looking for I had to first find the correct database.  Expand this to anything you want to learn or prove.  Want to show, for example, the religious beliefs of the founding fathers? You need their actualy writings.  An article by a historian, or worse, a blow-hard pundit isn't enough. 

What would be nice would be to search engine that find repositories rather than text strings.

History On Line

I recently watched the documentary "Titanic's Final Mystery" in which historian Tim Maltin uses historical sources to tease out new information about the tragedy of said ship's sinking nearly a century ago. To get a sense of weather conditions in the North Atlantic, in particular on the night of the sinking, Maltin reviews the logs of ships that crossed the ocean in April of 1912. NOAA's ships logs were, at least, transferred to computer punch cards in the 1970s. Records of German ships are still kept in their original log books at an archive in Germany. This is important because in historical research the only thing that counts is original records.

What jumped out at me was how he viewed them. In neither case were these records available on line. Maltin had to physically travel to their storage facilities to examine them.

Someday, I need to do a post about how 'proving' a point in the internet age still requires going beyond Google, sometimes way beyond Google. At least, that will be the case until all the paper in all the obscure archives is digitized.

Paper is one thing. Our intuition expects that paper will transferred to computer. But artifacts? This week I learned that museums are starting to put three dimensional scans of artifacts on line. Until now the public's access to archaeological relics has been restricted to views behind ropes and through panes of glass. Only top scholars get the privilege of turning these objects over and looking at them closely.

That isn't likely to ever change. But with online scans and now burgeoning virtual reality technology, exemplified by the recently released Oculus Rift, you and I will eventually get the chance to examine ancient artifacts.

The latest museum to enter the 3D club is the University College London's Petrie Museum of Egyptian Archaeology. Never heard of it? Among Etymologists, UCL ranks right behind the Cairo Museum and the British museum in the number and quality of ancient Egyptian artifacts.

You can view UCL's collection here. Meanwhile, I'll be looking for links to other collections. I hope someone has built a specialized search engine. I doubt this is on Google's radar yet.