Friday, May 22, 2009

Declining crime would be better news if data weren't corrupt, incomplete

As has been discussed many times on this blog, data in Texas about both ongoing crime and criminal histories are incomplete and often corrupt, with wide variations from county to county and department to department regarding what gets reported and how. Sometimes what looks like a big reduction in crime may result from minor changes in reporting rates, while conversely, simply reporting all the data instead of just a portion of it might give a false appearance crime has increased.

So when I see headlines like the ones yesterday claiming a 3% drop in crime statewide, I take it with a grain of salt. Much more important is the overall trend - the Dallas News says there have been drops in seven of the last ten years, which makes me comfortable concluding crime is indeed declining, even if I don't 100% accept the measuring stick being used.

Even so, much of this data is suspect. Dallas is claiming large crime reductions, for example; 10% down in the fiscal year measured. But reminiscent of scenes out of The Wire, they've been jiggering their definitions to give the appearance of less crime, regardless of the overall level of victimization. By contrast, San Antonio saw property crimes spike at rates that seem inexplicable given trends everywhere else in the state. Who knows what's behind that? It could result from more crime, but could also be explained by more reporting of crime or improved data processing by police. Ditto for crime spikes in Austin's suburbs.

The best approach IMO is not to react much to this particular metric: Taking credit for statistical crime reduction sets law enforcement up to take blame when crime goes up, when really the rate of offenses fluctuate independently from their actions based on many other variables, particularly for homicides and other violent crimes.

Relatedly, I'm hopeful the Legislature may address equally flawed reporting by counties of dispositions in criminal cases, that is, if the House can hustle through a whole bunch of bills ahead of it to get to SB 1061 by Sen. Florence Shapiro, which just made out of House Calendars. That bill (discussed earlier on Grits here) would require counties with rates of reporting case dispositions of lower than 90% to create a data improvement plan to rectify problems with case reporting. This is especially critical in cases where charges are dropped; often counties will report when initial charges are filed but don't follow up when the case is dismissed, leaving a record that invites misreading in the state criminal history database. Good job by Sen. Shapiro to proactively address it.

If she's still with us at the capitol then (she's announced she'll make a run for the US Senate if Kay Bailey Hutchison resigns), I wish Sen. Shapiro would pursue a similar program for improving Uniform Crime Reporting data from Texas law enforcement agencies. Without a better effort to standardize definitions and strong incentives for agencies to comprehensively report crime data, these numbers will continue to be quirky, incomplete and more or less incomparable city-to-city or year-to-year.

See the full report.

5 comments:

Adam said...

Corrupt data. In Texas? Now I have heard everything ROFLMAO!

Anonymous said...

Thanks for the continued reporting of this issue. I work for a company that often seeks to obtain court and criminal records. And I am often the one doing the searching.

It might be of interest to your readers to know that both the federal and state governments routinely sell this information to 3rd party vendors. These are typically very reputable companies that do their best to verify the accuracy of the data. And yet, because of the very real possibility of corrupt or fallicious data that is given them, a very lengthy disclaimer is given before a search even begins.

I'd also like to note that they provide a detailed description of the date range that is included in the data, as well as, the number of counties reporting within the state. Something that I found nowhere on the DPS site.

I'd also like to mention the technical deficiencies that are likely to occur. Without proper standards, indexing, or coding, even if the raw data is there, the likelihood that all data will be retrieved is suspect.

We often find it necessary to consult several different sources before a learned conclusion can be made. And even then, sometimes only a phone call will verify the information.

P.S. I'd like to know why TDCJ is selling data to a company that targets families of those who are incarcerated who are so overwhelmed by the mysteries of the parole process that they are willing to pay for information from a company that I can only describe as inflating their claims of providing accurate and prudent information! Riddle me that Grits!

Anonymous said...

"Those who rise by the numbers are beholden to them"

george said...

"P.S. I'd like to know why TDCJ is selling data to a company that targets families of those who are incarcerated who are so overwhelmed by the mysteries of the parole process that they are willing to pay for information from a company that I can only describe as inflating their claims of providing accurate and prudent information! Riddle me that Grits!"

WHEN WILL THE SHADINESS IN TEXAS STOP? EVER? IT SURE GETS TALKED ABOUT AND THEN EVERYONE JUST GOES ON THERE MERRY WAY. CROOKS...EVERYWHERE!!!!

Anonymous said...

Actually I think the politicians and cops are the most corrupt in this country. We only hear what they want us to know. Filtered news etc, What a joke newspapers and TV news is. Don't question their authority, you may be next on their list.