A Warning for Big Data, 3 Ways to Handle Citizen Feedback, and Why "Experts" Are Often Wrong
All the public-sector management news you need to know.
North Carolina’s Medicaid program had cost overruns in its last fiscal year amounting to some $250 million -- and it’s not alone. Back in December, the National Conference of State Legislatures reported that ten states (compared to just six in the previous year) were over budget in Medicaid and other public health spending.
As states close the fiscal year’s books, we wouldn’t be surprised if that number is somewhat higher because many states have been engaged in health-care cost control efforts in this last year and we know from experience that estimates of savings are often overly optimistic.
"I made a commitment to Pennsylvanians, and their elected officials, to create a map showing the benefits of additional transportation funding because Pennsylvanians have a right to see how their precious dollars are being invested," said the secretary of the state’s Department of Transportation, Barry Schoch, in a statement. Take a look at the map and see what you think. It lets users see the state’s projects and sort them by county, partners, state Senate district or state House district.
The map is a valuable tool and we congratulate the secretary for providing it. On the other hand, we can’t really see that it actually delivers on his claim of showing the benefits of individual projects beyond a brief document that makes the case for the state’s transportation funding plan couched in terms of “the cost of doing nothing.” Maybe state leaders are thinking about providing the benefits of specific projects in the future. Wouldn’t that be great?
"You can have all the facts you want but if you can’t tell a story, people won’t listen." – Shane Parrish on the Farnam Street blog.
City and states are awash with data. One obstacle to translating that data into information useful for management is an apples-and-oranges kind of phenomenon: Information in one dataset frequently can’t be effectively meshed and analyzed with that in another. When this challenge is overcome, though, good things can happen, and New York City’s tree pruning efforts illustrate this.
According to Network World, there are about 2.5 million trees in the Big Apple. In order to avoid major storm cleanups and fatal accidents from heavy limbs falling, the trees must be properly pruned. The idea that proper pruning was worth the effort made sense, but there was no statistical evidence to see just how much good it did. The city had lots of data about pruning and emergency calls, but it wasn’t able to see how they linked. When it turned to a contractor with the technological savvy to make both datasets function together, the city discovered that carefully pruning for specific hazards caused a 22 percent decline in the frequency with which the city had to send in an emergency cleanup crew.
“Big data” is all the rage, and we’re thoroughly excited by the notion that super powerful technology combined with huge databases and strong analysis can do a lot to manage well. But we’d throw a little cautionary note into the maelstrom of excitement: Big data by itself won’t do a bit of good for any government unless it’s actually used for management or policy. We’re worried about that given the way mountains of well-done analyses issued by the likes of legislative audit offices, comptrollers and so on have frequently been ignored by decision-makers.
In 2011, California began "realigning" its correctional system, shifting many nonviolent offenders from state prisons to county jails. It’s resulted in a 16.5 percent growth in jail population since May 2011, according to the Public Policy Institute of California. That much growth has led to overcrowding (which, ironically, was the reason for realignment), and the state is making $1.7 billion available for new county jail construction.
Here’s the hitch: Construction is less than 10 percent of the cost of operating a jail over its lifetime. So, concludes the report, this all “strengthens the incentives for counties to seek alternatives to incarceration.”
Mindmixer is an online platform where public-sector officials can hold conversations with citizens for a variety of purposes. Its CEO, Nick Bowden, is fascinated with his work -- and with ways that cities and states can interact with citizens online. We asked him to suggest three of the most useful ways for cities to make the best use of citizen feedback. Here’s what he suggested:
- “Ask questions that people want to respond to,” he says. The key here is to avoid language that won’t be easily understood by the audience the city wants to reach. A question about “land use,” for example, may well miss the mark for many, while one that relates to commercial, residential and industrial property won’t.
- There should be clarity of purpose and benchmarks for success. Cities need to know what they really want to get out of the conversation and have ways to see how successful they were at achieving those goals.
- Finally, it’s critical to let citizens know what the status of their feedback is. Even when the city can’t directly fix a problem, if a citizen knows he’s been heard and taken seriously, “that’s the reason people come back.”
Why are expert predictions inaccurate so frequently? Maybe it’s because there’s no certification people have to possess to be referred to as an “expert.” But here’s a thoughtful entertaining 15-minute video with political forecaster Nate Silver on why expert predictions are so often wrong and how adherence to theory can skew interpretation of data. This is one of many short and informative videos that we've enjoyed from the RSA, a British organization "committed to finding innovative practical solutions to today’s social challenges."
The Government Accountability Office published the results of a survey of federal managers in late June. One of the most interesting questions it asked was to what extent certain factors “hindered measuring performance or using the performance information.” The three factors that the most respondents thought hindered performance measurement “to a very great extent” were: lack of incentives, difficulty determining meaningful measures, and different parties using different definitions to measure performance. We can’t be sure, but it’s our bet that responses would be similar in states and cities.
State and county level inspectors general reports can be a wonderful source of information for anyone interested in understanding how things really work (or don’t work) in public-sector management. In fact, we wrote about this very topic in the July issue. If you’re intrigued, we recommend the website for the Association of Inspectors General.