We've been talking about measuring poverty here, here, and here. Measuring child poverty is even more difficult.

Some researchers have tried to discover better child poverty indicators through asking children for their point of view. This wasn't as helpful as they thought it might be -- these children in England ranked the lack of cell phones the key indicator of child poverty. (They may be right in some cultures -- how about the numbers of five-to-nine year olds with cell phones in Japan?)

In the United States, trying to understand what we know about child poverty indicators is hard work. Douglas Besharov's Poverty Update from September 27, 2007, is helpful in explaining what the numbers we use really mean, and why they aren't sufficient. His efforts to get behind the numbers yielded some interesting results, particularly on the racial/ethnic shifts of women employed in skilled blue-collar employment and the impact on racial disparities in child poverty.
I hadn't realized that Connecticut has mandated reducing child poverty by 50 percent by 2014, which is an interesting idea -- can you legislate away poverty? And Minnesota has launched a commission to end poverty by 2020 -- what comes out of these efforts may be incredibly useful for community work, depending on what they accomplish. Just declaring war on poverty didn't make it all go away.

Earlier this year, the United Nations adopted a new definition of child poverty, one that went beyond a family income definition. UNICEF added some thoughts that might be useful in exploring community indicators of child poverty:

Children’s well-being relies in large part on the availability and quality of basic services and an environment for play and leisure. Access to these does not always depend on family income but on the priorities and investments of the state. Lastly, income poverty assumes that all family members have an equal share of the family’s income, which is often not the case, particularly for girls. ... If poverty is understood as more than just income poverty, then responses need to address the broader picture of children’s experience of poverty.

How do you measure child poverty? Any suggestions for other communities?

I have to pass on one more item from Harvard's Social Science Statistics Blog. The article -- "How Do You Get 7,000,000 Cell Phone Records?" -- discussed this presentation by David Lazer.

Wow.

We've been talking about big numbers and new data sources and using traditional data in new ways and even using cell phones as data sources, but this goes way beyond.

Beyond beyond.

How about using the records of 7,000,000 cell phone users (including 49 trillion conversations!) to map social networks? What if you then followed up that work by tracking (participating) students for a month, paying attention to both phone conversations and physical proximity -- and used that data as a predictor of friendship patterns? What if you took that research next to Washington?

Amazing information that you need to check out, if only to put our community work and local indicator projects in context with the possible.

I had mentioned the Social Science Statistics blog earlier, but since the blog is more active now after summer vacation, I thought I'd bring it to your attention again.

It's from the nice people at Harvard's Institute for Quantitative Social Science. Here's why you should pay attention to it, from its own description:

This blog makes public the hallway conversations about social science statistical methods and analysis from the Institute for Quantitative Social Science and related research groups. Expect to see posts on trends in methodological thought, questions and comments, paper and conference announcements, applied problems needing methodological solutions, and methodological techniques seeking applied problems. Also included are summaries of papers and comments from a popular weekly research workshop held here and billed as a tour of Harvard's statistical innovations and applications with weekly stops in different disciplines.

Those are great conversations to listen in on, especially without paying tuition. A recent entry caught my eye and I thought community indicators practitioners might be particularly interested, since it deals with problems we have in comparing self-report health data to get a real sense of what's happening in communities.

The article is titled Health Inequities and Anchoring Vignettes, and it describes a technique for having survey respondents classify the health status of a series of short descriptions of someone else's health functioning before rating their own. In this way, differences in expression or culture aren't treated as real differences in health status.

Read the article, as well as these two earlier articles, to see how useful this technique might be for your community's health assessment -- and for a better understanding of how to use data from other communities as comparatives.

The good folks at Swivel have made it easire than ever to add graphs to a blog entry.

You begin by either creating or selecting a graph at Swivel. I've been talking with folks about global demographic trends and their potential impacts in local communities, so I chose fertility rates in OECD countries.

The next step is to look under the graph at "Share this graph". Under the send-to-e-mail link is "post to blog" -- they make it pretty easy. When I click that, I get a page with a helpful note pointing out which text to copy and paste into HTML, and some options on what size graph I want. I choose medium, and try below:



That seems a little cluttered, so I refocus on just U.S. fertility rates:



Then, because I'm interested in sparklines, I try a sparkline version of the same graph:

This turns out to be as simple as Swivel promised. Send me your examples of using graphs in your blogs. In the meantime, I'll gather a bit more information on the global trends we're looking at, and an upcoming conference to discuss how they effect local community planning effort.

He's questioned the data in the salsa v. ketchup debate. He's challenged the costs of cybercrime.

In Bringing Big Numbers Down to Size, the Numbers Guy takes on the concept of big numbers. How often do we use large numbers and statistics to make a point, and lose the audience because the numbers just don't make sense to them?

So we sometimes use analogies and descriptors and examples that describe dollar bills stretching to the moon or lines of people wrapping around the earth or similar kinds of things.

But there are some really innovative ways to portray large numbers. Check out how artist Chris Jordan depicts nine million children without health insurance.

And reader ChuckP adds that:

Most solutions to the difficulty conceptualizing numbers don’t work because they simply compound the problem. A stack of bills to the moon equals x dollars or the federal budget etc., etc.? Such images fall flat because it tries to solve the problem by introducing - more numbers!
Any solution probably requires two things - deeply held intuitive conceptions and a few SIMPLE numbers - but only if absolutely necessary.

My best attempt was an informational leaflet at describing how much more CEOs now make than the average Joe: “There are CEOs that make in one hour what takes many of us a whole year to earn.”

It’s very effective. People don’t forget it because it’s about one of the most important weekly events - the size of the paycheck for 40 hour’s work - and how it compares with the paychecks of other people we know. Imagine! - the example illustrates scale with no numbers at all!
It took me a long time to come up with the CEO example. It seems that every time I wanted to think about it, my mind automatically turned to the hopeless abstraction of numbers. I really had to work at getting into another frame of mind.

I know we've talked about the problems with using numbers to tell stories. And many times scale is even harder to show.

What are your best ways you've used in community indicators systems to help people understand large numbers?