Friday, April 26, 2013

Data link roundup (week of April 26, 2013)

The week's top data analysis links...

This week's theme: uncertainty and contradictory findings in research.


GAZE INTO THE CRYSTAL BALL...

Using the demographic shifts in Washington, D.C. as an example, Megan McArdle explains why forecasting is so difficult.


SAME DATA, OPPOSITE CONCLUSIONS

Paul T. Williams "achieved a feat that's exceedingly rare in mainstream science: He used exactly the same dataset to publish two opposing findings." Williams used data on exercise habits to show both that walking is better than running AND that running is better than walking. Read Daniel Engber's article in Slate to find out why "flip-flopping advice on exercise may not be as contradictory as it seems."


RETRACTIONS NEEDED?

Recently two famous papers generated great controversy when their findings were discovered to be false.

Inaccuracies in the Reinhart/Rogoff debt paper has serious public policy implications, just as undisclosed data sources and specious findings in the Regnerus paper has serious social ones.

Scatterplot lays out the debate about when to retract...


BEST CHART QUOTE OF THE WEEK: DEMOGRAPHY IS (NOT?) DESTINY

In a cleverly-written piece, Dr. John Weeks explains that demography is not destiny, but it does shape the future...
"the idea that demography is destiny does not mean that you can predict the future based on demographic trends. More subtly, and more importantly, demography shapes what the options for the future will be."


Friday, April 19, 2013

Data link roundup (week of April 19, 2013)

The week's top data analysis links...

Last week the Population Association of America conference wrapped up in New Orleans. This week's theme, of course: population models and methods.


NEW TECHNOLOGIES

One of the best sessions I attended was "Social Media, Digital Tracks and Demography." Papers showed how demographic researchers might use Twitter to collect research data, how cell phone positioning systems might be used to better understand human mobility patterns, and how Facebook connections could be used to understand the proximity of people in networks.


METHODOLOGY COMPENDIUM ONLINE

While this wasn't presented at the conference, it's usefulness for demographers is worth noting: The United Nations has made its series of demographic methods manuals available online.


... AND MY CONFERENCE POSTER

Estimating households by household size for small, sub-county areas using the Poisson distribution: