Person:

King, Gary

The increasing availability of digitized text presents enormous opportunities for social scientists. Yet hand coding many blogs, speeches, government records, newspapers, or other sources of unstructured text is infeasible. Although computer scientists have methods for automated content analysis, most are optimized to classify individual documents, whereas social scientists instead want generalizations about the population of documents, such as the proportion in a given category. Unfortunately, even a method with a high percent of individual documents correctly classified can be hugely biased when estimating category proportions. By directly optimizing for this social science goal, we develop a method that gives approximately unbiased estimates of category proportions even when the optimal classifier performs poorly. We illustrate with diverse data sets, including the daily expressed opinions of thousands of people about the U.S. presidency. We also make available software that implements our methods and large corpora of text for further analysis.

We highlight common problems in the application of random treatment assignment in large-scale program evaluation. Random assignment is the defining feature of modern experimental design, yet errors in design, implementation, and analysis often result in real-world applications not benefiting from its advantages. The errors discussed here cover the control of variability, levels of randomization, size of treatment arms, and power to detect causal effects, as well as the many problems that commonly lead to post-treatment bias. We illustrate these issues by identifying numerous serious errors in the Medicare Health Support evaluation and offering recommendations to improve the design and analysis of this and other large-scale randomized experiments.

Background: We assessed aspects of Seguro Popular, a programme aimed to deliver health insurance, regular and preventive medical care, medicines, and health facilities to 50 million uninsured Mexicans. Methods: We randomly assigned treatment within 74 matched pairs of health clusters—ie, health facility catchment areas—representing 118 569 households in seven Mexican states, and measured outcomes in a 2005 baseline survey (August, 2005, to September, 2005) and follow-up survey 10 months later (July, 2006, to August, 2006) in 50 pairs (n=32 515). The treatment consisted of encouragement to enrol in a health-insurance programme and upgraded medical facilities. Participant states also received funds to improve health facilities and to provide medications for services in treated clusters. We estimated intention to treat and complier average causal effects non-parametrically. Findings: Intention-to-treat estimates indicated a 23% reduction from baseline in catastrophic expenditures (1·9% points; 95% CI 0·14–3·66). The effect in poor households was 3·0% points (0·46–5·54) and in experimental compliers was 6·5% points (1·65–11·28), 30% and 59% reductions, respectively. The intention-to-treat effect on health spending in poor households was 426 pesos (39–812), and the complier average causal effect was 915 pesos (147–1684). Contrary to expectations and previous observational research, we found no effects on medication spending, health outcomes, or utilisation. Interpretation: Programme resources reached the poor. However, the programme did not show some other effects, possibly due to the short duration of treatment (10 months). Although Seguro Popular seems to be successful at this early stage, further experiments and follow-up studies, with longer assessment periods, are needed to ascertain the long-term effects of the programme. Funding: Mexican Ministry of Health, the National Institute of Public Health of Mexico, and Harvard University Institute for Quantitative Social Science

In the election for President of the United States, the Electoral College is the body whose members vote to elect the President directly. Each state sends a number of delegates equal to its total number of representatives and senators in Congress; all but two states (Nebraska and Maine) assign electors pledged to the candidate that wins the state's plurality vote. We investigate the effect on presidential elections if states were to assign their electoral votes according to results in each congressional district,and conclude that the direct popular vote and the current electoral college are both substantially fairer compared to those alternatives where states would have divided their electoral votes by congressional district.

Social science data collected in the United States, both historically and at present, have often not been placed in any public archive -- even when the data collection was supported by government grants. The availability of the data for future use is, therefore, in jeopardy. Enforcing archiving norms may be the only way to increase data preservation and availability in the future.

In response to its poor performance during the 2012-2013 flu season, Google Flu Trends (GFT) engineers announced a redesign of the GFT algorithm. Two changes were made: (1) dampening anomalous media spikes and (2) using ElasticNet, rather than regression, for estimation. This paper identifies several problems that persist in the new algorithm. First, the transparency problems identified in our earlier Science paper appear to have, if anything, become worse. Second, there are reasons to doubt whether a spike in media attention was the only, or primary, cause of GFT's errors. Finally, there is strong evidence that GFT is still not using all the information at its disposal to make accurate measurements of flu prevalence. While it is too early to give a complete evaluation of the new algorithm, these results are discouraging.

The financial viability of Social Security, the single largest U.S. government program, depends on accurate forecasts of the solvency of its intergenerational trust fund. We begin by detailing information necessary for replicating the Social Security Administration’s (SSA’s) forecasting procedures, which until now has been unavailable in the public domain. We then offer a way to improve the quality of these procedures via age- and sex-specific mortality forecasts. The most recent SSA mortality forecasts were based on the best available technology at the time, which was a combination of linear extrapolation and qualitative judgments. Unfortunately, linear extrapolation excludes known risk factors and is inconsistent with long-standing demographic patterns, such as the smoothness of age profiles. Modern statistical methods typically outperform even the best qualitative judgments in these contexts. We show how to use such methods, enabling researchers to forecast using far more information, such as the known risk factors of smoking and obesity and known demographic patterns. Including this extra information makes a substantial difference. For example, by improving only mortality forecasting methods, we predict three fewer years of net surplus, $730 billion less in Social Security Trust Funds, and program costs that are 0.66% greater for projected taxable payroll by 2031 compared with SSA projections. More important than specific numerical estimates are the advantages of transparency, replicability, reduction of uncertainty, and what may be the resulting lower vulnerability to the politicization of program forecasts. In addition, by offering with this article software and detailed replication information, we hope to marshal the efforts of the research community to include ever more informative inputs and to continue to reduce uncertainties in Social Security forecasts.

Existing research on the extensive Chinese censorship organization uses observational methods with well-known limitations. We conducted the first large-scale experimental study of censorship by creating accounts on numerous social media sites, randomly submitting different texts, and observing from a worldwide network of computers which texts were censored and which were not. We also supplemented interviews with confidential sources by creating our own social media site, contracting with Chinese firms to install the same censoring technologies as existing sites, and—with their software, documentation, and even customer support—reverse-engineering how it all works. Our results offer rigorous support for the recent hypothesis that criticisms of the state, its leaders, and their policies are published, whereas posts about real-world events with collective action potential are censored.

We introduce a method for estimating incidence curves of several co-circulating infectious pathogens, where each infection has its own probabilities of particular symptom profiles. Our deconvolution method utilizes weekly surveillance data on symptoms from a defined population as well as additional data on symptoms from a sample of virologically confirmed infectious episodes. We illustrate this method by numerical simulations and by using data from a survey conducted on the University of Michigan campus. Last, we describe the data needs to make such estimates accurate.