Differential privacy is widely accepted as a powerful framework for providing strong, formal privacy guarantees for aggregate data analysis. A limitation of the model is that the same level of privacy protection is afforded for all individuals. However, it is common that the data subjects have quite different expectations regarding the acceptable level of privacy for their data. Consequently, differential privacy may lead to insufficient privacy protection for some users, while over-protecting others. We argue that by accepting that not all users require the same level of privacy, a higher level of utility can often be attained by not providing excess privacy to those who do not want it. We propose a new privacy definition called personalized differential privacy (PDP), a generalization of differential privacy in which users specify a personal privacy requirement for their data. We then introduce several novel mechanisms for achieving PDP. Our primary mechanism is a general one that automatically converts any existing differentially private algorithm into one that satisfies PDP. We also present a more direct approach for achieving PDP, inspired by the well-known exponential mechanism. We demonstrate our framework through extensive experiments on real and synthetic data.

Compared to the extensive study on the effects of Medicare payment in hospital markets and on physician behavior, almost no work has been done on the effects of Medicare hospital payment on the labor market for nurses. This study deals with the hospital wage index (HWI) adjustment to Medicare payments, a geographic adjustment intended to compensate hospitals in high-cost labor markets, and examines two potential consequences of HWI adjustment on nurses labor. First, we examine whether hospitals in highly-concentrated markets exploit their ability to influence their own wage index by paying nurses more. Second, we examine whether the Occupational Mix Adjustment (OMA) to the HWI affected nurse employment. We test for both consequences in U.S. hospitals between 1999 and 2009 and in empirical models that also allow us to examine support for classic monopsony power. We find no evidence that hospitals wield monopsony power to reduce nurses wages or employment during our study period, nor do we find that existing HWI adjustment methods lead to higher wages. However, we do find evidence that hospitals responded to the implementation of the OMA by hiring fewer highskilled nurses, which implies that hospitals were gaming Medicare wage adjustment rules prior to 2005.

In 1942, the United States incarcerated all Japanese Americans on the West Coast, including children, in internment camps. Using non-West Coast Japanese Americans and non-Japanese Asians as control groups, I estimate the effect of attending a War Relocation Authority school on educational attainment. Non-linear difference-in-differences estimates suggest that attending school within the internment camps decreased the probability of receiving a post-collegiate education by approximately 4 to 5 percentage points and decreased the probability of receiving a college degree by between 2 and 7 percentage points. I find some evidence that attending a WRA school may have decreased the returns to education as well. By using un-incarcerated birth cohorts and races, placebo tests find no evidence that the identifying assumptions are violated.

Purpose: This paper aims to marry Michael Porter’s industrial cluster theory of traded and local clusters to Richard Florida’s occupational approach of creative and routine workers to gain a better understanding of the process of economic development. Design/methodology/approach: Combining these two approaches, four major industrial-occupational categories are identified. The shares of US employment in each – creative-in-traded, creative-in-local, routine-in-traded and routine-in-local – are calculated, and a correlation analysis is used to examine the relationship of each to regional economic development indicators. Findings: Economic growth and development is positively related to employment in the creative-in-traded category. While metros with a higher share of creative-in-traded employment enjoy higher wages and incomes overall, these benefits are not experienced by all worker categories. The share of creative-in-traded employment is also positively and significantly associated with higher inequality. After accounting for higher median housing costs, routine workers in both traded and local industries are found to be relatively worse off in metros with high shares of creative-in-traded employment, on average. Social implications: This work points to the imperative for the US Government and industry to upgrade routine jobs, which make up the majority of all employment, by increasing the creative content of this work. Originality/value: The research is among the first to systematically marry the industry and occupational approaches to clusters and economic development.

Job loss in the years before retirement has been found to increase risk of cardiovascular disease (CVD), but some studies suggest that CVD mortality among older workers declines during recessions. We hypothesized that recessionary labor market conditions were associated with reduced CVD risk among persons who did not experience job loss and increased CVD risk among persons who lost their jobs. In our analyses, we used longitudinal, nationally representative data from Americans 50 years of age or older who were enrolled in the Health and Retirement Study and surveyed every 2 years from 1992 to 2010 about their employment status and whether they had experienced a stroke or myocardial infarction. To measure local labor market conditions, Health and Retirement Study data were linked to county unemployment rates. Among workers who experienced job loss, recessionary labor market conditions at the time of job loss were associated with a significantly higher CVD risk (hazard ratio = 2.54, 95% confidence interval: 1.39, 4.65). In contrast, among workers who did not experience job loss, recessionary labor market conditions were associated with a lower CVD risk (hazard ratio = 0.50, 95% confidence interval: 0.31, 0.78). These results suggest that recessions might be protective in the absence of job loss but hazardous in the presence of job loss.

The objective of the AIDA project is the development of a dynamic and adaptive digital atlas on the history of Europe as well as the Mediterranean region for research and education purposes. Dynamic maps enable the visualization of temporal and spatial localization as well as their changes concerning objects and events and mediate historical processes over a long ranged period. Through the inclusion of text and image the map as a system of symbols and rules becomes a key medium for the collection of multimedia data. The conjunction of space and time and the free combination and connection of very different content opens new questions and enlarges the data basis of the system embedding user driven projects. This paper focuses on the architecture of the AIDA – project as well as the Meme Media – technology “Webbles” used for its implementation. The visualization and analysis of spatiotemporal information will be thematised as well as didactic surplus of the system concluding with the presentation of three pilot projects.

Although negotiated prices in consumer markets are rare in North America today, two important exceptionshousing and automobilesmake up the biggest purchases in most consumers lives. Negotiated prices in these markets create the possibility of large disparities in outcomes among consumers and we establish that, in the new car market, there is huge variation in final transaction prices. We show that average differences between demographic groups explains at least 20% of this variation. Our results suggest that the complex nature of vehicle transactions leads to price dispersion in this market. Older consumers perform worse in new car negotiations, even after controlling for all observable aspects of the transaction, and the age premium is greater for women than for men. We then show that the worst performing groups are also the least likely to participate in this market.

In November 2014, Barack Obama announced several executive actions to make progress in mending our broken immigration system. One part of his effort centered on broadening the requirements to be part of the Deferred Action for Childhood Arrivals (DACA) program that was announced in 2012 and has provided some relief to so-called DREAMERs brought to this country at an early age; our estimates, similar to those of the Migration Policy Institute (MPI), suggest that nearly 290,000 additional immigrants might benefit from this program expansion.

Efforts to accurately predict health outcomes with a focus on informing policy makers of where to best spend limited resources have been made in the past. This paper builds on the efforts of those studies in an attempt to build an accurate predictor of health from readily available data. The American Time Use Survey (2010, 2012, and 2013) provides the majority of the data from which this model is built, and it is then tested via several methods.

This article describes a new method for inferring the gender of personal names using large historical datasets. In contrast to existing methods of gender prediction that treat names as if they are timelessly associated with one gender, this method uses a historical approach that takes into account how naming practices change over time. It uses historical data to measure the likelihood that a name was associated with a particular gender based on the time or place under study. This approach generates more accurate results for sources that encompass changing periods of time, providing digital humanities scholars with a tool to estimate the gender of names across large textual collections. The article first describes the methodology as implemented in the gender package for the R programming language. It goes on to apply the method to a case study in which we examine gender and gatekeeping in the American historical profession over the past half-century. The gender package illustrates the importance of incorporating historical approaches into computer science and related fields. Please see the lmullen/gender-article GitHub repository for the code used to create this article.

Arguments about national tax policy have taken center stage in U.S. politics in recent times, creating acute dilemmas for Democrats. With Republicans locked into antitax agendas for some time, Democrats have recently begun to push back, arguing for maintaining or even increasing taxes on the very wealthy in the name of deficit reduction and the need to sustain funding for public programs. But the Democratic Party as a whole has not been able to find a consistent voice on tax issues. It experienced key defections when large, upward-tilting tax cuts were enacted under President George W. Bush, and the Democratic Party could not control the agenda on debates over continuing those tax cuts even when it enjoyed unified control in Washington, DC, in 2009 and 2010. To explain these cleavages among Democrats, we examine growing pressures from small business owners, a key antitax constituency. We show that organizations claiming to speak for small business have become more active in tax politics in recent decades, and we track the ways in which constituency pressures have been enhanced by feedbacks from federal tax rules that encourage individuals to pass high incomes through legal preferences for the self-employed. Comparing debates over the inception and renewal of the Bush tax cuts, we show how small business organizations and constituencies have divided Democrats on tax issues. Our findings pinpoint the mechanisms that have propelled tax resistance in contemporary U.S. politics, and our analysis contributes to theoretical understandings of the ways in which political parties are influenced by policy feedbacks and by coalitions of policy-driven organized economic interests.

In this thesis we describe a number of data visualization and optimization techniques for urban planning. Many of these techniques originate from contributions to the Social Computing Group's "You Are Here" project, which publishes maps intended to be viewed as a blend between art and urban planning tools. In particular these maps and this thesis focus on the topics of education and transportation. Eventually we hope to evolve these maps into social technologies that make it easier for communities to create the change they seek.

Population growth will result in a significant anthropogenic environmental change worldwide through increases in developed land (DL) consumption. DL consumption is an important environmental and socioeconomic process affecting humans and ecosystems. Attention has been given to DL modeling inside highly populated cities. However, modeling DL consumption should expand to non-metropolitan areas where arguably the environmental consequences are more significant. Here, we study all counties within the conterminous U.S. and based on satellite-derived product (National Land Cover Dataset 2001) we calculate the associated DL for each county. By using county population data from the 2000 census we present a comparative study on DL consumption and we propose a model linking population with expected DL consumption. Results indicate distinct geographic patterns of comparatively low and high consuming counties moving from east to west. We also demonstrate that the relationship of DL consumption with population is mostly linear, altering the notion that expected population growth will have lower DL consumption if added in counties with larger population. Added DL consumption is independent of a countys starting population and only dependent on whether the county belongs to a Metropolitan Statistical Area (MSA). In the overlapping MSA and non-MSA population range there is also a constant DL efficiency gain of approximately 20km2 for a given population for MSA counties which suggests that transitioning from rural to urban counties has significantly higher benefits in lower populations. In addition, we analyze the socioeconomic composition of counties with extremely high or low DL consumption. High DL consumption counties have statistically lower Black/African American population, higher poverty rate and lower income per capita than average in both NMSA and MSA counties. Our analysis offers a baseline to investigate further land consumption strategies in anticipation of growing population pressures.

We examine the interconnectedness of demand for regionally and locally varying public goods using a residential sorting model. We propose a version of the model that describes household choices at the city (MSA) level and, conditional on city, the neighborhood (census tract) level. We use a two-stage budgeting argument to develop an empirically feasible sorting model that allows us to estimate preferences for regionally varying air quality while accounting for sorting at the local level. Our conceptual and empirical approach nests previous sorting models as special cases, allowing us to assess the importance of accounting for multiple spatial scales in our predictions for the cost of air pollution. Furthermore our preferred specification connects the city and neighborhood sorting margins to the upper and lower elements of a nested logit model, thereby establishing a useful correspondence between two stage budgeting and nested logit estimation. Empirically we find that estimates from a conventional model of sorting across MSAs imply a smaller marginal willingness to pay for air quality than estimates from our proposed model. We discuss how the difference is attributable in part to the omitted variable problems arising when tract level sorting is ignored.

I test whether sample-selection bias explains the industrialization puzzle. It is widely believed that this decline in average stature and stature-implied standards of living in the presence of rising output and real wages was caused by the deleterious health effects of industrialization, such as increased urbanization and changes in the availability and relative price of food; but it has recently been suggested that the puzzle may be an artifact of sample-selection bias, stemming from a reliance on records of military volunteers and other self-selected samples for stature data. In this paper I test this argument by using a semiparametric sample-selection model to estimate selection-corrected trends in average stature in the United States. This estimation is based on a novel data set of my construction, consisting of military data including staturefor the birth cohorts of 18471860, linked to census data. These data are combined with similar data from the Union Army project for the birth cohorts of 18321846. Identification is aided through incorporation of voting data from the presidential elections of 1856 and 1860, which are argued to measure political motives for military enlistment. I find that the industrialization puzzleat least in the United Statesis robust to these corrections, and therefore is not an artifact of sampleselection bias. A decrease in average stature of approximately one inch in the 1830s and 1840s is present despite the correction. These results, however, do not imply that sample-selection bias is unimportant in understanding trends in average stature. On the contrary, the degree of sample-selection bias is shown to vary over time, and accounting for sample selection meaningfully alters the trend in average stature. This result shows generally that accounting for sample-selection bias can affect conclusions drawn from stature data and thus the importance of taking it into account in studying samples of stature that can be reasonably believed to not be representative of the population of interest. It also shows specifically that the decline in stature during industrialization may have been less severe than indicated by the raw data.

For the past several decades, politicians and economists thought that high levels of inequality were good for the economy. But because America’s middle class is now so weak, the US economy suffers from the kinds of problems that plague less-developed countries. As Hollowed Out explains, to have strong, sustainable growth, the economy needs to work for everyone and expand from the middle out. This new thinking has the potential to supplant trickle-down economics—the theory that was so wrong about inequality and our economy—and shape economic policymaking for generations.

This article investigates the boom and bust in U.S. homeownership rates over the 2000–2010 period. Using individual‐level census data, we first estimate 204 homeownership regressions stratified by household age (21, 22, …, 89) and survey year (2000, 2005 and 2009). Shift‐share methods confirm that changes in the model coefficients that reflect household attitudes, lending standards and other market conditions—but not population socioeconomics—were the primary driver of the boom and bust in homeownership over the decade. This pattern holds for nearly all age groups and is more pronounced for recent movers. Results also suggest that homeownership rates may have come close to bottoming out in early 2013 at 65% after falling roughly four percentage points from their peak in 2006. This suggests little lasting effect of the grand homeownership policy experiment of recent decades.

Objectives The drip and ship model is a method used to deliver thrombolysis to acute stroke patients in facilities lacking onsite neurology coverage. We sought to determine whether our drip and ship population differs from patients treated directly at our stroke center (direct presenters). Methods We retrospectively reviewed consecutive patients who received thrombolysis at an outside facility with subsequent transfer to our center between 2009 and 2011. Patients received thrombolysis after telephone consultation with a stroke specialist. We examined demographics, vascular risk factors, laboratory values, and stroke severity in drip and ship patients compared with direct presenters. Results Ninety-six patients were identified who received thrombolysis by drip and ship compared with 212 direct presenters. The two groups did not differ with respect to sex, ethnicity, vascular risk factors, or admission glucose. The odds ratio (OR) of arriving at our hospital as a drip and ship for someone 80 years or older was 0.31 (95% confidence interval [CI] 0.15–0.61, P < 0.001). Only 21% of drip and ship patients were black versus 38% of direct presenters (OR 0.434, 95% CI 0.25–0.76, P = 0.004). Even after stratifying by age (<80 vs ≥80), a smaller proportion of drip and ship patients were black (OR 0.44, 95% CI 0.24–0.81, P = 0.008). Furthermore, we found that fewer black patients with severe strokes arrived by drip and ship (OR 0.33, 95% CI 0.11–0.98, P = 0.0028). Conclusions Our study showed that a smaller proportion of blacks and older adults arrived at our center by the drip and ship model. This may reflect differences in how patients are selected for thrombolysis and transfer to a higher level of care.

This paper assesses the extent to which Rhode Islands workforce lacks skills that are in demand among the states current and potential employers and, if so, whether such a skills gap or labor market mismatch significantly restrains employment growth in the state. Using an index developed by Sahin et al. (2014), we find that occupational mismatch in Rhode Island increased leading up to and during the Great Recession and fell back to pre-recession levels by mid-2013. In 2015:Q2, occupational mismatch restrained quarterly employment growth in Rhode Island by an estimated 0.03 percentage points. However, since 2013 Rhode Islands employment growth rate has exceeded its long-run average, suggesting that the current extent of mismatch in the state is no greater than its long-run average level. Nonetheless, the share of Rhode Island-based jobs filled by out-of-state workers has increased significantly since 2001, a trend driven by jobs that employ college-educated workers. This evidence agrees with some Rhode Island employers claims that they have trouble filling skilled jobs with Rhode Island resident workers. While this finding suggests that at the broader regional level, occupational mismatch is not a serious problem, regional welfare might yet be improved by training Rhode Island workers to fill such positions.