Hi All --
Our speaker this Wednesday at Applied Stats will be Krista Gile who is a Professor in the Department of Mathematics and Statistics at the University of Massachusetts. She will be giving a talk entitled New methods for inference from Respondent-Driven Sampling Data. The abstract is included below.
Professor Gile's research focuses on developing statistical methodology for social and behavioral science research, particularly related to making inference from partially-observed social network structures. Most of her current work is focused on understanding the strengths and limitations of data sampled with link-tracing designs such as snowball sampling, contact tracing, and respondent-driven sampling.
As per usual, the talk will be held at 12 noon in CGIS K354. Lunch will be served.
I hope to see you all there!
Tess
---------
Title:
New methods for inference from Respondent-Driven Sampling Data
Abstract:
Respondent-Driven Sampling is type of link-tracing network sampling used to study hard-to-reach populations. Beginning with a convenience sample, each person sampled is given 2-3 uniquely identified coupons to distribute to other members of the target population, making them eligible for enrollment in the study. This is effective at collecting large diverse samples from many populations.
Current estimation relies on sampling weights estimated by treating the sampling process as a random walk on the underlying network of social relations. These estimates are based on strong assumptions allowing the data to be treated as a probability sample. In particular, existing estimators assume a with-replacement sample with an ideal initial sample. We introduce two new estimators, the first based on a without-replacement approximation to the sampling process, and the second based on fitting a social network model (ERGM), and demonstrate their ability to adjust for biases due to the finite population and initial convenience sample. Our estimators are based on a model-assisted design-based approach, using standard errors based on a parametric bootstrap. We conclude with an application to data collected among injecting drug users, including extension to observable features of the sampling process.
Hello Everyone!
I hope you all had a lovely spring break! Our speaker at Applied Stats this week will be Herman van Dijk, a Bayesian econometrician and visiting fellow at IQSS, who will giving a presentation entitled "Forecasting with Many Models in Finance and Economics using Large Data Sets and Parallel Computing."
As per usual, we will meet in CGIS K354 at 12 noon on Wednesday and lunch will be served. Here is the abstract for the talk:
We propose a Bayesian combination approach for multivariate predictive densities which relies upon a distributional state space representation of the combination weights. Several specifications of multivariate time-varying weights are introduced with a particular focus on weight dynamics driven by the past performance of the predictive densities and the use of learning mechanisms. In the proposed approach the model set can be incomplete, meaning that all models can be individually misspecified. A Sequential Monte Carlo method is proposed to approximate the filtering and predictive densities. The combination approach is assessed using statistical and utility-based performance measures for evaluating density forecasts of simulated data, US macroeconomic time series and surveys of stock market prices. Simulation results indicate that, for a set of linear autoregressive models, the combination strategy is successful in selecting, with probability close to one, the true model when the model set is complete and it is able to detect parameter instability when the model set includes the true model that has generated subsamples of data. Also, substantial uncertainty appears in the weights when predictors are similar; residual uncertainty reduces when the model set is complete; and learning reduces this uncertainty. For the macro series we find that incompleteness of the models is relatively large in the 1970’s, the beginning of the 1980’s and during the recent financial crisis, and lower during the Great Moderation; the predicted probabilities of recession accurately compare with the NBER business cycle dating; model weights have substantial uncertainty attached. With respect to returns of the S&P 500 series, we find that an investment strategy using a combination of predictions from professional forecasters and from a white noise model puts more weight on the white noise model in the beginning of the 1990’s and switches to giving more weight to the professional forecasts over time. Information on the complete predictive distribution and not just on some moments turns out to be very important, above all during turbulent times such as the recent financial crisis. More generally, the proposed distributional state space representation offers great flexibility in combining densities.
The corresponding paper is attached. In addition there is another paper that provides some background for those who are interested.
See you all on Wednesday!
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com
I forgot to include the abstract in the last email!
Ecometrics in the Age of Big Data: Measuring and Assessing Neighborhood Characteristics Using Administrative Records
The collection of large-scale administrative records in electronic form by many cities provides a new opportunity for the measurement and longitudinal tracking of neighborhood characteristics, but one that will require novel methodologies that convert such data into research-relevant measures. The current paper illustrates these challenges by developing measures of physical disorder from Boston’s “Constituent Relationship Management” (CRM) system. A sixteen-month archive of the CRM database contains more than 300,000 address-based requests for city services, many of which reference physical incivilities (e.g., graffiti removal). The work seeks to solve three challenges presented by the raw database: 1) identifyingcontent pertinent to the measure of interest; 2) assessing the validity of the data using objective audits; and 3) establishing reliability criteria for. This generated a multi-dimensional measure of physical disorder that could be measured repeatedly for virtually no cost every 2-6 months, representing an important new resource in research on urban disorder. The process also generated some additional ecometrics regarding civic engagement and care for the public space. Ways to extend this methodology to new data sets, locales, and research questions are discussed.
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com
Hi Everyone!
Our speaker at Applied Stats tomorrow (Wednesday, March 12) will be Daniel O'Brien (Harvard Sociology). Daniel's talk is entitled "Ecometrics in the Age of Big Data: Measuring and Assessing Neighborhood Characteristics Using Administrative Records." The abstract is included below.
As per usual, we will meet in CGIS K354 at 12 noon and lunch will be served.
See you all there!
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com
Hi Everyone!
I am incredibly excited to announce this week's speaker: Professor Ned Hall from the Harvard Department of Philosophy. The working title of Professor Hall's talk is "In Praise of Causal Mechanisms." As per usual, the talk will be held in CGIS K354 on Wednesday (3/5) at 12 noon and lunch will be served.
You might be asking yourself: "Why has Tess invited a philosopher to Applied Stats?" Well, as Kurt Lewin, (sometimes considered the father of modern social psychology) once said: "There is nothing so practical as a good theory." When I need some practical theory about causality, I turn to Professor Hall's work -- especially Structural equations and causation<http://link.springer.com.ezp-prod1.hul.harvard.edu/article/10.1007/s11098-0…>, and his recently released edited volume, Causation and Counterfactuals<http://mitpress.mit.edu/books/causation-and-counterfactuals>.
Here's Professor Hall's abstract:
Consider two theses about causation: (1) Causes are connected to their effects by way of mediating causal mechanisms or processes. (2) Scientific inquiry aims (at least in part) at discerning and describing the causal structure of our world. Some of the best contemporary work on causation claims—often implicitly, but sometimes quite explicitly—that, in giving an account of causation, we should sacrifice (1) for the sake of producing an account that makes the best sense of (2). I will first try to show why this claim is quite attractive, and then obstreperously argue against it: I will aim to show that the best way to make sense of (2) is, in fact, by means of an account of causal structure that fully vindicates (1).
Looking forward to seeing you all on Wednesday,
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com