Discover and read the best of Twitter Threads about #JSM2018
Most recents (24)
#JSM2018 Tobias Schmidt Looking at interviewer experience and interview duration 
#JSM2018 Schmidt In this survey, duration linked to interviewer salaries.
#JSM2018 Schmidt Looking at interviewer experience over the course of survey and respondent experience within survey and experience over repeated surveys. Looking in particular at experience within panel survey for both Iers and Rs
#JSM2018 Celine Wuyts from KU Leuven up next.
#JSM2018 Wuyts Interested in within-survey workload. Use call history data and interview time data. Some Measure workload by fixed measures of experience and interview order cumulated over the field period. They use actual number of cases assigned at time t in field period
#JSM2018 Wuyts Use Paradata to create new measures of interview workload, based on sample units assigned on given day
#JSM2018 Rebecca Powell from @RTI_Intl talking about an experiment on Add Health shifting from interviewer administered to self administered survey
#JSM2018 Powell moved to a 55 self-administered survey from 90 minutes interviewer administered. Worried about response burden with this length of self-admin survey. Randomized n=7600 into either full 55 minute survey or 2 modules- one 35 minutes then 20 minutes.
#JSM2018 Powell Could select to continue on the web. In paper, had to first complete module A, then sent module B. Cover letters told about modules in the incentive part, but not up front. $55 incentive total in each condition
#JSM2018 The brilliant Susan Murphy is this year’s Fisher Lecture award recipient!
#JSM2018 Murphy Lab does sequential experimentation in improving health. Some for companies.
#JSM2018 Murphy Experimentation and continual optimization is key. How do we use learning as an experiment is put into the field to improve outcomes for individuals? Mobile interventions are key here. Intervention may be either a push intervention or pull intervention
#JSM2018 Next up Hubert Hamer from NASS talking about NASS Small Area Estimation
#JSM2018 Hamer NASS has Agriculture Loss Coverage County Option program. Payments triggered based on county crop revenue falling below program guarantee. NASS surveys used to make this decision, along with other data
#JSM2018 Hamer Program paid out $3.7 billion on 2016. Small changes can affect payments
#JSM2018 Peter Miller appearing as a Northwestern University emeritus professor, providing comments on the CNSTAT reports
#JSM2018 Miller Survey paradigm vs multiple data source paradigm. Surveys may become irrelevant b/c they are slow, not granular, not nimble, costly, not sustainable
#JSM2018 Miller Multiple Data sources require new: methods, computing resources, privacy protections, training, data quality frameworks. Not cheap. What does this give us?
#JSM2018 Panel on CNSTAT report on Federal Statistics, Multiple Data Sources, and Privacy Protection, with @fraukolos kicking off The discussion
#JSM2018 @fraukolos Goal of panel to evaluate combining data sources to possibly replace / augment surveys. Two reports out of the panel.
#JSM2018 @fraukolos Conclusions: Current Federal Statistical Agencies face threats from falling response rates, rising costs, increased desire for granularity and timelines
#JSM2018 Waiting for John Eltinge from @uscensusbureau to deliver the Deming Lecture: “Improving the Quality and Value of Statistical Information: 14 Questions on Management”
#JSM2018 Who is John Eltinge?
#JSM2018 Eltinge Deming and his work shaped by his time-1900 to 1993. Big changes in industry, government, how citizens related to science.
.@AllenDowney on "Inference in 3 hours" - slides at tinyurl.com/inference3 and a relevant blog post allendowney.blogspot.com/2016/06/there-… #JSM2018
.@AllenDowney: "You can get through most important ideas of statistical inference in 3 hours, especially if you avoid black holes." #JSM2018
The infer package draws inspiration from @AllenDowney's post too! infer.netlify.com
#JSM2018 A highly distinguished all-women panel for the 1 year update on the Commission on Evidence-Based Policy Making. Katharine Abraham, Nancy Potok, Amy O’Hara, and Julia Lane
#JSM2018 Abraham Commission co-sponsored by Paul Ryan & Patty Murray. 2/3 appointed based on substantive issues re: data, 1/3 appointed based on expertise in privacy.
#JSM2018 Abraham Had one year to do report - 7 public meetings, public hearings held across US, put out request for comments, talked to federal offices. 8 months of data gathering, 4 months of writing. Wrote a report that had unanimous support from all commissioners.
#JSM2018 Don Rubin as discussant up next.
#JSM2018 Don Rubin as Yoda, care of Jorg Dreschler
#JSM2018 Don Rubin talking about developing multiple Imputation on organization and industry codes at Census. Codes written on 3x5 cards, not computer readable cards, so hard to do anything with it
#JSM2018 Rod Little up next, talking about multiple Imputation for causal inference
#JSM2018 Little Statistics is prediction. Multiple Imputation is an all-purpose tool for prediction.
#JSM2018 Little Using Penalized Spline of Propensity Prediction (PSPP). Estimate propensity to respond to Y model, then include propensity as spline in impute for Y with other covariates. Now, extend this to causal inference
#JSM2018 Raghu Talking about using Multiple Imputation for sample design. Cites @jameswagner254 2010 Stats in Medicine paper.
#JSM2018 Raghu Have benchmark data from large national survey or admin data. Have population frame with sample units. Draw replicate sample from frame such that distribution is similar to population
#JSM2018 Standing room only for this session. With Jorg Dreschler, Raghu, Rod Little, and Don Rubin as presenters, easy to understand why
#JSM2018 Jorg Dreschler up first. Classification systems for industries change frequently, no consistent coding over time. Treat problem as missing data problem.
#JSM2018 Dreschler Has Establishment History Panel from all establishments in Germany with 1+ employee covered by social security. Lots of data on each establishment. 4 different classification systems in use. Two systems are jointly observed in one year.
#JSM2018 Lepkowski reminds us to be careful in looking at average relative bias with proportions as the denominator can really affect conclusions (do we get same conclusions if we look at q rather than p?).
#JSM2018 Lepkowski wonders what reputation means in context of surveys. If more information is out there about our survey, telling potential Rs that our “cause is good” will they do our survey? Depends on how we tell them about it. No acronyms!
#JSM2018 Tuba Suzer-Gurtekin from @umisr next speaker. Are there time-dependent differences in the Survey of Consumers, breaking the time series? Need to care about this when transitioning to new modes
#JSM2018 Suzer-Gurtekin Verbal likelihood on better off/ worse off than a year ago; other response scales in survey. Verbal likelihood shown with 3 points in mail, but only two points read on phone, with “same” being volunteered on phone.
#JSM2018 Suzer-Gurtekin Differences between modes could be due to q’n format, visual vs auditory differences in primacy and recency effects, and probing differences
#jsm2018 Paul Schulz from @umisr up next, talking about mixed mode surveys on the Index of Consumer Sentiment from the Surveys of Consumers
#JSM2018 Schulz Parallel modes for SoC - current mode is cell only; parallel study is ABS sample with mode choice for mail or web
#JSM2018 Schulz Web/mail is older, more female, lower income than phone. Can weight these differences away. But other differences may exist. Phone is considered benchmark in this study, so want to have web/mail match the phone.
#JSM2018 Up next is Don Dillman. Challenges in web-push studies of the general public.
#JSM2018 Dillman inspired by JSM2007 presentations on using DSF ABS surveys for in person surveys, so why not mail? Mail only surveys from 2007 to 2014 for higher response rates than web-push. But things are changing, and web push is out there in lots of important surveys
#JSM2018 Dillman Telephone communication is changing, especially with local presence dialing and number spoofing. Voice interviews being replaced with visual communication. Mixed mode is effective with multiple contact modes.
#JSM2018 Joe Sakshaug up first in the session on mixed mode surveys
#JSM2018 Sakshaug talking about sequential mixed mode surveys and nonresponse and measurement error bias. Isolating these errors in hard. Had CATI and mail as starting modes, and then switched modes, to look at ordering of modes and NR/ME errors
#JSM2018 Sakshaug looking at 2006 study on driving experiences of young adults; random assignment to mail or CATI starting mode in wave3 of longitudinal study
#JSM2018 Last speaker Rita Ko from The Hive - Humanizing Data about refugees, use same type of models for presidential elections to predict supporters of refugee crisis
#JSM2018 Ko How to get people to support a cause if they haven’t thought about it previously?
#JSM2018 Ko Building predictive model to identify potential supporters. Don’t go for easiest folks who look just like current supporters
#JSM2018 Mark Hansen up next. Best title slide so far
#JSM2018 Hansen Data are not neutral . Data journalists are contributing to data literacy in the public
#JSM2018 Hansen teaches a computational journalism class. What makes something trend? What is behind recommender systems? Use Jupyter notebooks and teach R and Python to journalists
#JSM2018 Richard Coffin from USAFacts up next. USAFacts - mission: make government spending data useful and accessible. Create a 10-K for government.
#JSM2018 Coffin Organize all Federal, state and local revenue and spending in categories, link to outcomes
#JSM2018 Coffin Goals for their Data
#JSM2018 Next up - The Future of Statistics and the Public
David Spiegelhalter #JSM2018 BBC report that unemployment fell by 3000 people w/ a MoE of 77,000
