ePrivacy and GPDR Cookie Consent by Cookie Consent
   
AAPOR
The leading association
of public opinion and
survey research professionals
American Association for Public Opinion Research

Webinar Details

Applications of Predictive Modeling to Survey Design & Operation in Address-based Samples


Thursday, March 17, 2022
1:00 - 2:00 PM Eastern Time

ASA-SRMS members will receive AAPOR member pricing on webinars when registering for live webinars or purchasing recordings of webinars.
If you are an ASA member, click here to purchase.

Speaker: Cameron McPhee, Chief Methodologist at SSRS

adam-zammit-22.jpgAbout Cameron:
Cameron McPhee, Chief Methodologist at SSRS, brings 20+ years of experience in survey research, data collection, and quantitative analysis to her role providing methodological and statistical support throughout the organization. She directs SSRS’s advanced analytics and methodology team, delivering high-quality, statistically sound solutions developed using experimentally validated methods including rigorous, and efficient methods for sampling, measuring, and weighting diverse populations. Skilled in conceptualizing and managing survey data collection from design to dissemination, Cameron has extensive expertise in methodological experimentation, sampling, weighting, and adaptive design.

Prior to joining SSRS, Cameron spent 12 years at the American Institutes for Research directing AIR’s work supporting a wide range of federal education surveys conducted by the National Center for Education Statistics including the National Household Education Survey and multiple longitudinal cohort studies.

Cameron holds a Master’s degree in Survey Methodology from the University of Maryland and a Bachelor’s degree in Sociology from Swarthmore College.

Description:
As survey response rates continue to decline, the risk of bias grows due to the correlation between individuals’ willingness to respond to the survey request and the substantive constructs being measured. Over the last decade, survey researchers have grown their arsenal of techniques for improving survey response and representativeness including the use of carefully tailored designs (Dillman, et. al., 2014), mixed modes for recruitment and response, and adaptive and responsive designs (Groves & Heeringa, 2006;  Peychev et. al., 2009; Wagner, 2021; West et. al., 2015; ) that incorporate results from previous surveys or earlier waves to target resources and/or interventions.

The effectiveness of many of these techniques for improving response and/or representativeness of sample surveys can be enhanced through the use of propensity modeling. Broadly, propensity modeling is an empirical process that identifies a multivariate statistical model to predict the likelihood (propensity) of observing a particular characteristic or behavior in a sampled element. Propensity modeling can be applied to many aspects of the data collection process including identifying members of harder-to-reach subgroups for sampling purposes, tailoring communication and survey materials to subsets of sampled cases, and targeting higher-cost interventions to those sampled units most needed to reduce bias or improve response rates at the start of data collection or responsively in later waves (Jackson et. al., 2019; Jackson et., al., 2021; Peychev et. al., 2020).

This webinar will discuss the various ways propensity models can be used to improve the design and fielding of surveys using address-based samples (ABS). While the focus of the webinar will be on applications for ABS, much of the information can be applied to other sampling frames including RDD, RBS, and other list frames. We will give a brief, non-technical, overview of the statistical methods used to create propensity models.

However, the majority of the hour will focus on how the models can be applied to different aspects of the survey cycle and will provide examples from recent studies. Specifically, we will discuss using propensity models to oversample lower-responding subgroups or sample units meeting certain eligibility criteria, differentially allocating cash incentives, tailoring mailing materials to appropriate sample units, varying response mode based on predicted preferences, and selectively utilizing more costly nonresponse follow-up strategies based on predicted response and bias propensities. Each example will include a discussion of the tradeoffs associated with the implementation of the propensity modeling.

The session will conclude with an open discussion of the application of propensity modeling for survey research and give participants an opportunity to ask questions relevant to their own research.


Want to be added to the notification list? Email [email protected].

Note: If you purchased the yearly webinar subscription with your membership renewal, you are already registered for this webinar and do not need to re-register. You will automatically receive the login instructions before the event. Click here for the Webinar Package FAQs.