“Survey Research and Sampling in the New Digital Era – Challenges and Innovative Approaches”
Presented in Partnership with the Eagleton Institute of Politics.
Chintan Turakhia, Executive Vice President of Strategic Research and Innovations at SSRS
Mansour Fahimi, Ph.D., Executive Vice President & Chief Data Scientist at Marketing Systems Group
Wednesday April 14, 2021 from 5:00 PM to 6:00 PM Eastern (US)
Virtual Meeting via Zoom
Free for Members and Students; $20 for new members (2021 membership included)
SURVEY RESEARCH IN THE DIGITAL AGE
Presented by CHINTAN TURAKHIA
2020 can easily be classified as an inflection point in the evolution of digital age. COVID-19 and resulting technological evolution has potentially led to permanent changes in how we think and behave. Technology is driving many aspects of our lives- from purchasing behaviors to how we communicate and express our opinions. What does it all mean for survey research? In this presentation, we will attempt to examine what recent changes to the digital landscape mean for the future of survey research.
SPEAKER BIO: Chintan Turakhia has over 26 years of survey research experience and has expertise in all phases of survey research. Chintan is a nationally renowned expert on probability based on-line panels and data collection platforms. His extensive experience includes establishment of first ever face-to-face recruited internet survey platform (FRISP) in 2008-2009, assisting with the design and establishment of the nationally representative on-line American Trends Panel for Pew Research Center (pewresearch.org), and recently creating a large-scale probability-based panel for Annenberg Public Policy Center at the University of Pennsylvania. He presently leads the probability based SSRS Opinion Panel at SSRS.
SURVEY SAMPLING IN THE ACCELERATED DIGITAL AGE
Presented by DR. MANSOUR FAHIMI
Conducting credible survey research in the 21st century is an endeavor subject to growing challenges that require thinking outside of the traditional survey sampling toolbox. The statistical machinery developed by Neyman (1934) has made it possible to produce measurable inferences about target populations based on samples of modest size, however, that paradigm relies on a set of assumptions that are becoming exceedingly difficult to fulfill. Under that traditional approach, samples must be selected from complete frames; all sampling units should carry known selection probabilities; and surveys need to secure near-perfect rates of response. While most surveys conducted these days struggle to satisfy these fundamental assumptions, still, representative sampling and diligent survey administration; effective analysis of survey data; and coherent interpretation of the results are nonnegotiable principles that should guide all surveys.
It is from the above perspective that rigorous survey research programs are designed and implemented to minimize the so-called Total Survey Error. Recognizing the truism that organic representation of the target population is exclusive to probability-based samples, however, due to cost and infeasibility concerns there is a growing interest among researchers to consider alternative methods for survey design and administration. In particular, online panels are gaining popularity as an efficient platform for sample surveys. However, virtually all online panels are compiled using convenience methods of sampling, exposing the resulting opt-in panels subject to unknowable coverage problems. During this presentation we will review some of the current best practices for dealing with these challenges. Specifically, a two-pronged approach will be discussed that relies on improved sample representation and enhanced weighting adjustments developed for samples with compromised coverage properties.
SPEAKER BIO: For over 30 years, Dr. Mansour Fahimi has been providing statistical expertise and hands-on support for projects from design to delivery, investigating innovative refinements for market and survey research methods, and mentoring staff. Mansour works on THE design and administration of complex surveys for both public and private sector clients, as well as data enhancement methods, process optimization procedures, and program evaluation tasks. He has extensive experience with advanced data analysis techniques, particularly multivariate methods for analysis of weighted data from complex surveys. In recent years Dr. Fahimi has been working on online data collection methods where his research and publications have focused on improving the inferential integrity of data from compromised samples. Mansour has developed innovative calibration techniques to compensate for the limitations of data with unspecified representational properties, as well as data quality initiatives and procedures for elimination of fraudulent respondents. Dr. Fahimi has numerous publications and presentations to his credit and has served on technical advisory boards for reviewing government surveys and program evaluations.