National Survey Response Rates Down – Will Big Data Fill the Gap?
|June 19, 2013||Posted by M. P. under Federal Government, News, Policy, Research||
An April 2013 briefing to Congress on surveys and statistics focused on the problematic trend of declining response rates for federal surveys, including the American Community Survey and the National Survey of Child Health. The briefing, Policy Makers & Businesses Need Reliable Information and Data: The Impact of Falling Response Rates to Social Surveys and What Can Be Done, organized by The American Academy of Political and Social Science (AAPSS), outlined the risks to research and the impact on policy-making if response rates to surveys on health, employment and household continue to subside. The largest risk is that of biased results. Other issues:
- Nonresponse rates currently range from 30 to more than 60 percent. This is an all-time high.
- Over 60 percent of nonresponses were refusals, while approximately another 1/4 were due to the inability to contact the intended recipient.
- Young single-person households, minorities, renters and the poor were less likely to respond.
- One-time surveys have higher nonresponse rates than more complex longitudinal studies that follow the same group of respondents over period of time.
While incentives (such as a gift card or a small amount of money) for completing and returning a survey would boost response rates, it would also increase costs – a risky proposition in an atmosphere of austerity. The authors of a related paper, Where Do We Go from Here? Nonresponse and Social Measurement, published in the January 2013 volume of AAPSS’s The Annals, suggest that a solution to this growing problem is a strategic outreach plan to inform both politicians and the public of the purpose of national surveys. Clear explanation of what the data is used for, as well as the regulations and protocols in place to protect it from being presented other than in aggregate form could have a favorable impact on perception. Unfortunately for these and other large-scale surveys, the recent news of metadata collected absent suspicion may have even the most tech-savvy survey-loving among us rethinking issues of privacy, transparency and information storage and retrieval.
Perhaps in the future these surveys that, by the way, inform funding decisions on infrastructure, education, and transportation to name a few, will be deemed too intrusive and/or obsolete and left behind. Funding and other governing decisions can then be made based on variables extracted from all that we have uploaded onto the digital data heap. So, will big data replace big surveys? Will traditional statistical methods be successful in tracking, analyzing and accurately reporting big data to inform policies at the federal, state and local level?