SURVEYS THAT CHANGED THE WORLD
Surveys are supposed to begin and end. A month, for instance, should be all it takes to do the actual collection of data. That is what makes the PSID survey a strange specimen. This survey has been going on for 45 years.
Its full name is the Panel Study of Income Dynamics or simply PSID. If that sounds like the sort of title that a group of government bureaucrats or academics might give to a survey, it is because it is, in fact, a child of the US government and the University of Michigan.
But why would anyone want to have a never-ending research survey?
How the Immortal PSID Survey Was Born
Like most government plans, this one was not intended to become what it became. The year was 1966, and the US government’s Office of Economic Opportunity (OEO) wanted to be able to tell the President, Lyndon Johnson, how well his War on Poverty was faring. So, the OEO asked the Bureau of the Census to do a field survey on the President’s project, which was a pet project of his.
The result was the Survey of Economic Opportunity (SEO) of 1968. A sample of over 18,000 individuals living in 5,000 families, mostly from the poorer segment of the population, were sampled from the national census list and then interviewed. They were asked about their status and type of employment, their income, wealth, expenditures, health, marriage, childbearing, child development, and their education.
That should have been the end of it. But, as it happened, the government functionaries were intrigued by the mine of data. This was the first time anyone had thought of directly checking with families on how they were doing. So, after dutifully impressing the President with the data, the OEO went looking for the Survey Research Centre (SRC) of the University of Michigan and offered them money to continue the survey.
That is how Professor James Morgan got the job that defined his life. As the study’s first director, the professor made the fateful decision of asking the government people to let him add into the survey sample some non-poor households from the university’s national population samples. The idea was to make the survey more nationally representative. The professor may also have been cleverly exploiting an opportunity to piggy ride the university’s research work on the incomparably healthier bank account of a government department.
Anyway, Professor James Morgan and his research team added 3,000 new households to the survey sample and threw in some more sophisticated survey techniques like a new questionnaire and probability-of-selection weights to each sample family.
And that is how the world’s first household standard-of-living tracking survey was born. Every year, the sampled families were asked how they were doing as compared to the last time. The same families would be visited each time. When it became clear that some families were immigrating to other parts of the country – and the kids were growing up and getting married – somebody came up with the bright idea that the survey should follow ‘its people’ wherever they went and take on board any new family that was formed from the marriage of a child from an original sample family.
Along the way, it also became clear that Latinos were not adequately represented in the survey sample, even by their 1968 ratio. So a sample of 2,043 Latino households from the Latino National Political Survey at Austin’s University of Texas was thrown into the mix.
How the PSID Survey Is Done
From that 1968’s almost accidental survey, the PSID now annually tracks over 70,000 individuals in over 10,000 families across the United States. There is a ‘core’ sample covering all the families except those added in the 1997 ‘refresher sample’ of immigrants. At every survey taking, the sample is further categorized into a ‘re-interview sample’ (consisting of families that were interviewed the previous year) and a ‘recontact sample’ (families that missed out the previous year’s survey). Then there are the ‘split-off’ families, made of families of kids who grew up and started their own families.
Response rates are calculated separately for each sample category before being summed into the overall response rate.
Among the questions asked are sources of the previous year’s income including type of job, inheritance, unemployment payments, and child support, plus major cost of living costs such as pension and rent payments. Families also get to state the time they spent at work and on housework, including the money saved on performing household repairs including car repairs.
The usual social status data is also gathered, such as age, residential location, marital and education status, and changes in family composition including deaths and marriages. This includes questions about income sources and amounts, employment, and family composition changes. Occasionally, the survey also throws in questions about family wellbeing or happiness.
Criticism and Impact of the PSID Survey
Perhaps because of the subject it covers, the Panel Study of Income Dynamics is one of the few major surveys that have not attracted serious criticism. The nearest to a major criticism is that the survey has lost approximately 50% of its original sample from the attrition of time and changing circumstances. The loss has been more common among families with more economically shaky situations, marriages, and migration histories.
But studies over the years have shown that this attrition has not seriously distorted the surveys sample.
And what about the survey’s impact? The PSID survey has achieved stupendous success. Its data cache provides a unique glimpse into the link between economic situation with health, education, and even emotional wellbeing across generations. No other types of survey can provide such clarity on changes within families over generations.
As an indicator of just how the PSID data has become central to social science research in the US, there are more than 3,200 peer-reviewed publications based on the survey. Further, nine U.S. Federal Agencies plan their work around the survey’s findings. And 5 million people visit the PSID website every year - 2,500 of which download in total about 25,000 parts of the survey.
Another indicator of the survey’s success is that its major funder - the U.S. National Science Foundation (NSF) – ranks the survey in its list of 60 finest achievements, alongside such shattering scientific milestones as the development of the Google search engine and nanotechnology. But perhaps best of all, virtually all developed countries – and an increasing number of developing countries as well – have borrowed from the PSID playbook and started their own household tracking surveys.
As surveys go, there can be little doubt that this accidental survey – initially meant to please a president but then just couldn’t end – easily ranks among the greatest surveys ever conducted.
May it never end.
Do you know of an important survey that you would like us to cover? Feel free to send us your suggestions!
Photo credit: Futurilla (thanks, Futurilla)
Fitzgerald, J., Gottschalk, P., & Moffitt, R. A. (1998) An analysis of sample attrition in panel data: The Michigan Panel Study of Income Dynamics
McGonagle, K.A., Schoeni, R.F., Sastry, N. & Freedman, V.A.. (2012) The Panel Study of Income Dynamics: overview, recent innovations, and potential for life course research, Longitudinal and Life Course Studies, 3 (2), p268 – 284
PSID (2014) PSID: A national study of socioeconomics and health over lifetimes and across generations, Available at: http://psidonline.isr.umich.edu/, Last accessed 16th Feb 2014
You might also like...
The Kinsey Reports: Surveys That Questioned America's Sexuality
The Lancet Surveys: Surveys That Increased Iraq's Death Rate By 150%
The TI Corruption Index: 184 Surveys To Measure Corruption
Feb 18, 2014