Dickens Reconsidered or Data Under Siege? 1

Data! Data?

As several Bulletin of Science, Technology & Society (BSTS) readers know, I can be something of a “Big Data” person, working most extensively with the National Science Foundation’s Surveys of Public Understanding of Science and Technology. This database is far from the only informative one for those with STS interests. There is an increasing accumulation of databases in all disciplines (see Sabina Leonelli’s, 2013, insightful article in last year’s BSTS). Below I describe other data sources, some criticisms, and then, more critically, threats that I see to U.S. data collection and data usage.

We now have potentially magnificent data available to us, with a slew of acronyms: PISA, TIMSS, NLS (at many levels), NSF and more in archives such as the University of Michigan’s Inter-University Consortium for Political and Social Research (ICPSR), the University of Connecticut’s Roper Center, or the University of North Carolina’s Odum Center. As never before, these annals allow us to compare cultures or subpopulations and to assess temporal change and stability. They enable us to begin to separate aging from cohort effects. This wealth of materials makes it possible to build multi-item indices (which we hope are more reliable or valid—or both—than earlier single-item versions) and to invoke a heavenly host of control variables.

Some of us employ these databases as a teaching tool, in statistics, methods, and substantive topic courses. Others use them as instructional illustrations. We reference them in policy statements and as consultants to academia, government, and industry.

Researchers can also create multiple culture, age, or year “big data” files. Many commonly used statistical packages provide resources to create and/or merge files. “Megafiles” make the kind of analyses mentioned above surprisingly easy and blindingly fast. More than that, they make analyses possible that even by hand were difficult or even unimaginable 30 years ago. Even if we are more interested in the social, behavioral, and economic sciences, these kinds of databases are proliferating in all disciplines.

Better yet, for many of us, the cost is low or even free (except for the U.S. tax dollars paid to collect, create, and archive the databases). One of my statistics students downloaded all the General Social Survey (starting in 1972) “free” in less than 15 minutes on broadband using a university computer. If your institution subscribes to Michigan’s ICPSR or to Connecticut’s Roper Center, you have access to several thousand data sets. Others can be downloaded from the U.S. National Center for Educational Statistics or the Centers for Disease Control and Prevention (CDC). Even when there is a charge for the data, it is surprisingly modest: for several hundred dollars, one can acquire a data set that took millions to acquire, merge, and/or create.

So these data archives have imminent possibilities to tell us where we have been, where we are now, and where we are headed. All this sounds like the “best of times.” What’s the downside? What are the “worst of times”? First, criticisms of “big data” are plentiful (e.g., Langer, 2014, describes several from multiple critics). “Big” does not necessarily equal quality. Instrumentation can be imprecise; samples, although large, may be sloppy; and interpretation of the results can be overly optimistic. In some cases, estimation can be erratic. ²

Second, recent years have seen coalitions of ideologists, politicians, and governmental personnel, who appear motivated to restrict these potentially fruitful archives and/or halt the further collection of data to expand them. For example, I’m still smarting over losing an “affordable” Statistical Abstract of the United States. Federal funding to compile and support publication of the Abstracts ended in 2011; 2012 was the last edition published at truly nominal cost to consumers (less than $50). Furthermore, its tables and prose could be selectively accessed and downloaded from the Internet for free.

Long ago, I lost track on just how much I used Statistical Abstracts. I cited its tables for classes as diverse as Lifespan Development (births, deaths, marriages, divorces) and Research Methods (rates, ratios, data collection methods, and more). Data on U.S. household income brought my gender students down to earth on earnings they could expect as college-educated adults—as well as estimates of the cost of children. I created new data files from successive editions to illustrate time series analysis and provide longitudinal exercises for statistics students. The new 2014 edition, through private industry, is $170.00.

Political opposition to “Big Data” appears grounded in a mix of funding and ideology. Statistical Abstracts lost funding because the U.S. government decided to “save money,” although the program, relatively speaking, of course, cost little. Large databases, such as those funded by the National Science Foundation (NSF) or the National Center for Educational Statistics (NCES), cost millions. Lacking a strong rationale—from us perhaps?—governments can be stymied to defend collecting and compiling such data.

Ideologically, groups exist that see educational, social, economic, and behavioral research as politically and socially threatening. Such research may be seen as “contradicting nature” in areas such as gender or ethnic equality or challenging “common sense,” such as attitudes toward the U.S. federal Congress. In “kill the messenger” scenarios, some politicians designate data, such as the “crisis of confidence” series on falling support for many U.S. institutions, as virtually unpatriotic. As educational, social, economic, and behavioral researchers in many countries, we must educate politicians about the importance of these data initiatives and resist the state and federal legislators who appease their constituencies that complain about funding, collecting, analyzing, and disseminating these data by eliminating these databases.

Last year I reported on the federal “Continuing Appropriations Act of 2013,” which limits funding “political science research” by the NSF unless a project is certified as “promoting national security or the economic interests of the United States.” So far, it’s unclear exactly what constitutes “political science research,” although it seems to encompass attitudes toward American congressional personnel.

Thus far this year, the state of Texas wants to make participation in the Census’ American Community Survey voluntary. Government, researchers, and industry extensively use this mandatory data-collection tool in the aggregate, among other things, for economic forecasting.

Then, this March, two American federal representatives introduced H.R. 4186, the Frontiers in Innovation, Research, Science and Technology Act of 2014 (“FIRST Act”), to reauthorize programs at the NSF. As the American Educational Research Association (AERA) noted in an email,

Previous iterations of this bill, the America COMPETES Act of 2007 and 2010, provided an important foundation for fostering innovation in STEM education and research, but the new bill creates a low ceiling for NSF funding and specifically targets the SBE Directorate (Social, Behavioral and Economic Sciences) for significant cuts.

The “low ceiling” term is an oxymoron: the bill would cut funding in the SBE division for the next two fiscal years by nearly half. It would also appear to partially substitute political judgment for peer review in all NSF areas.

It is no accident that organizations such as the American Association for the Advancement of Science immediately opposed such legislation. Are we are so naïve as to believe that life or physical science funding, or the CDC will be immune to political chicanery? After all, we know that science education, physical anthropology, biology, meteorology, chemistry, and geology have all provided fodder for recent political activism, for example, on evolution and/or climate change.

For an illustrative lesson on government suppression—or even benign neglect—of science and technology look no further than Alexander Terekhov’s (2013) article in the last BSTS issue. He describes how Russia’s neglect and low funding of nanotechnology during and immediately following Perestroika severely damaged research and industry. A generation of “nanotech” scholars either were never created in the first place or decamped from Russia for more encouraging and financially generous countries.

We can support organizations such as the Consortium of Social Science Association, AERA, or AAAS, among others, who have spoken to support all the sciences. They remind us to remain vigilant and not to take these data resources for granted. Scholars from outside the United States are welcome to submit informative articles, summaries, and letters to the editor describing the status of science and technology in their home countries.

In This Issue

In a continuation of a dialogue begun in our prior issue (e.g., Andersson, 2013), what kind of knowledge does the general public need to contribute to public science discourse and policy? To set the stage, Susanna Priest reviews the state of “critical science literacy.” Next are three articles (by Eric Plutzer; John Besley; and Aviram Sharma and Saradindu Bhaduri) that address the public and science, including reactions when technology goes awry, and trust in the processes of science and technology. Continuing with a BSTS long-standing interest in technology, Sora Park examines what happens over time when students receive a gift: their very own mobile tablet! BSTS receives (and publishes) a lot of international submissions; science and technology “work differently” in postindustrial, industrial, and industrializing countries (see also Sharma and Bhaduri in this issue). Aditya Poudyal and Jukka V. Paatero examine rural energy planning from a social and economical perspective.

Thanks To

In addition to the BSTS Editorial Board, these guest reviewers gave generously of their time and thought for this issue:

Look for more book reviews in the next BSTS issue!