Skip to main content

Currently Skimming:

5 Data Sources and Methods
Pages 115-136

The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.


From page 115...
... In particular, federal statistical agencies have been the predominant source for national-level collection of economic, employment, and demographic data about individuals and businesses. Additional data are increasingly being collected by other means, both in the private sector and by academic researchers.
From page 116...
... In addition, federal agencies obtain administrative data as byproducts of various programs and activities, such as household and business tax collection and related filings or other transactions or record keeping. Core administrative data are housed at the Internal Revenue Service and Social Security Administration, both of which have statistical arms that provide important measures of economic activity; much of these data are also provided to the Census Bureau.
From page 117...
... In recent years, some repositories of key administrative data have been made available to the research community on a restricted-use basis. For example, the Federal Statistical Research Data Centers, funded jointly by the Census Bureau and the National Science Foundation, allow approved research institutions to access census and NCHS data for statistical use; other agencies, such as BLS, are scheduled to join this system.
From page 118...
... generated from administrative data on all private, nonfarm establishments in the United States. CES data are generally used to evaluate monthly employment and earnings growth rates, rather than absolute levels, and are annually benchmarked (or normalized)
From page 119...
... , both collected by the Census Bureau. The much larger samples in the Decennial Census and the ACS permit cross-classifying workforce trends by demographics, 3  T
From page 120...
... data at Census Bureau tracks job creation and destruction statistics annually using administrative data on U.S. businesses in the private, nonfarm sector.
From page 121...
... For example, Abraham et al.4 use the integration of the person-level CPS and the LEHD-matched employer-employee data to study changing labor market trends. They find, for example, that many individuals in the CPS state that they are working as wage and salary workers, but the administrative data underlying the QWI and other related sources at the Census Bureau show that they are working as independent contractors.
From page 122...
... Tracking Changes in Technology Using U.S. Federal Data Sources The federal statistical agencies have a wide array of business surveys that provide information on the changing nature of technology and workplace organization at U.S.
From page 123...
... The research community has been using this as an important source for tracking technological change for years, but these data are now increasingly being integrated into the other survey and administrative data discussed above. Research projects with participants from the academic community, the patent office, and the statistical agencies are integrating patent and other related data on innovation to longitudinal business data (i.e., the microdata underlying the BDS)
From page 124...
... While O* NET is quite valuable for understanding current jobs and associated requirements, researchers face some challenges when aiming to use its content to help assess changing skills requirements and the likely impacts of technology on a given occupational field.
From page 125...
... However, such data can be costly and time-consuming to collect, especially from offline surveys. However, the digitization of economic and labor market transactions has created new opportunities for tracking and evaluating workforce trends.
From page 126...
... They are available more quickly, often in real time, and they are available in many forms and types.11 Big data are commonly described as differing from traditional data sources due to the volume, velocity, and variety of these data.12 The following types of naturally occurring data -- that is, data that would exist in digital form whether or not someone sought them -- provide key examples of analytical opportunities for researchers. • Individual worker profiles.
From page 127...
... For example, Burning Glass Technologies analyzes millions of job postings to track skills gaps in real time.13 Real Time Macroeconomics works similarly, making use of Google Trends and scraping online data from job postings, layoff announcements, and wage reports to develop economic indicators that supplement government data.14 Additional information, such as retail or online purchasing trends (tracked via bar-code scanners in brick-and-mortar stores or product numbers in online purchases) , search queries (inputs into online search engines)
From page 128...
... Although they constructed the model to be simpler than the one used by the National Association of Realtors, by using real-time data, they were able to make predictions more accurately than the experts.17 Federal statistical agencies have begun to use big data sources in their evaluations. For example, the BLS currently uses "web-scrape characteristics for hedonics" to help calculate the Consumer Price Index (CPI)
From page 129...
... . Although private sector data sets are clearly of significant value to the companies that collect them, there may well be motivations on both sides to share and aggregate data and to make aggregate statistics available to researchers and the public.23 Second, large-scale collection and analysis of nonpublic information, such as credit-card transactions, or even some publicly available data, such as social media information, and much of the administrative data described above could have significant privacy or ethical implications, which will not be explored here.
From page 130...
... Of these methods, participant observation and ethnographic interviewing have contributed the most to an understanding of the changing nature of work. Participant observation (or fieldwork)
From page 131...
... DATA SOURCES AND METHODS 131 ducted in mines,26,27 factories,28,29 offices,30 construction sites,31,32 and other work settings.33,34 In fact, field studies such as these provided the situated, contextual insights that enabled sociologists to elaborate on theories of bureaucratic organizing as well as the grounding for large-scale survey research on the nature of work and work life throughout the remainder of the century. Ethnographic research continues to provide grounded understandings of the changing nature of work in the 21st century.
From page 132...
... Recent approaches to analyzing the types of jobs subject to automation have involved matching current technological capabilities with the skills or tasks associated with common occupational fields, in particular, as identified in the O* NET system (see the section "Tracking Changing Occupational and Skill Requirements" above for a discussion of O*
From page 133...
... A full report of these results is currently in development for The Oxford Handbook of Skills and Training.50 Another analysis came from the McKinsey Global Institute, whose researchers evaluated which of 2,000 occupational activities characterized by the O* NET system could be automated through some application of current technologies.
From page 134...
... 239, https://www.whitehouse.gov/sites/default/files/docs/ ERP_2016_Book_Complete%20JA.pdf. Data from Bureau of Labor Statistics; C.B.
From page 135...
... However, collection and curation of such statistics is resource-intensive, and data sets must be updated periodically, while limited update frequencies can be a challenge for researchers. There is potential for technology to reduce some of these costs, for instance, via automated AI telephone interview systems, large-scale digital surveys, and more effective use of administrative data and the increasing amounts of digital data created for other purposes.
From page 136...
... The use of microscale social science methods, such as interview, field work, and other ethnographic approaches, can help researchers bridge these critical gaps by testing conclusions from quantitative data and lead to new hypotheses and study, survey, or experiment design. There are few methods for quantifying technological progress or foundational advances in science and engineering that can be used for developing indicators of impending changes to the labor economy.


This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More information on Chapter Skim is available.