National Academies Press: OpenBook

A Consumer Food Data System for 2030 and Beyond (2020)

Chapter: 1 Introduction

« Previous: Summary
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

1

Introduction

1.1. WHY MEASURE THE POPULATION’S FOOD INTAKE AND NUTRITION?

Patterns of food consumption and nutritional intake strongly affect the population’s health and well-being in the United States, as in every other country. The Economic Research Service (ERS), in part through its Consumer Food Data System (CFDS), advances our understanding of these impacts.1 Food and nutrition intake influence diverse outcomes, including risk of chronic disease, risk of death, and economic costs. The economic burden of diet-related diseases amounts to trillions of dollars annually: negative outcomes associated with obesity and overweight alone are estimated to cost $1.42 trillion every year ($428 billion in direct expenditures and $989 billion in lost productivity); cardiovascular diseases cost $316 billion ($190 billion in direct expenditures and $126 billion in lost productivity); and type 2 diabetes costs $320 billion ($112 billion in direct expenditures

___________________

1 The agency’s mission is “to anticipate trends and emerging issues in agriculture, food, the environment, and rural America and to conduct high-quality, objective economic research to inform and enhance public and private decision making. ERS shapes its research program and products to serve those who routinely make or influence public policy and program decisions. Key clientele include White House and U.S. Department of Agriculture (USDA) policy officials; the U.S. Congress; program administrators and managers; other federal agencies; state and local government officials; and organizations, including farm and industry groups and those studying food assistance. ERS research provides context for and informs the decisions that affect the agricultural sector, which in turn benefits everyone with efficient stewardship of our agricultural resources and the economic prosperity of the sector.” See https://www.ers.usda.gov/about-ers.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

and $208 billion in lost productivity) (Milken Institute, 2016; Benjamin et al., 2017; Centers for Medicare & Medicaid Services, 2017). Consequently, understanding why people choose foods and how food assistance programs affect these choices is crucial.

ERS research influences real-world policy making and public spending for health insurance and food and nutrition assistance programs, which account for roughly $100 billion in federal spending (Oliveira, 2017). U.S. Department of Agriculture’s (USDA’s) expenditures on the Supplemental Nutrition Assistance Program (SNAP), formerly the Food Stamps Program, were $70.8 billion in 2016. During a typical month, 42.2 million people participated in SNAP. In the same year, expenditures for the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) were about $6 billion, and the program served an average of 7.3 million people per month, including some of the country’s most vulnerable populations. The program is restricted to pregnant, postpartum, and breastfeeding women and children under age 5. Expenditures on the National School Lunch Program were $14 billion (for an average daily participation of 30 million people), and expenditures on the National School Breakfast Program were $4 billion (for an average daily participation of 15 million people).

Another perspective is offered by thinking about how many people are participating in these programs across their life cycle and not simply at a given point in time. From this perspective, the school meals programs and WIC have very large footprints, with all hot meals at schools being subsidized by the school meals programs and with WIC serving more than half of all infants. With this level of spending and impacts on this many people at stake, it is essential that the design of outcomes-driven policies be as well informed as possible. Investments in the data used to inform those policy choices can yield large returns in program effectiveness.

In an array of health-related policy areas, ERS research on agriculture, food, food assistance programs and the food environment, and nutrition programs advances the public good.

1.2. GOALS OF A CONSUMER FOOD DATA SYSTEM

ERS’s vision for its Food Economics Division (FED) is to build a comprehensive, integrated data system to efficiently deliver credible evidence for informing research and policy. Data collection and sampling designs should always be motivated by the important research and policy questions to be answered and take into account possible variations in policy that may affect outcomes either by design or by accident and the characteristics of the targeted and ultimately affected populations. For example, if it is anticipated that research will employ instrumental variables (or other econometric

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

methods for causal inference), then data collection for potential instruments should be considered part of the research task. If it is anticipated that natural experiments based on existing policy variations will be employed, then data on those policy variations become essential.

Along the spectrum of data uses, the CFDS is designed for “monitoring, identifying, and understanding changes in food supply, purchases, and consumption patterns” for individuals, households, and markets (Larimore et al., 2018). Components of the CFDS include population surveys, either stand-alone or as modules added to extant surveys, many of which are fielded by other statistical agencies. They also include administrative data residing within USDA programs and proprietary commercial data, as well as products created by blending across all these sources. The desirable characteristics and qualities of a CFDS, and recommendations for achieving them, are examined in detail in Chapter 4.

The CFDS is structured to track sequential elements of the food supply system, focusing on consumer acquisition. In the food supply system, food (commodities) moves from the agricultural sector (farmers) through food processors and distributors to grocery stores and restaurants (retailers) to reach the ultimate consumer (see Figure 1.1). In return, money flows from the consumer through retailers, processors, and distributors, eventually to reach the ultimate producer. Food assistance programs can alter the relationship between consumers and retailers in a variety of ways, for example depending on whether they offer vouchers for food or directly provide it. Changes

Image
FIGURE 1.1. Food supply system (chain).
NOTES: TEFAP = The Emergency Food Assistance Program.
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

in supply are reflected through the chain to consumers, and changes in consumption are reflected back through the supply chain to producers. Food losses occur at each transition. Food acquisition obviously affects food consumption, which is the direct link to health and well-being. Thus, understanding the food supply system is a critical part of understanding the health of the economy as well as the health and well-being of the people.

Both the supply and the prices associated with food commodities (e.g., as produced by farmers, represented in Figure 1.1) are measured by the USDA’s National Agricultural Statistics Service, and both are used as inputs to ERS’s Food Availability Data System, a data summary of the food supply chain. Economic aspects of food processors, distributors, grocery stores, and restaurants are monitored by the U.S. Department of Commerce agencies (the Census Bureau and the Bureau of Economic Analysis) and by the U.S. Department of Labor’s Bureau of Labor Statistics (BLS). ERS’s CFDS (the main topic of this report) aims to illuminate the consumer part of the chain, including acquisition as well as consumption. The health and well-being of the population are measured and monitored by various agencies within the U.S. Department of Health and Human Services (HHS), but without funding from USDA these agencies would lack data on food insecurity, for example. In their efforts to understand the population’s health, their focus is not on the food environment or the food assistance landscape either. Hence, the responsibility for measuring and monitoring the food supply system and its impacts, including the ways it interacts with food assistance programs, falls to many different agencies. Fulfilling that responsibility requires significant collaborative efforts.

Broadly speaking, this data system for tracking sequential elements of the food supply system, focusing on consumer acquisition, is called upon to fulfill descriptive or monitoring needs, some of them essential to developing official statistics. It is also called upon to support research to examine program impact or address other causal questions. In later chapters of this report, we examine details of the current CFDS infrastructure and propose solutions for improving it. A big part of the solution involves integrating survey, administrative, proprietary commercial, and other kinds of data sources in order to exploit the strengths of each type.2

Key policy areas for the CFDS, which fall within the agency’s purview, are headlined by:

  1. Agriculture and the food system. How do upstream factors (such as the agroecological environment, agriculture policy, innovations in food manufacturing, new product development, and labor policy)

___________________

2 CFDS integration of multiple data sources is discussed in detail in Chapter 2.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
  1. influence consumer food outcomes? Conversely, how are changes in consumer tastes and preferences about the food system communicated back up the supply chain?

  2. The food retail environment. How do the location and competitiveness of small food retailers, large supermarkets and superstores, and restaurants influence consumer food outcomes? Conversely, how are changes in consumer outcomes and preferences concerning the retail environment communicated back up the demand chain to influence retail competition and location decisions?
  3. Healthfulness of U.S. diets at all income levels. For all of these factors that influence consumer food purchase and acquisition, what is the ultimate effect on nutrition, health, chronic disease, and mortality risk?
  4. Economic determinants of consumer demand. How do incomes and consumer preferences influence food choices and how are prices related to food choices?
  5. Food and nutrition safety net programs. How do SNAP, WIC, school meals, the Child and Adult Care Food Program (CACFP), and other programs affect (i) food acquisition and use by people at all income levels and, in turn, (ii) the food security of and nutritionally healthy consumption by the population?

Understanding the Food Environment and Its Relationship to Health

Food choices and diet quality are influenced by the many opportunities, constraints, and challenges that consumers face in the food environment. The Centers for Disease Control and Prevention (CDC) defines the food environment to be “the physical presence of food that affects a person’s diet; a person’s proximity to food store locations; the distribution of food stores, food service, and any physical entity by which food may be obtained; or a connected system that allows access to food.”3

The types of food outlets that are accessible to consumers dictate the product availability, quality, and prices those consumers face. Outlets exist across a range of types, from supermarkets to convenience stores to restaurants. A consumer’s (geographical) food access reflects the proximity and types of restaurants and stores present in his or her local environment, and where retail food stores are concerned an important feature is whether a store is authorized to participate in one of the USDA food assistance programs. Food is also provided by schools, child care providers, pantries, and nursing homes, and all these can affect the food environment.

___________________

3 See https://www.cdc.gov/healthyplaces/healthtopics/healthyfood/general.htm.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

Within this topic area of the role the food environment plays in people’s food choices, key descriptive and monitoring questions include these:

  • How many people have limited access to sources of healthy and affordable foods?
  • Does ease of access matter for nutritional quality of purchases?
  • Where do people buy and consume food?
  • How does food preparation affect food safety?
  • Do food assistance programs affect these choices?

Causal impact questions (with program policy implications) include these:

  • How do food store access, access to restaurants, and the larger food environment impact food choices, diet, and diet-related health?
  • How do food access and regional price variation jointly affect these outcomes?
  • How do consumers respond to new information and product attributes?
  • How do other factors, such as income, time resources, and consumers’ preferences and knowledge, affect food consumption decisions; and how have these factors and connections changed over time?
  • How do food assistance programs affect these choices?

Concerning ease of access, USDA has provided mapping tools in applications such as the Healthy Food Finance Initiative and the Food Access Research Database. There has also been some debate over new SNAP stocking requirements, so these two data projects are useful for community and local planning use.

Surveys currently serve as one data source to address many of these questions. In particular, ERS’s National Household Food Acquisition and Purchase Survey (FoodAPS)—described in detail in Chapter 2—generates information not captured elsewhere about spending by consumers at specific retailers as well as access to other sources of food (including food in-kind from local organizations, family, neighbors, or friends), about distance to primary food retailers, and about consumer attitudes and opinions regarding food retail access for a point in time. The central role of surveys notwithstanding, the use of both proprietary commercial data (e.g., NPD Group data on the locations of restaurants and Nielsen price data for retailers) and administrative data (e.g., from SNAP and WIC programs) is expanding rapidly as new opportunities emerge. At the same time, surveys are becoming less viable as the sole source of information for reasons of cost and quality.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

Improved access to administrative and proprietary data is opening new opportunities, though it is worth noting that surveys are key sources of data covering the entire population. For example, due to survey problems with underreporting and, to a lesser extent, with over-reporting, administrative data on use of any single program are typically superior for measuring participation. However, such administrative data cannot provide data on the universe of individuals who could participate in a program but are instead restricted to those who have participated. Without also knowing about nonparticipants or participants in other programs, key questions about policy effects, program take-up, and impacts of programs on health and nutrition outcomes cannot be answered.

Finally, it is difficult to generate causal estimates of the effects of programs outside randomized control trials without contextual information about program rules at the state and local levels. The panel’s recommendations for advancing ERS’s CFDS (Chapter 4) are largely focused on survey collections, enabling linkages between survey and nonsurvey data sources, establishing and maintaining searchable policy databases, and monitoring the quality and coverage of proprietary data sources.

Supporting Program Policy and Administration for the Food and Nutrition Safety Net

Food assistance programs serve a large share of the population. At some point during a given year, about one in four Americans participate in at least one of USDA’s 15 domestic food and nutrition assistance programs. As indicated in the budget figures cited at the beginning of the chapter, these programs accounted for $98.6 billion in spending in 2017—more than two-thirds of USDA’s annual budget, but well below historical highs (Oliveira, 2018, p. iii). The largest of these programs in terms of spending is SNAP, which serves as the foundation for the country’s nutrition safety net. While at around $6 billion annually WIC has lower expenditures than several other programs, it touches around half of all infants. Similarly, the school meals programs subsidizes a large share of meals at schools every day.

Policy makers, other stakeholders, and the public are interested in understanding the impacts of these substantial investments. This requires accurate information about program participation; the factors that affect take-up of programs among those eligible to participate; the profiles of program participants; and the food choices, nutrition, and health outcomes associated with participation. Importantly, the influence of program participation should be modeled in a way that makes it possible to study causal relationships and to allow researchers and policy makers to monitor the performance of these programs. Questions, some descriptive and

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

some causal, that need to be regularly answered for affective administration of safety net programs include the following:

  • Who participates in USDA food assistance programs? And in what ways are SNAP and low-income households similar to or different from the overall population along different dimensions? (descriptive)
  • Among those who are eligible, who does not participate? (descriptive)
  • Among those who do not participate, why not? Is this driven by policy or other factors? (causal)
  • What are the program participation rates, both as a percentage of the relevant categorically eligible populations and as a percentage of all eligible persons? (descriptive)
  • How are participation rates affected by program rules? (causal)
  • What are the program entry rates, exit rates, average spell durations for cohorts of new entrants, and average spell durations for a cross-section of current participants? (descriptive)
  • What is the dietary quality profile of the U.S. population? What foods do people buy, how much do they pay, where do they shop, and what is the nutritional quality of food expenditures? (descriptive)
  • What is the dietary quality profile of food expenditures for SNAP participants, low-income non-SNAP participants, and higher-income non-SNAP participants? What is the profile for those people who are joint SNAP and WIC participants? For those who do not participate? And how do these programs interact with CACFP and the school meals programs in eligibility, participation, and effects? (descriptive)
  • What is the dietary profile of food intake (rather than food expenditures) for the different groups described above, including those participating and those not participating in the different programs? (descriptive)
  • How do the respective programs affect food intake? (causal)

Since most food assistance programs are funded and administered through the USDA’s Food and Nutrition Services (FNS), it is natural that measuring and monitoring the impact of those programs should fall to a separate group within USDA, such as FED. For policy purposes, the causal answers to the questions listed above and below are even more important than the descriptive findings above.

Key additional causal questions exploring program impact include these:

  • How do SNAP, WIC, the school meals programs, and other programs affect food spending and dietary intake?
  • How do the programs affect nutrition, food security, health, and use of health care systems?
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
  • How do food spending and food intake respond to prices, income, and program benefits in a demand-systems framework consistent with economic theory?
  • How does the safety net function during economic contractions?
  • Does the food assistance safety net have unintended consequences?

To answer these questions and those envisioned going forward, researchers will rely on survey, administrative, and, increasingly, proprietary commercial sources. For example, FoodAPS has been successfully used to assess food expenditure and acquisition at a point in time; the National Health and Nutrition Examination Survey (NHANES) measures food and nutrition intake; the Current Population Survey’s Food Security Supplement measures food security; and the Survey of Income and Program Participation (directed by the Census Bureau) and the Panel Study of Income Dynamics (PSID, directed by faculty at the University of Michigan) produce longitudinal data that can be used to study program participation.4 Examples of administrative sources used to help answer questions about safety net programs include FNS data on program participation that originates at the state level; state-provided individual-level data on participants’ and firms’ program benefits and their use; SNAP quality control data; and the WIC Participant Characteristics Data (a census of WIC participants constructed with administrative records in April of even-numbered years).

An example of contextual data that allow researchers to study causal questions about SNAP is the USDA/ERS SNAP Policy Database. Examples of proprietary data that have been used to study how people make choices about food acquisition are scanner data collected by IRI Worldwide and Nielsen on a panel of households or from a set of retailers. Analyses employing integrated or linked survey, commercial, and administrative approaches can take advantage of the wide-ranging outcome variables in surveys and the large sample sizes and geographical disaggregation of administrative data with high-frequency data from commercial sources. These combinations have also enhanced researchers’ capacity to measure “small area estimates” of outcomes such as food security by county. Administrative data also provide the ability to measure participation in many of these programs with considerably less error than survey data. Commercial data provide some high-frequency data for relatively low cost, and when combined with the other two sources, can be even more useful.

In Chapter 2, we review examples of these strategies from the literature, along with the barriers they present. Barriers include sample size limitations

___________________

4 Current data solutions and their strengths and weaknesses are described in detail in Chapter 2; panel-envisioned solutions are advanced in Chapter 4. For example, the well-known problem of misreporting program participation is discussed there.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

and self-reported program participation indicators in surveys; lack of outcome variables and a corresponding measure of the population in administrative data sources; and lack of coverage and benchmarks about quality in commercial data. Recommendations (Chapter 4) focus on the following: (i) continued targeted investment in surveys; (ii) expanded coordination and access to administrative sources; (iii) expanded use and continued quality assessment of commercial data; and (iv) expanded tracking of state and local eligibility and implementation across all USDA food assistance programs, as well as tracking of stores where benefits are redeemed.

Supporting Research on the Healthfulness of U.S. Diets (at all income levels and for all types of people)

Making headway in understanding the complex links between diet, nutrition, and health outcomes is critical for informing government food policy strategies to improve a population’s well-being (Duffey et al., 2010; Olson, 1999; Marshall, Burrows, and Collins, 2014). A key example is the relationship between poor diets, coupled with physical inactivity, and obesity, which is a leading cause of preventable death and disability in the United States and in many other countries. Gorski and Roberto (2015) describe the ways in which current food environments exploit biological, psychological, social, and economic vulnerabilities that encourage overeating, and they review recent public health policies to promote healthier diet patterns, including mandates, restrictions, economic incentives, marketing limits, information provision, and environmental defaults. The authors (p. 81) point out, “unhealthy diet patterns, including high intake of added sugars, trans fats, and excess sodium intake are linked with obesity, heart disease, type 2 diabetes, cancer, high blood pressure, and stroke.” Yet we know little about the causal link between these policy levers and changes in long-run health outcomes such as obesity. The CFDS offers data that may provide insight into the causal links between these policy levers and diet as well as links to longer-run outcomes, and it also allows for surveillance.

The 2015–2020 Dietary Guidelines Advisory Committee report stated, “health and optimal nutrition and weight management cannot be achieved without a focus on the synergistic linkages and interactions between individuals and their environments, and understanding the different domains of food-related environmental influences” (U.S. Departments of Agriculture and Health and Human Services, 2015, p. 1).5 Evaluation of food-related health policies to determine how well they are accomplishing their goals requires access to high-quality consumer-level panel data. While the Guidelines

___________________

5 This advisory committee serves HHS and USDA, which jointly publish the Dietary Guidelines for Americans (Dietary Guidelines) every 5 years.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

process typically uses NHANES data (which has the weakness of 24-hour recall surveys) in its analyses, rather than data generated by ERS, FoodAPS might provide alternative data that could be useful in assessing existing U.S. dietary patterns as well as adjustments that might be made to diets (again recalling the weaknesses of 24-hour-recall surveys such as NHANES).6

A partial list of questions about food-related health policy including diets, nutrition, obesity, and health care—some of which are causal in nature and some of which are descriptive—includes the following:

  • What foods do households buy? What is the nutritional quality of the foods they acquire? What about the food they consume? How much are they willing to pay and where do they shop?
  • What impact has the application of federal nutritional standards for all foods and beverages served in schools had on overweight and obesity among school-age children?
  • What are the impacts of SNAP and other programs on food purchases, food consumption, diet quality, food insecurity, overweight and obesity, and other health outcomes?
  • What effects have food assistance and nutritional educational programs had on the nutritional quality of diets of those served by the programs?

Data sources available to help to provide answers to these questions include nationally representative surveys as well as data from proprietary sources, which are typically not nationally representative. Key surveys for tracking food acquisition and consumption include NHANES,7 sponsored primarily by the National Center for Health Statistics (NCHS) of CDC, which uses a dietary recall survey to collect information about food intake; the Consumer Expenditure Survey, sponsored by BLS, which collects data on expenditures for food at home and food away from home using two 1-week diary surveys; the food expenditure questions in PSID, funded by ERS; and the above-noted FoodAPS, sponsored by ERS, which collects household food expense data by asking selected households to scan their food receipts, as well through food diaries and telephone interviews.

Other nonsurvey data sources that are increasingly being used to measure food acquisition include proprietary data, such as the Consumer Network by IRI Worldwide and HomeScans by Nielsen, which provide con-

___________________

6 The same parent department, USDA, comprises both ERS and the Center for Nutrition Policy and Promotion, which is the department’s lead agency on the Dietary Guidelines for Americans.

7 NHANES collects a wide variety of information other than food intake as well. See http:///www.cdc.gov/nchs/nhanes/about_nhanes.htm.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

sumer data from a panel of volunteers, and Infoscan from IRI Worldwide, which provides scanner data from food stores. Other surveys, such as the Current Population Survey’s Food Security Supplement, National Health Interview Survey (NHIS), NHANES, and PSID, all track food security.

Facilitating Answers: ERS’s Data Collection Approach

Increasingly, ERS’s approach to answering the types of research and policy questions described above is to emphasize an integrated mix of data sources. To address many of these major public health policy issues, research on causal effects of programs on nutrition and health outcomes is particularly important. To advance research in topic areas that fall within its purview, FED is actively engaged in developing a data collection strategy that draws on a wide variety of sources. Bringing together all these sources makes it possible to combine survey, administrative, and proprietary data on the food environment, production, processing, and food items (from retailers and restaurants) with related information on the consumers, including their nutritional intake and status, the affordability of their food purchases, and their health outcomes.

Larimore and colleagues (2018) describe three specific FED initiatives reflecting this multiple-data-source approach that were born, in part, out of recommendations from an earlier Committee on National Statistics report (NRC, 2005): (i) expanding the use of proprietary data; (ii) developing the Next Generation Data Platform; and (iii) creating an innovative consumer acquisition survey, realized as FoodAPS, which was first fielded in 2012.8

ERS has had extensive experience (relative to most statistical agencies) with commercial data, including acquiring it, assessing its quality, and using it to answer questions about food acquisition. For the most part, these data fall into one of three categories: proprietary retail scanner data,9 household panel and scanner data, and food store and restaurant name and location data.10 Retail scanner data are collected in stores during customer checkout, while household scanner data are collected using hand-held scanners provided to participating households. An advantage of scanner data (over survey data, for example) is that reading devices detect and record exactly which product is purchased and sometimes, though not always, also collect its price (Larimore et al., 2018, p. 8).;11Chapters 2 and 4 discuss the

___________________

8 Information in this section is drawn from Larimore et al. (2018) and from presentations by ERS staff and others at the workshops described in Appendix A.

9 A scanner uses a laser to read the Universal Product Code (UPC) on a store item’s label.

10 Presentation to the panel by Mary Muth, September 21, 2018.

11 Advantages and limitations of these commercial sources are covered in detailed in Chapter 2, section 2.4.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

insights that have and can be drawn from scanner data, and also issues—such as their frequent lack of critical geographic coverage or of identifiers needed to link stores across datasets—that need further attention in order to improve returns on future investments in such resources.

Commercial data have been usefully applied by ERS and other researchers to policy-oriented matters, such as identifying the composition of food purchases by WIC household versus non-WIC households (e.g., types of products, such as breakfast cereals); identifying the use of WIC benefits (by identified food items); measuring the effects of WIC program participation on food purchases; and evaluating the effects of program changes on food purchases over time.12 However, this WIC example also suggests the value of linking administrative data to such proprietary data, if possible, because otherwise WIC use has to be inferred from self-reports or from the food item scanning records, both of which are likely to lead to measurement error.

The Next Generation Data Platform, discussed in detail in Chapter 2 (section 2.2), is a strategic partnership formed in 2012 with USDA’s FNS and the Census Bureau. This joint project is a long-term effort to acquire state-level administrative data for USDA nutrition assistance programs—especially SNAP and WIC—and to make those data available for linkage to other administrative files and surveys. In this work, FNS contacts state SNAP and WIC offices to encourage them to share their USDA administrative data for the project, and the Census Bureau negotiates a data-sharing agreement that provides mutual benefits for all parties.13 For example, one anticipated research application of the program is the ability to evaluate SNAP and WIC participation and nonparticipation by county within a state, as well as by various demographic and other variables captured in the Census Bureau’s American Community Survey. These data will also permit research into how food program rules affect the take-up of programs as well as other outcomes.

As noted earlier (and as detailed in Chapter 2), FoodAPS, which was designed in collaboration with FNS, was conceived of in response to the recognized limitations of U.S. consumption and expenditure surveys. For example, dietary recall data, which are generated by food consumption surveys, are collected to learn about patterns of individual-level food consumption and about the nutrient content of foods consumed, but these data convey no information about where the foods were acquired. Consumer expenditure surveys provide information for learning about household food expenditures. Both of these traditional sources fail to provide a complete

___________________

12Muth et al. (2016) provide a full accounting of the application of commercial data, particular scanner data, to food policy research.

13 As of mid-2017, 20 state SNAP agencies (including 39 counties in California) and 11 state WIC agencies were partners in the Next Generation Data Platform. See Prell (2018) for a summary of state-level participation.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

picture of the amount and types of foods that households acquire and how those acquisitions are affected by food prices, the local food environment, and participation in USDA’s food and nutrition assistance programs.

FoodAPS, which was fielded from April 2012 through January 2013, was the first nationally representative survey designed to collect comprehensive data on foods that households purchase or acquire from all sources whether obtained by money or for free. It is notable in capturing data on the way most households also tap into “non-purchased” or so-called free sources—such as food pantries and food supplied by friends and relatives as well as by employers, schools, and child care providers—to supplement their bought food, and these foods do not appear in expenditure surveys (Larimore et al., 2018, p. 9).

FoodAPS data have been used to address important policy-relevant issues with both descriptive and causal approaches. These issues include where households acquire food in a typical week, which foods they acquire, and how much they pay (Todd and Scharadin, 2016); which factors affect households’ decisions about where to shop for food (Ver Ploeg et al., 2015; Ver Ploeg, Larimore, and Wilde, 2017); which household characteristics are associated with increased childhood obesity risks (Jo, 2017); how SNAP benefits are used over the course of the benefit month (Smith et al., 2016); price sensitivity among WIC households (Dong et al., 2016); and how price variation across geographic areas affects the adequacy of SNAP benefits (Basu, Wimer, and Seligman, 2016). While the acquisition data that FoodAPS provides are rich, they limit a researcher’s ability to study some causal questions because the data derive from food acquisition at a single point in time.

Elsewhere on the survey front, ERS has actively expanded its portfolio by sponsoring or cosponsoring modules on surveys conducted by other agencies. Among the noteworthy modules that have been developed are the Food Security Supplement, which has been added to many surveys (see the list in Box 2.1); the Flexible Consumer Behavior Survey, a module which has been added to NHANES, conducted by NCHS; and the Eating and Health Module, which has been added to the American Time Use Survey, part of the Current Population survey conducted by BLS.

Improving researchers’ access to data is another important aspect of ERS’s CFDS data strategy (discussed in detail in Chapter 4). ERS collaborates with researchers in academia and with research organizations through grants and cooperative agreements. ERS has sponsored FoodAPS research by external researchers through a National Bureau of Economic Research grants program and a University of Kentucky Center for Poverty Research (UKCPR) grant program.14 Additional research to enhance

___________________

14 ERS also sponsored the UKCPR grant program to conduct research on NHIS and PSID.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

the nation’s nutrition assistance programs has been sponsored by USDA through the Tufts University/University of Connecticut Research Innovation and Development Grants in Economics Program, established to “address national objectives for improved food security and dietary quality.”15

ERS has also broadened public access to FoodAPS by removing identifying information about survey participants and posting the edited files and documentation on the ERS Website. ERS has made confidential survey data from FoodAPS available to researchers through a secure data enclave at NORC, an independent research organization at the University of Chicago. Beyond this arrangement, the agency is working with the Census Bureau, FNS, and state partners to make confidential administrative data and linked data available through the national network of Federal Statistical Research Data Centers. Data from modules and supplements cosponsored by ERS are available through the access procedures provided by the agency that collects the data. Commercial scanner data on people and stores are available for collaborative work with ERS researchers, although (as discussed later) the commercial entities providing such data impose limitations that mean these data are not always available to those at public institutions.

1.3. CHARGE TO THE PANEL; REPORT THEMES AND STRUCTURE

In 2017, ERS’s FED asked the National Academies of Sciences, Engineering, and Medicine’s Committee on National Statistics (CNSTAT) to provide guidance for further development of its consumer food data system over the next decade. The mission of FED, as described to the panel,16 is to evaluate contemporary and anticipated food policy and program objectives, as well as market trends and dynamics; to develop the necessary data and information infrastructure to examine evolving questions; and to produce the right products and information for the administration, the Congress, and the public on consumer food choice behaviors and outcomes such as nutrition and health. In support of this mission, according to its Webpage:

FED conducts economic research and analysis on policy-relevant issues related to the food sector (food safety, food prices, and markets); consumer behavior related to food choices (food consumption, diet quality, and nutrition); and food and nutrition assistance programs (Supplemental Nutrition Assistance Program (SNAP), Women, Infants, and Children’s

___________________

15 See https://ridge.nutrition.tufts.edu/research-grants/2019.

16 Presentation to the panel by Mark Denbaly, April 16, 2018.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

Program (WIC), National School Lunch Program). Food and Economics Division also provides data and statistics on food prices, food expenditures, and the food supply chain.17

FED adopted guidance from an earlier report by CNSTAT (NRC, 2005) to create a blueprint for enhancing its consumer data program—a portfolio of data resources that measure the country’s food and nutrition conditions and the factors that affect those conditions.

This report is intended to reconsider how FED’s consumer food data collections are conceived, how it adapts over time to challenges, and how it exploits new opportunities. The testimony that the panel heard during its public meetings was striking in its portrayal of the challenges faced by traditional survey approaches, of the new opportunities (and problems to overcome) in exploiting administrative and commercial data, and of the benefits of blending all three types of data sources, that is, survey, administrative, and commercial data.

No single data source—or even single data type—can provide all the information needed to understand the food sector, including consumption, diet, and nutrition. Policy makers and researchers who rely on FED data include those within USDA as well as those in other agencies whose responsibilities are related to food outcomes, including those within HHS and the Environmental Protection Agency. Other stakeholders include state and local policy makers, food producers, food retailers, consumer groups, think tanks, nonprofit groups, and academic researchers. The multiplicity of data sources and distributed food-related responsibilities make collaborative efforts imperative for reducing duplication and gaps. In particular, the panel that produced the 2005 report (NRC, 2005) supported collaborative interagency activities to create linkages between surveys, administrative data, and other data; to develop food-related modules to be used on relevant federal surveys; and to evaluate use of proprietary data (collected, owned, and made available by commercial firms).

Aims and Focus of This Report

This report is intended to provide a blueprint for ERS’s Food Economics Division for its data strategy over the next decade. ERS leadership specifically asked that the panel address the following questions:18

  • Are the current data collected or supported by ERS delivering policy-relevant evidence that is as credible and insightful as possible?

___________________

17 See https://www.ers.usda.gov/about-ers/agency-structure/food-economics-division-fed.

18 Presentation to the panel by Jay Variyam, April 16, 2018.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
  • Is the current multiprong data approach—and particularly the balance between the use of survey, administrative, and commercial sources—the correct one, or is there a better practical use of resources?
  • Given key current and emerging policy questions, which kinds of data are anticipated to become the most valuable to researchers, policy makers, and program administrators going forward?
  • Should the nation have a comprehensive food acquisition survey like FoodAPS and, if so, how frequently should it be conducted? If not, what are the alternative uses of resources now used to fund this survey?
  • Considering the new data opportunities made possible by the Web, by wearable devices, by mobile technologies, by apps, and by big data, which ones should ERS be considering?

Motivated by the goal to improve the data infrastructure supporting research and policy in the topic areas outlined above, the authoring panel of this report was charged with addressing the statement of task (see Box 1.1). In addressing this charge, the panel identified key questions that CFDS data are used to address. Data are needed to produce descriptive content, to serve monitoring functions, and to support causal and other kinds of policy research spanning topic areas ranging from the food environment, to informing program policy, to understanding the healthfulness of people’s diets. The present report first describes the current ERS data infrastructure—which includes survey, administrative, commercial, and combined data elements—and then proposes data solutions to better answer questions that, as of now, cannot be satisfactorily addressed.

One prominent part of this charge, although certainly not the only one, is to provide guidance to ERS on directions for future iterations of the FoodAPS survey. Specifically, the panel aimed to answer these questions: Should FoodAPS be a permanent data collection effort? Is FoodAPS as currently constructed worth the investment or should it be pared back? Can FoodAPS be better combined with other administrative data? What are the alternative data investment options? Would alternative investments generate similar or greater research and policy content or not? Recommendations about the future direction of FoodAPS, formally presented in Chapter 4, include three messages: (1) a caution on costs—that USDA should note the expense of FoodAPS and be careful about not displacing other data sources and staff activities through over-investment; (2) the need to make the survey cycle predictable—if budgets permit continued investment in FoodAPS, the survey should be fielded on a consistent time interval; and (3) the importance of continued cost reduction—FoodAPS can be streamlined and its quality enhanced simultaneously through continued investment in linkage

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

across survey and administrative and commercial data sources, and early planning for how those data sources will be used in public use files.

The panel was also asked for guidance about building the agency’s broader data infrastructure. Relevant questions here are: What is the feasibility of Web-based data collection methods? How can expanded investment in food data (e.g., UPC product dictionaries and restaurant menu

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

databases) complement existing data resources? How can development of regional food price indices be enhanced using retail scanner data, and how can geographic components be enhanced with area-based local demographics and policy characteristics?

The process by which the panel met its charge included four open meetings in a workshop format to gather information. These meetings were intended to inform the panel as it began shaping a strategy for producing a report that fully addresses its charge. Workshop topics of interest included the current and potential use of commercial and other nongovernment, nonsurvey data sources; users’ perspectives on directions for ERS’s FoodAPS survey; issues with data quality; and the linking of data sources. At later meetings, researchers presented ideas for improving food and nutrition data—including the integration of commercial and administrative data—to inform key policy issues. Among the topics discussed were the following:

  1. The value (and limits) of linking SNAP or other food assistance administrative data not to surveys but to other types of administrative data to provide a “universe” of people affected by the programs. Possible universe files could be provided by data such as state unemployment insurance system data on workers covered by the system; Medicaid enrollment and claims data, which would cover a large share of low-income individuals; and public K–12 education data, which would cover a large share of families with children.
  2. The limits of existing survey data and suggestions about how they could be made more useful.
  3. Use of data from retail loyalty-card customers and other commercial data linked to state administrative records across most public programs to analyze a wide range of questions including how SNAP benefits are spent and what evidence is needed to design a “smarter SNAP.”
  4. Food consumption data needs for studying the determinants of diet quality and health-related outcomes (healthy eating index scores and body mass indexes are examples of outcomes indicators).

The panel also heard from experts on the potential of data integration and linkages for policy research and the use of administrative data. Practices being developed by the statistical agencies for combining data sources were also discussed, including the Next Generation Data Platform—a Census-ERS-FNS collaboration that links SNAP (19 states and 39 counties in California participated in 2017) and WIC (11 states) data to Census survey data and administrative data. Challenges discussed included the fact that not all states have participated in the Next Generation program. Since administrative data are often missing key pieces of information necessary

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

to produce thorough descriptive or causal analysis, linked data can resolve some missing data or measures. However, efforts to improve the quality and comprehensiveness of existing administrative data resources are an important first step. Also, administrative data are often not available for some of the most vulnerable geographic areas or communities. Finally, as recommended in Chapter 4, many researchers currently have difficulty accessing linked administrative data, and efforts are needed to broaden who is able to access such resources.

Additionally, the panel heard several presentations about data needs from those in the policy environment studying nutrition or food assistance programs and those running the programs at a more local level. A topic of particular interest was the U.S. Commission on Evidence-Based Policymaking’s guidance on the importance of states making individual-level participant data from federally funded programs available for research by the Federal Statistical System. Workshop presentations are summarized in detail in Appendix B of this report.

Outline of the Report

The remainder of this report describes ERS’s current consumer food data system, assesses remaining gaps for research and policy use, and outlines guidance for filling those gaps. Chapter 2 reviews the survey data sources relied on by ERS and other relevant statistical agencies and describes the purposes these data are intended to fulfill. The chapter then describes how survey sources and program administrative records have been combined to improve the accuracy and coverage of data used for statistical purposes. Next, it documents ERS’s use of commercial data and its assessments of the coverage and quality of those data, along with how those multiple data sources have been combined by ERS to produce useful resources for stakeholders. In some cases, the far-ranging data sources are used to generate statistics for descriptive or monitoring purposes; in others, these sources are used for research—ideally, to better understand causality—into program impacts and into the links between food/diet and health.

In Chapter 3, data and knowledge gaps in the areas of food, nutrition, and safety net research are identified. The chapter discusses the progress made by ERS to date in modernizing data infrastructure along with those policy and research questions that remain difficult to answer with the given data options. Specific data and measurement needs for addressing these questions are described.

Chapter 4 lays out strategies to advance the CFDS infrastructure. Most of the panel’s recommendations for moving ERS data forward are presented and supported here. Desirable characteristics and qualities of a consumer food data system are discussed, and then a path is laid out for development

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

of a forward-looking research- and policy-driven data infrastructure that will necessarily require integrating different kinds of data sources—most notably, survey and administrative data. Here, FoodAPS and complementary and alternative data sources are considered. Implications for the survey component of the CFDS are discussed alongside opportunities and challenges associated with expanding the use of administrative records and commercial data sources. Finally, the chapter discusses the issues of data access and confidentiality constraints as they relate to a statistical system that is increasingly based on multiple sources of data, acknowledging that overcoming these constraints will require investment.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×

This page intentionally left blank.

Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 15
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 16
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 17
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 18
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 19
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 20
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 21
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 22
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 23
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 24
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 25
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 26
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 27
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 28
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 29
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 30
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 31
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 32
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 33
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 34
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 35
Suggested Citation:"1 Introduction." National Academies of Sciences, Engineering, and Medicine. 2020. A Consumer Food Data System for 2030 and Beyond. Washington, DC: The National Academies Press. doi: 10.17226/25657.
×
Page 36
Next: 2 ERS's Current Consumer Food and Nutrition Data Infrastructure »
A Consumer Food Data System for 2030 and Beyond Get This Book
×
Buy Paperback | $70.00 Buy Ebook | $54.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Patterns of food consumption and nutritional intake strongly affect the population's health and well-being. The Food Economics Division of USDA's Economic Research Service (ERS) engages in research and data collection to inform policy making related to the leading federal nutrition assistance programs managed by USDA's Food and Nutrition Service. The ERS uses the Consumer Food Data System to understand why people choose foods, how food assistance programs affect these choices, and the health impacts of those choices.

At the request of ERS, A Consumer Food Data System for 2030 and Beyond provides a blueprint for ERS's Food Economics Division for its data strategy over the next decade. This report explores the quality of data collected, the data collection process, and the kinds of data that may be most valuable to researchers, policy makers, and program administrators going forward. The recommendations of A Consumer Food Data System for 2030 and Beyond will guide ERS to provide and sustain a multisource, interconnected, reliable data system.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!