Putting the Levels Together
Key Messages Noted by Participants
Following the sessions on the four levels of socioecological influence, Pamela Morris, professor of applied psychology at the New York University Steinhardt School of Culture, Education, and Human Development, sought to put the levels back together. Although much of Morris’s research has been on antipoverty programs and their effects on families and children rather than on obesity per se, she was trained with Urie Bronfenbrenner at Cornell University to look at systems-level processes, giving her a broad perspective to integrate the interactions across levels.
From an outsider’s perspective, Morris said, the research on food insecurity and obesity has several prominent gaps that limit the ability to inform interventions. First, what are the relevant aspects of food insecurity that are central to the relationship? Is it the level of food insecurity, as measured by hunger or lack of food? Is it the quality of food that low-income individuals and families eat? Or is it the volatility of access to food that influences obesity?
Each of these factors points toward a different policy response. If the level of food insecurity matters, then families need more resources. If it is quality that matters, families need access to more nutritious food. If the critical factor is volatility, the question becomes how to smooth consumption over a monthly period so that families do not have times of great need and less need.
Gaps also exist in the research methods that can be applied to these topics, said Morris. In particular, better ways are needed to measure sources of influence across levels and causal processes.
Volatility and Food Access
Volatility may be the most understudied of these issues, Morris said. Volatility is also a topic that has not been much attended to in poverty research. In child development research, income is typically annualized so that short-term fluctuations are smoothed out (Duncan and Brooks-Gunn, 1997; Cancian and Danziger, 2009; Gennetian et al., 2010). Economics research has looked at earnings volatility; for example, earnings volatility has been linked with material hardship and food insecurity (Bania and Leete, 2007; Jolliffe and Ziliak, 2008; Mills and Amick, 2010). Yet poverty research generally has examined the volatility of residential stability or family structure (Wu and Martinson, 1993; Ackerman et al., 1999) and not income instability. Also, this research tends to look at a single event rather than a chronic pattern of events.
If a low-income family cannot predict its income levels, it may experience more distress than in adjusting to a routinely low income. Variations in how much parents can purchase for their children could have a variety of deleterious effects, from disruptions of predictable food cycles to reductions in academic achievement. Volatility could reduce parents’ time and money investments in children’s human development (Becker, 1981). Volatility also can lead to stress, raising the possibility that stress and not just deprivation leads to behavioral outcomes (McLoyd, 1990). In this regard, animal research has demonstrated enormous effects of unpredictable feeding and social relationships on parental and offspring behaviors.
Another understudied subject, according to Morris, is the effect of developmental stage on the relationship between food insecurity and obesity. Food insecurity can have very different effects on an infant versus a preschooler versus a school-aged child versus an adolescent. A young child may be protected against hunger by the family or by food assistance programs, while an older child may have more autonomy in selecting foods to consume.
Children also may go through critical periods where food insecurity has especially dramatic effects. If so, interventions might be structured to protect children during these periods.
Finally, said Morris, causality is a critical factor. If a relationship is not causal, interventions will not necessarily make a difference to the desired outcome. Ways to determine causality include state-level variation, natural experiments, sibling and longitudinal fixed-effect studies that factor out unobserved influences, and the study of instrumental variables with randomized experiments, which essentially discards the naturally occurring variation in the predictor of interest and estimates causal relations.
In addition to theory, methods are important in research on food insecurity and obesity. Multiple factors need to be measured at multiple levels to understand the relationship from a biological, psychosocial, and environmental perspective. Only through these measurements can models and theories be tested in the context of individuals, households, neighborhoods, and policies.
AN INTEGRATIVE FRAMEWORK
Individuals are embedded within systems at various levels, from the microlevel to the eco-level to the macrolevel, said Morris, but the influences of factors at these various levels are not additive or even interactive but truly integrative. In addition, biological factors, including genetic influences, interact with environmental factors, producing outcomes for the developing person.
The standard Bronfenbrenner model can be represented by an individual in the middle of concentric circles symbolizing the increasing levels of the environment surrounding that individual. The mutually influencing effects of genes and the environment and their interactions produce outcomes for the developing person. An extension of this model developed by Sameroff (2010) includes macrolevel influences that constrain these levels
(Figure 9-1). It is a biopsychosocial model, because individuals have both a biology and a psychology that help create the person. Changes in the outer circles can have effects on lower-order systems, and these effects may be heterogeneous across individuals with different biological or psychological processes or starting states.
As an individual develops over time, the elements of the model expand through the integrated effects of genes and the environment (Figure 9-2). Over time, the individual interacts with and influences more of the environment, producing the self embedded within the large contextual framework. The advantage of this model, said Morris, is that a change in one aspect can “reverberate” throughout the system. “By being interactive and integrative, we could change one thing and see ramifications,” which allows such models to inform interventions.
Key measurements that are needed to inform such models are physiological as well as behavioral health indicators, contextual features of households and neighborhoods, and policy variables.
Conditional Cash Transfer Programs
One intervention that Morris discussed extensively is the use of conditional cash transfers in antipoverty programs. These transfers have two major goals. First, they reduce current poverty by transferring cash to low-income families in return for behavioral changes. Second, they reduce long-term poverty by tying those transfers to parental investments in their children through actions such as adherence to well-baby visits, immunizations, nutritional regimens, and school attendance.
The idea has been pioneered in Mexico and is now operating in about 30 lower- and middle-income countries, and “there is some evidence of its effects in improving health,” said Morris. These programs also have been controversial, and Morris encouraged workshop participants to discuss such strategies during the discussion period.
New York City is testing a version of a conditional cash transfer program called Opportunity New York City/Family Rewards. Developed by MDRC and sponsored by the New York City Center for Economic Opportunity, the intervention is layered on top of the existing social safety net. Low-income families in New York get access to all of the benefits that they would otherwise receive, along with additional benefits. Participating
families can receive $4,000 to $6,000 per year in the form of rewards for certain behavioral changes. They might receive $50 a month for their children going to school regularly, $25 for getting a library card, and several hundred dollars for a child receiving a high score on a standardized test. For older children, the rewards are split between the child and the family. The program offers rewards in three domains: children’s education, family preventive health care, and parents’ work and training.
Low-income families in six communities in New York City with a fourth, a seventh, or a ninth grader have been randomly assigned to experimental or control groups, and results from a 5-year evaluation thus far cover the first 2 years, including startup (Riccio et al., 2010). According to Morris, the program is having substantial effects on low-income families. Families in the program have seen an 11 percentage point reduction in being below the poverty level compared with control families (from 70.0 percent to 58.9 percent), and they have experienced a 13 percentage point reduction in being below half of the poverty level (from 30.0 percent to 16.7 percent).
In addition, the program has had “remarkable effects” on food insecurity, according to Morris. Control families reported a 22.1 percent rate of food insufficiency, while families in the program had a 14.8 percent rate. “Here you have a policy that is actually changing food insecurity for families.”
As mentioned earlier by Sonya Jones, interventions can have a wide variety of effects. For example, conditional cash transfers can have the effect of reducing work efforts. However, that has not been the case with the New York program, Morris observed. The program has produced greater use of healthcare services, particularly dental care, and it has produced small benefits for the academic achievement of older children who are already among the most proficient students. A question Morris posed to the workshop is how such programs could be used to address health and obesity issues specifically.
Incentives may not be enough to change some behaviors, Morris acknowledged. Low-income families face many sources of stress. It can be very complicated for such families to figure out how to buy better foods for their children. However, Morris agreed with Jones that behaviors can be nudged. “Can we … change the defaults or reduce the psychic pressures for low-income families in a way to make a big difference for behavior?” For example, is it possible to redesign supermarkets or food assistance programs to change purchasing or consumption patterns? As another example, low-income families are more likely to apply for financial aid when a child is preparing to enter college if financial aid forms are filled out for them in advance. “Could we imagine doing the same thing with the food stamp program?”
Finally, if volatility is a major issue, interventions may have to be targeted specifically to volatility, said Morris. Many income support policies do not try to reduce volatility, and some of them might increase income volatility. Do families need savings accounts or rainy day funds to weather income instability? Can the timing of food assistance distribution protect families from food instability? Such policy innovations also could protect families from lenders who otherwise could contribute to their poverty.
The application of models needs to be strategic. Where are the points of intervention that have synergistic effects? How can additional resources most effectively reduce poverty? How can food resources be made available to low-income families in ways that address both food insecurity and obesity? Policies should support rather than undermine health-related outcomes for low-income families and children, Morris concluded.
Moderator: Edward Frongillo, Jr.
During the group discussion period, points raised by participants included the following:
Conditional Cash Transfer Programs
Frongillo noted that conditional cash transfer programs have met with resistance wherever they have been proposed and asked how people can be persuaded that the idea is viable and worthy of testing. Morris responded that a major challenge is framing. Such programs should not be described as paying people to do what you want them to do anyway. Rather, they should be described as compensating families to overcome the barriers in their lives. For example, when middle-income families need medical care, they make an appointment and lose just a few hours from work, but when low-income families go a clinic, they can lose half a day or more of work and possibly a job if their work absences are too frequent or protracted. “The idea is to compensate families to make up for those times of challenge.”
Marion Standish of The California Endowment asked whether conditional cash transfer programs may add to the stress of families in environments where it may be very difficult or impossible to achieve certain goals. Morris acknowledged that this was a concern in establishing the Family Rewards program in New York City. In response, the program has sought to design the incentives to be large enough to be beneficial but not so large that they would create undue pressure on families. A related concern has been to ensure that the goals are worth achieving. For example, schools should not be dangerous for or unsupportive of students striving to attend them.
John Cook also observed that programs often seek to screen out people and asked how conditional cash transfers would differ from something like the Temporary Assistance for Needy Families (TANF) program. Morris pointed out that considerable stigma is associated with programs such as TANF. Such programs also have mixed goals. The predecessor to TANF was originally created for widowed women and then became a program to support people while they were not working and at the same time encouraging them to find work. Conditional cash transfer programs, in contrast, try to work outside the traditional welfare system.
In response to a question about how families receive the funds from conditional cash transfer programs, Morris said that the money goes into a bank account and can be spent without constraints by the recipient. It would be possible to designate the money for particular purposes, such as savings for college attendance or purchases of certain foods. Frongillo added that antipoverty programs need to be integrative, because changing a single aspect of the environment tends not to be effective.
Workshop participants also discussed the ways in which antipoverty and food assistance programs can contribute to income volatility and food insecurity. Programs can change eligibility requirement from year to year or even go in and out of existence. Programs also can have cutoffs where recipients suddenly lose benefits. For example, Barbara Lohse pointed out that schools can participate in the Supplemental Nutrition Assistance Program for Education (SNAP Ed) in Pennsylvania if more than 50 percent of their students receive free or reduced-price lunches, which means that some schools can participate some years and not other years. In this case, volatility occurs at the community level, not just at the individual or household level.
Ackerman, B. P., J. Kogos, E. Youngstrom, K. Schoff, and C. Izard. 1999. Family instability and the problem behaviors of children from economically disadvantaged families. Developmental Psychology 35(1):258-268.
Bania, N., and L. Leete. 2007. Income volatility and food insufficiency in U.S. low-income households, 1992-2003. Discussion Paper #1325-07. Madison, WI: Institute for Research on Poverty.
Becker, G. S. 1981. A treatise on the family. Cambridge, MA: Harvard University Press.
Cancian, M., and S. Danziger. 2009. Changing poverty and changing antipoverty policies. Discussion Paper no. 1364-09, Focus 26(2). Madison, WI: Institute for Research on Poverty.
Duncan, G. J., and J. Brooks-Gunn. 1997. The consequences of growing up poor. New York: Russell Sage.
Gennetian, L. A., N. Castells, and P. A. Morris. 2010. Meeting the basic needs of children: Does income matter? Children and Youth Services Review 32(9):1138-1148.
Jolliffe, D., and J. P. Ziliak (Eds.). 2008. Income volatility and food assistance in the United States. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.
McLoyd, V. C. 1990. The impact of economic hardship on black families and children: Psychological distress, parenting, and socioemotional development. Child Development 61(2):311-346.
Mills, G., and J. Amick. 2010. Income volatility, savings, and material hardship. Ideas 42 Working Paper. Cambridge, MA: Institute for Quantitative Social Sciences, Harvard University.
Riccio, J., N. Dechausay, D. Greenberg, C. Miller, Z. Rucks, and N. Verma. 2010. Toward reduced poverty across generations: Early findings from New York City’s conditional cash transfer program. New York: MDRC.
Sameroff, A. 2010. A unified theory of development: A dialectic integration of nature and nurture. Child Development 81(1):6-22.
Wu, L. L., and B. C. Martinson. 1993. Family structure and the risk of a premarital birth. American Sociological Review 58(2):210-232.