Order Now

Childhood Obesity

Childhood obesity rates have dramatically increased during the recent 30 years: in 1980 the rate of obese or overweight children between the ages of 6 and 11 years old was 6.5%, and in 2004 this rate increased to 18.8% (Franzini et al., 2009). Child obesity is the cause of many physical and mental problems, which can persist into adult obesity and related diseases. The issues of childhood obesity are thus highly important for nursing and health care professionals as well as development of methods and policies against childhood obesity.

The article “Influences of physical and social neighborhood environments on children’s physical activity and obesity” is devoted to the investigation of the relation between social and physical neighborhood environments and the levels of physical activity and obesity of fifth-grade students. The research hypothesis states that physical activity levels, correlating with childhood obesity, are negatively associated with the changes of physical environment, expressed as “more traffic, more physical disorder, low residential density and primarily residential neighborhood” (Franzini et al., 2009), and are positively associated with the social environment (measured as social cohesion and safety) (Franzini et al., 2009).

The researchers analyzed the factors which may contribute to childhood obesity at both individual and contextual level, and have outlined several groups of factors. At individual level, sociodemographic characteristics of the child are important, and at the contextual level factors were organized into two groups: neighborhood physical and social environment. Previous studies have focused mostly on the factors relating to the physical environment. Thus, the research by Franzini et al. (2009) is highly important for health care professionals, since it outlines critical factors causing low levels of physical activity and allows to develop policies and interventions for reducing childhood obesity.

Research data was collected as part of the Healthy Passages study Phase 1, which is a cross-sectional study of children’s health. The sample includes data collected for 650 fifth-grade students and for one of their primary caregivers (one of the parents, in most cases). Data were gathered between May and September 2003 at the University of California (Los Angeles), the University of Alabama (Birmingham) and the University of Texas Health Sciences Center (Houston) (Franzini et al., 2009). The final sample includes 205 Hispanics, 236 non-Hispanic Blacks, 157 non-Hispanic Whites and 52 representatives of other racial/ethical groups.

The size of the sample can be considered representative for a large population, and is reliable both for the population of fifth-graders and for the population of children of school age. This conclusion is based on the estimates using the formula for determining minimal sample size: , where n is the minimal size sample, N – total size of the population, d – precision level, Z – the number of standard deviation units of the sampling distribution corresponding to the desired confidence level (Jackson, 2011).

In 2003, the number of elementary school students (grades 1 through 4) constituted 15.9 million, elementary school students (grades 5 through 8) – 16.6 million and high school (grades 9 through 12) – 17.1 million (U.S. Census Bureau – School Enrollment, 2011). Thus, N is approximately 4.15 million for the population of fifth graders, and 49.6 million for the whole school students. Acceptable precision level is 0.05, and Z-factor for this precision level is 1.96 (Jackson, 2011). According to the formula for minimal sample size, for both populations n=385. Thus, the sample containing 650 participants is appropriate for the chosen population with a precision level of 0.05, or with 95% confidence.

Instruments for data collection and measurement

For the whole Healthy Passage research data collection protocols and materials were standardized and included validation procedures, field and training manuals. Field interviewers collected anthropomorphic data of the child (height, weight, BMI), child and parent also completed a computer-assisted personal interview and audio self-interview (also computer assisted, without the interviewer). Neighborhood observation of the child’s environment was performed by two trained observers. Physical activity was measured using the Youth Risk Behavior Survey, developed by the Centers for Disease Control and Prevention. Measures for neighborhood environment were based on the questionnaire of Chicago Neighborhood Community Survey, created in the context of the Project on Human Development (Franzini et al., 2009).

Reliability and validity of the measurements are mentioned in the “Measures” section. Margin errors are listed for the instruments of measuring anthropometric data (portable stadiometer and electronic digital scale), and Cronbach internal reliability coefficients are listed for all scales. The level of scale reliability varies from 0.66 to 0.86 (Franzini et al., 2009), with most scales having reliability higher than 0.76. The scale for social ties is questionable since its reliability is only 0.66.

The issues of validity of the chosen instruments were not explicitly mentioned in the article; however, it can be suggested that validity checks were performed during the selection of instruments for the Healthy Passages project. Overall, both reliability and validity measures are highly important for the research, since reliability reflects the consistency of the measurements, and validity shows the accuracy of the assessments (Jackson, 2011).

Forms of data display

The article contains two tables and two figures, which significantly help to visualize the research and illustrate the model developed by the authors. Figure 1 shows the theoretical background of the research and helps to understand the interaction of different groups of factors affecting childhood obesity. Table 1 contains descriptive statistics and reliability estimates for all factors measured during the experiment, with additional explanation of scales and scoring. Table 2 illustrates the results of the research, and contains regression coefficients for the individual and neighborhood factors listed for 8 structural equation models. Figure 2 contains the structural equation model and the relation between individual and neighborhood factors, childhood obesity status and mediating physical activity. The forms of data display chosen by the authors are very appropriate, and allow to understand the whole complex framework, instruments and models used for the research.

Statistical methods and limitations

The article contains a separate section devoted to discussing the limitations of the research. The authors mention theoretical restrictions of the model and the lack of fundamental social determinants in it. Another limitation is the unavailability of sociodemographic characteristics of census data and lack of information on amenities (Franzini et al., 2009). Here I agree with the authors that these two issues are major limitations of the research; however, I would not agree with the suggestion regarding the absence of fundamental social variables in the social model. Current variables can be significantly related to the base variables, and might reflect the real situation; however, this suggestion should be proved statistically.

Limitations of the statistical methods were not explicitly discussed. In my opinion, the selection of two latent variables such as physical and social neighborhood and the selection of indicators for these variables is a significant limitation of the study. In addition to this, the absence of relation between childhood obesity and physical environment indicate that the selection of factors and/or scales measuring physical environment might not be associated with children’s physical activity.
Importance and generalizability of the results

The results of the study show that favorable neighborhood social environment was positively associated with physical activity, which is, in its turn, negatively associated with child obesity (after controlling individual sociodemographic factors). No association was found between the physical environment and different measures of physical activity. Structural equation model developed by the authors allows to determine the child obesity status (ordinal variable) basing on neighborhood social environment, physical environment and individual sociodemographic factors. Among the most significant neighborhood factors affecting child obesity status are socialization of children, collective efficacy, exchange and safety.

If I were to conduct this study, I would reconsider the measures chosen for evaluating the latent variable for physical environment, and would also include genetic factors to the sociodemographic group.

The results of the study can be generalized, since the size of the sample and its diversity are sustainable. This research is of great importance, since it has shown that there is a strong relationship between the child’s social environment and obesity. This study and its further implications can be used to develop policies and interventions aimed at reducing childhood obesity.

Personal usefulness

This article was very useful for understanding the idea and statistical background of the structural equation modeling. Evaluation of the minimal sample size for large populations and analysis of models available for such evaluation will definitely be used for further research. The generation of estimates for latent variables basing on a set of indicators for the further use in the equation model is a valuable concept, which added to my understanding of research of complex social phenomena. Among other statistical approaches which were new for me is the creation of composite scores using standardized z scores and the use of various fit measures such as root mean square error of approximation, standardized root mean square residual and comparative fit index. Finally, the effective idea of increasing comparability of the variables using standardized regression coefficients expanded my knowledge of application of regression analysis.


Jackson, S.L. (2011). Research methods and statistics: a critical thinking approach. Belmont: Cengage Learning, 2011.
Franzini, L., Elliot, M.N., Cuccaro, P., Schuster, M., Grumbaum, J.A., Franklin. F., et al. (2009). Influences of physical and social neighborhood environments on children’s physical activity and obesity. American Journal of Public Health, 99(2), 271-278.
U.S. Census Bureau – School Enrollment. (2011). Available from http://www.census.gov/population/www/socdemo/school.html, accessed at June 29, 2011.