Socio-economic inequalities in women`s fruit and vegetable intakes: a multilevel study of individual, social and environmental mediators

Objective: This study employed a multilevel design to test the contribution of individual, social and environmental factors to mediating socio-economic status (SES) inequalities in fruit and vegetable consumption among women. Design: A cross-sectional survey was linked with objective environmental data. Setting: A community sample involving 45 neighbourhoods. Subjects: In total, 1347 women from 45 neighbourhoods provided survey data on their SES (highest education level), nutrition knowledge, health considerations related to food purchasing, and social support for healthy eating. These data were linked with objective environmental data on the density of supermarkets and fruit and vegetable outlets in local neighbourhoods. Results: Multilevel modelling showed that individual and social factors partly mediated, but did not completely explain, SES variations in fruit and vegetable consumption. Store density did not mediate the relationship of SES with fruit or vegetable consumption. Conclusions: Nutrition promotion interventions should focus on enhancing nutrition knowledge and health considerations underlying food purchasing in order to promote healthy eating, particularly among those who are socio-economically disadvantaged. Further investigation is required to identify additional potential mediators of SES–diet relationships, particularly at the environmental level.

Personal, social and environmental determinants of educational inequalities in walking: a multilevel study

Objective: To investigate the contribution of personal, social and environmental factors to mediating socioeconomic (educational) inequalities in women’s leisure-time walking and walking for transport.
Methods: A community sample of 1282 women provided survey data on walking for leisure and transport; educational level; enjoyment of, and self-efficacy for, walking; physical activity barriers and intentions; social support for physical activity; sporting/recreational club membership; dog ownership; and perceived environmental aesthetics and safety. These data were linked with objective environmental data on the density of public open space and walking tracks in the women’s local neighbourhood, coastal proximity and street connectivity.
Results: Multilevel modelling showed that different personal, social and environmental factors were associated with walking for leisure and walking for transport. Variables from all three domains explained (mediated) educational inequalities in leisure-time walking, including neighbourhood walking tracks; coastal proximity; friends’ social support; dog ownership; self-efficacy, enjoyment and intentions. On the other hand, few of the variables examined explained educational variations in walking for transport, exceptions being neighbourhood, coastal proximity, street connectivity and social support from family.
Conclusions: Public health initiatives aimed at promoting, and reducing educational inequalities in, leisure-time walking should incorporate a focus on environmental strategies, such as advocating for neighbourhood walking tracks, as well as personal and social factors. Further investigation is required to better understand the pathways by which education might influence walking for transport.

A multilevel trust management framework

Lack of trust in e-commerce transactions has been identified by researchers as one of the main factors that hamper e-commerce from reaching its full potential. Various trust-related supporting features for online transactions are available to improve trust management. However, most of these existing approaches have insufficient conditions to establish online trust among businesses and customers. There are many relevant factors that influence potential buyers to make decisions. In this paper, we identify several desirable properties of an ideal trust management system that existing trust management systems do not support. A multilevel trust management framework is proposed to improve the support for existing trust management in e-commerce.

Urban area disadvantage and physical activity: a multilevel study in Melbourne, Australia

Objective: To estimate variation between small areas in the levels of walking, cycling, jogging, and swimming and overall physical activity and the importance of area level socioeconomic disadvantage in predicting physical activity participation.

Methods: All census collector districts (CCDs) in the 20 innermost local government areas in metropolitan Melbourne, Australia, were identified and ranked by the percentage of low income households (<$400/week) living in the CCD. Fifty CCDs were randomly selected from the least, middle, and most disadvantaged septiles of the ranked CCDs and 2349 residents (58.7% participation rate) participated in a cross sectional postal survey about physical activity. Multilevel logistic regression (adjusted for extrabinomial variation) was used to estimate area level variation in walking, cycling, jogging, and swimming and in overall physical activity participation, and the importance of area level socioeconomic disadvantage in predicting physical activity participation.

Results: There were significant variations between CCDs in all activities and in overall physical participation in age and sex adjusted models; however, after adjustment for individual SES (income, occupation, education) and area level socioeconomic disadvantage, significant differences remained only for walking (p = 0.004), cycling (p = 0.003), and swimming (p = 0.024). Living in the most socioeconomically disadvantaged areas was associated with a decreased likelihood of jogging and of having overall physical activity levels that were sufficiently active for health; these effects remained after adjustment for individual socioeconomic status (sufficiently active: OR 0.70, 95% CI 0.55 to 0.90 and jogging: OR = 0.69, 95% CI 0.51 to 0.94).

Conclusion: These research findings support the need to focus on improving local environments to increase physical activity participation.

Abbreviations: SES, socioeconomic status; CCD, census collector district

Weight and place: a multilevel cross-sectional survey of area-level social disadvantage and overweight/obesity in Australia

Objective: To estimate variation between small areas in adult body mass index (BMI), and assess the importance of area level socioeconomic disadvantage in predicting BMI.

Methods: We identified all census collector districts (CCDs) in the 20 innermost Local Government Areas in metropolitan Melbourne, Australia, and ranked them by the percentage of low income households (<$400/week). In all, 50 CCDs were randomly selected from the least, middle and most disadvantaged septiles of the ranked list and 4913 residents (61.4% participation rate) completed one of two surveys. Multilevel linear regression was used to estimate area level variance in BMI and the importance of area level socioeconomic disadvantage in predicting BMI.

Results: There were significant variations in BMI between CCDs for women, even after adjustment for individual and area SES (P=0.012); significant area variation was not found for men. Living in the most versus least disadvantaged areas was associated with an average difference in BMI of 1.08 kg/m² (95% CI: 0.48–1.68 kg/m²) for women, and of 0.93 kg/m² (95% CI: 0.32–1.55 kg/m²) for men. Living in the mid versus least disadvantaged areas were associated with an average difference in BMI of 0.67 kg/m² (95% CI: 0.09–1.26 kg/m²) for women, and 0.43 kg/m² for men (95% CI: -0.16–1.01).

Conclusion: These findings suggest that area disadvantage is an important predictor of adult BMI, and support the need to focus on improving local environments to reduce socioeconomic inequalities in overweight and obesity.

Fast food purchasing and access to fast food restaurants : a multilevel analysis of VicLANES

Background : While previous research on fast food access and purchasing has not found evidence of an association, these studies have had methodological problems including aggregation error, lack of specificity between the exposures and outcomes, and lack of adjustment for potential confounding. In this paper we attempt to address these methodological problems using data from the Victorian Lifestyle and Neighbourhood Environments Study (VicLANES) – a cross-sectional multilevel study conducted within metropolitan Melbourne, Australia in 2003.
Methods : The VicLANES data used in this analysis included 2547 participants from 49 census collector districts in metropolitan Melbourne, Australia. The outcome of interest was the total frequency of fast food purchased for consumption at home within the previous month (never, monthly and weekly) from five major fast food chains (Red Rooster, McDonalds, Kentucky Fried Chicken, Hungry Jacks and Pizza Hut). Three measures of fast food access were created: density and variety, defined as the number of fast food restaurants and the number of different fast food chains within 3 kilometres of road network distance respectively, and proximity defined as the road network distance to the closest fast food restaurant. Multilevel multinomial models were used to estimate the associations between fast food restaurant access and purchasing with never purchased as the reference category. Models were adjusted for confounders including determinants of demand (attitudes and tastes that influence food purchasing decisions) as well as individual and area socio-economic characteristics.
Results : Purchasing fast food on a monthly basis was related to the variety of fast food restaurants (odds ratio 1.13; 95% confidence interval 1.02 – 1.25) after adjusting for individual and area characteristics. Density and proximity were not found to be significant predictors of fast food purchasing after adjustment for individual socio-economic predictors.
Conclusion : Although we found an independent association between fast food purchasing and access to a wider variety of fast food restaurant, density and proximity were not significant predictors. The methods used in our study are an advance on previous analyses.

Multilevel model of trade union membership

Presents a new theory of why people join or leave trade unions. Shows that a major influence on this decision is the quality of social relationships in the workplace. In particular, workplaces with a workforce of diverse occupational groups are not conducive to trade unionism.

The influence of parents, siblings and peers on pre- and early-teen smoking : a multilevel model

Introduction and Aims. Despite considerable success in tobacco control, many teenagers in Australia and other industrialised countries still smoke tobacco. There is mixed evidence on the relative influence of proximal social networks (parents/siblings/peers) on pre- and early-teen smoking, and no research has examined how these influences compare after accounting for school- and community-level effects.The aim of this study was to compare the relative influences of parents, siblings and peers, after accounting for school- and community-level variation in smoking.

Design and Methods. A cross-sectional fixed and random effects model of smoking prevalence was used, with individuals (n = 7314) nested within schools (n = 231) nested within communities (n = 30). Grade 6 and 8 students (modal ages 11 and 13 years) completed an on-line survey. Key variables included parent/sibling/peer use. Controls included alcohol involvement, sensation seeking, pro-social beliefs, laws/norms about substance use and school commitment.

Results.There was significant variation in smoking at both the school and community levels, supporting the need for a multilevel model. Individual-level predictors accounted for much of the variance at higher levels. The strongest effects were for number of friends who smoke, sibling smoking and alcohol involvement. Smaller significant effects were found for parent smoking. At the community level, socioeconomic disadvantage was significant, but community-level variance in pro-social and drug-related laws/norms was not related to smoking.

Discussion and Conclusions. Cross-level interactions were generally non-significant. Early teenage smoking was best explained by sibling and peer smoking, and individual risks largely accounted for the substantial variation observed across schools and communities. In terms of future tobacco control, findings point to the utility of targeting families in disadvantaged communities.[Kelly AB, O'Flaherty M, Connor JP, Homel R, Toumbourou JW, Patton GC, Williams J. The influence of parents, siblings and peers on pre- and early-teen smoking: A multilevel model.

Reduced food access due to a lack of money, inability to lift and lack of access to a car for food shopping : a multilevel study in Melbourne, Victoria

Objective: To describe associations between demographic and individual and arealevel socio-economic variables and restricted household food access due to lack of money, inability to lift groceries and lack of access to a car to do food shopping.
Design: Multilevel study of three measures of restricted food access, i.e. running out of money to buy food, inability to lift groceries and lack of access to a car for food shopping. Multilevel logistic regression was conducted to examine the risk of each of these outcomes according to demographic and socio-economic variables.
Setting: Random selection of households from fifty small areas in Melbourne, Australia, in 2003.
Subjects: The main food shoppers in each household (n 2564).
Results: A lack of money was significantly more likely among the young and in households with single adults. Difficultly lifting was more likely among the elderly and those born overseas. The youngest and highest age groups both reported reduced car access, as did those born overseas and single-adult households. All three factors were most likely among those with a lower individual or household socio-economic position. Increased levels of area disadvantage were independently associated with difficultly lifting and reduced car access.
Conclusions: In Melbourne, households with lower individual socio-economic position and area disadvantage have restricted access to food because of a lack of money and/or having physical limitations due difficulty lifting or lack of access to a car for food shopping. Further research is required to explore the relationship between physical restrictions and food access.

AdaBoost.MRF: boosted Markov random forests and application to multilevel activity recognition

Activity recognition is an important issue in building intelligent monitoring systems. We address the recognition of multilevel activities in this paper via a conditional Markov random field (MRF), known as the dynamic conditional random field (DCRF). Parameter estimation in general MRFs using maximum likelihood is known to be computationally challenging (except for extreme cases), and thus we propose an efficient boosting-based algorithm AdaBoost.MRF for this task. Distinct from most existing work, our algorithm can handle hidden variables (missing labels) and is particularly attractive for smarthouse domains where reliable labels are often sparsely observed. Furthermore, our method works exclusively on trees and thus is guaranteed to converge. We apply the AdaBoost.MRF algorithm to a home video surveillance application and demonstrate its efficacy.

Neighborhood disadvantage, individual-level socioeconomic position, and self-reported chronic arthritis : a cross-sectional multilevel study

Objective To examine the association between individual- and neighborhood-level disadvantage and self-reported arthritis.

Methods We used data from a population-based cross-sectional study conducted in 2007 among 10,757 men and women ages 40–65 years, selected from 200 neighborhoods in Brisbane, Queensland, Australia using a stratified 2-stage cluster design. Data were collected using a mail survey (68.5% response). Neighborhood disadvantage was measured using a census-based composite index, and individual disadvantage was measured using self-reported education, household income, and occupation. Arthritis was indicated by self-report. Data were analyzed using multilevel modeling.

Results The overall rate of self-reported arthritis was 23% (95% confidence interval [95% CI] 22–24). After adjustment for sociodemographic factors, arthritis prevalence was greatest for women (odds ratio [OR] 1.5, 95% CI 1.4–1.7) and in those ages 60–65 years (OR 4.4, 95% CI 3.7–5.2), those with a diploma/associate diploma (OR 1.3, 95% CI 1.1–1.6), those who were permanently unable to work (OR 4.0, 95% CI 3.1–5.3), and those with a household income <$25,999 (OR 2.1, 95% CI 1.7–2.6). Independent of individual-level factors, residents of the most disadvantaged neighborhoods were 42% (OR 1.4, 95% CI 1.2–1.7) more likely than those in the least disadvantaged neighborhoods to self-report arthritis. Cross-level interactions between neighborhood disadvantage and education, occupation, and household income were not significant.

Conclusion Arthritis prevalence is greater in more socially disadvantaged neighborhoods. These are the first multilevel data to examine the relationship between individual- and neighborhood-level disadvantage upon arthritis and have important implications for policy, health promotion, and other intervention strategies designed to reduce the rates of arthritis, indicating that intervention efforts may need to focus on both people and places.

State of multilevel research perspective in the information systems discipline

Existing distinctions among macro and micro approaches have been jeopardising the advances of Information Systems (IS) research. Both approaches have been criticized for explaining one level while neglecting the other; thereby, the current situation necessitates the application of multilevel research for revealing the deficiencies. Instead of studying single level (macro or micro), multilevel research entails more than one level of conceptualization and analysis, simultaneously. As the notion of multilevel is borrowed from reference disciplines, there tends to be confusions and inconsistencies within the IS discipline, which hinders the adoption of multilevel research. This paper speaks for the potential value of multilevel research, by investigating the current application status of multilevel research within the IS domain. A content analysis of multilevel research articles from major IS conferences and journals is presented. Analysis results suggest that IS scholars have applied multilevel research to produce high quality work ranging from a variety of topics. However, researchers have not yet been consistently defining “multilevel”, leading to idiosyncratic meanings of multilevel research, most often, in authors’ own interpretations. We argue that a rigorous definition of “multilevel research” needs to be explicated for consistencies in research community.