Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “multilevel analysis”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 181 records · Page 10Linked to original sources
PubMed2003

Changes in heavy drinking over the third decade of life as a function of collegiate fraternity and sorority involvement: a prospective, multilevel analysis.

Although affiliation with a fraternity or sorority is an important risk factor for heavy drinking, recent research indicates that this risk may be limited to the college years. Random coefficient growth modeling was used to track changes in patterns of heavy drinking over the course of 11 years as a function of gender and collegiate Greek involvement (N=318). Overall, greater cumulative exposure to the Greek system led to increased heavy drinking during the college years, particularly among men. Shortly after leaving college, heavy drinking levels dropped markedly and remained low through approximately age 30. Inclusion of peer alcohol use norms in the model reduced the influence of Greek involvement. Implications for models of heavy drinking and health risks are discussed.

Adolescent↗
PubMed1997

A multilevel analysis of school factors associated with health-related fitness.

School factors associated with health-related fitness, using the National Children and Youth Fitness Survey II (NCYFS II) data, have been examined by Pate and Ross (1987), but the hierarchical structure of the data was ignored in the data analyses. The purpose of this study was to reanalyze the NCYFS II data using the hierarchical linear model and reexamine the effects of school physical education (PE) and other factors on children's 1-mile run/walk performance. Only two of five "significant" school factors were confirmed by this study, suggesting that heightened probabilities of Type I errors might have occurred in the previous study. Two confirmed key characteristics of school PE programs that led to improvement of children's cardiovascular endurance were PE specialists and the administration of fitness tests. For a hierarchical data structure, both experimental and observational units should be considered in the data analysis. The hierarchical linear model, which not only provides more accurate individual prediction but also takes group effect into account, proved to be an appropriate analytical model in analyzing the hierarchical data.

Age Factors↗
PubMed2006

Comparing individual- and area-based socioeconomic measures for the surveillance of health disparities: A multilevel analysis of Massachusetts births, 1989-1991.

The absence of individual-level socioeconomic information in most US health surveillance data necessitates using area-based socioeconomic measures (ABSMs) to monitor health inequalities. Using the 1989-1991 birth weight data from Massachusetts, the authors compared estimates of health disparities detected with census tract- and block group-level ABSMs pertaining to poverty and education, as well as parental education, both independently and together. In separate models, adjusted for infant's sex, mother's age, and parents' race/ethnicity, worst-off categories of census tract ABSMs and parental education had a comparable birth weight deficit of approximately 70 g. Similar results were observed for low birth weight (<2,500 g), with worst-off categories of census tract ABSMs and parental education having an odds ratio of approximately 1.37 (p < 0.001). In mutually adjusted models for birth weight and low birth weight, census tract ABSMs still detected an effect estimate nearly 50% of that detected by parental education. Additionally, census tract ABSMs detected socioeconomic gradients in birth weight among births to mothers aged less than 25 years, an age group in which educational attainment is unlikely to be completed. These results suggest that aptly chosen ABSMs can be used to monitor socioeconomic inequalities in health. The risk, if any, in the absence of individual-level socioeconomic information is a conservative estimate of socioeconomic inequalities in health.

Adolescent↗
PubMed2006

Linking resident satisfaction to staff perceptions of the work environment in assisted living: A multilevel analysis.

PURPOSE: This study examines the relationship between resident satisfaction and staff perceptions of the work environment in assisted living. Staff perceptions were assessed at the facility level, using aggregate measures of staff job satisfaction, organizational commitment, and views of organizational culture. DESIGN AND METHODS: The sample included 335 residents and 298 staff members in 43 assisted living facilities. Data were collected by means of self-administered questionnaires. Two-level hierarchical linear models were used to test the hypotheses. With resident age, gender, and education controlled for, the relationships between resident satisfaction (Level 1 variable) and staff perceptions of the work environment (Level 2 variables) were assessed. RESULTS: Greater resident satisfaction in the facility was associated with higher staff job satisfaction and more positive staff views of organizational culture (e.g., greater teamwork and participation in the decision making). From resident characteristics, only education was significantly related to satisfaction. More educated residents were less satisfied with assisted living. IMPLICATIONS: These findings suggest that a good quality of work environment for the staff contributes to a better quality of care for the residents. More research is needed to examine the causal nature of this relationship.

Assisted Living Facilities↗
PubMed2006

A multilevel analysis of the relationships between selected examinee characteristics and United States Medical Licensing Examination Step 2 Clinical Knowledge performance: revisiting old findings and asking new questions.

BACKGROUND: This study examines: (1) the relationships between examinee characteristics and United States Medical Licensing Examination Step 2 Clinical Knowledge (CK) performance; (2) the effect of gender and examination timing (time per item) on the relationship between Steps 1 and 2 CK; and (3) the effect of school characteristics on the relationships between examinee characteristics and Step 2 CK performance. METHOD: A series of hierarchical linear models (examinees-nested-in-schools) predicting Step 2 CK scores was fit to the data set. The sample included 54,487 examinees from 114 U.S. Liaison Committee on Medical Education-accredited medical schools. RESULTS: Consistent with past examinee-level research, women generally outperformed men on Step 2 CK, and examinees who received more time per item generally outperformed examinees who received less time per item. Step 1 score was generally more strongly associated with Step 2 CK performance for men and for examinees who received less time per item. School-level characteristics (size, average Step 1 performance) influenced the relationship between Steps 1 and 2 CK. CONCLUSION: Both examinee-level and school-level characteristics are important for understanding Step 2 CK performance.

Female↗
PubMed2003

A multilevel analysis of the interpersonal behavior of socially anxious people.

We investigate the interpersonal behavior of socially anxious (SA) and non-socially anxious (NSA) individuals at three different levels of analysis, focusing on the dimensions of warmth and dominance. Study 1 examines self-reported general interaction styles, Study 2 explores behavior occurring within the context of a single interaction, and in Study 3 we focus on the performance of a single conversational act (a disagreement). Studies 1 and 2 adopt the framework of Interpersonal Circumplex Theory (IPC; Kiesler, 1983), which is well suited for studying trait-level and interaction-level social behaviors, while Study 3 is grounded in Politeness Theory (PT; Brown & Levinson, 1987), which can be used to analyze individual acts at the microstructural level. The potential mutual relevance of PT and IPC is also discussed.

Adolescent↗
PubMed2006

Multilevel analysis of group-randomized trials with binary outcomes.

OBJECTIVES: Many dental studies have assessed the effectiveness of community- or group-based interventions such as community water fluoridation. These cluster trials, of which group-randomized trials (GRTs) are one type, have design and analysis considerations not found in studies with randomization of treatments to individuals (randomized controlled trials, RCTs). The purpose of this paper is to review analytic methods used for the analysis of binary outcomes from cluster trials and to illustrate these concepts and analytical methods using a school-based GRT. METHODS: We examine characteristics of GRTs including intra-class correlation (ICC), their most distinctive feature, and review analytical methods for GRTs including group-level analysis, adjusted chi-square test and multivariable analysis (mixed effect models and generalized estimating equations) for correlated binary data. We consider two- and three-level modeling of data from a cross-sectional cluster design. We apply the concepts reviewed using a GRT designed to determine the effect of incentives on response rates in a school-based dental study. We compare the results of analyses using methods for correlated binary data with those from traditional methods that do not account for ICC. RESULTS: Application of traditional analytic methods to the dental GRT used as an example for this paper led to a substantial overstatement of the effectiveness of the intervention. CONCLUSIONS: Ignoring the ICC among members of the same group in the analysis of public health intervention studies can lead to erroneous conclusions where groups are the unit of assignment. Special consideration is needed in the analysis of data from these cluster trials. Randomization of treatments to groups also should receive more consideration in the design of cluster trials in dental public health.

Algorithms↗
PubMed2004

A multilevel analysis of socioeconomic (small area) differences in household food purchasing behaviour.

STUDY OBJECTIVE: To examine the association between area and individual level socioeconomic status (SES) and food purchasing behaviour. DESIGN: The sample comprised 1000 households and 50 small areas. Data were collected by face to face interview (66.4% response rate). SES was measured using a composite area index of disadvantage (mean 1026.8, SD = 95.2) and household income. Purchasing behaviour was scored as continuous indices ranging from 0 to 100 for three food types: fruits (mean 50.5, SD = 17.8), vegetables (61.8, 15.2), and grocery items (51.4, 17.6), with higher scores indicating purchasing patterns more consistent with dietary guideline recommendations. SETTING: Brisbane, Australia, 2000. PARTICIPANTS: Persons responsible for their household's food purchasing. MAIN RESULTS: Controlling for age, gender, and household income, a two standard deviation increase on the area SES measure was associated with a 2.01 unit increase on the fruit purchasing index (95% CI -0.49 to 4.50). The corresponding associations for vegetables and grocery foods were 0.60 (-1.36 to 2.56) and 0.94 (-1.35 to 3.23). Before controlling for household income, significant area level differences were found for each food, suggesting that clustering of household income within areas (a composition effect) accounted for the purchasing variability between them. CONCLUSIONS: Living in a socioeconomically advantaged area was associated with a tendency to purchase healthier food, however, the association was small in magnitude and the 95% CI for area SES included the null. Although urban areas in Brisbane are differentiated on the basis of their socioeconomic characteristics, it seems unlikely that where you live shapes your procurement of food over and above your personal characteristics.

Adult↗
PubMed2006

Acceptability of domestic violence against women in the European Union: a multilevel analysis.

STUDY OBJECTIVE: The acceptability of domestic violence against women (DVAW) plays an important part in shaping the social environment in which the victims are embedded, which in turn may contribute either to perpetuate or to reduce the levels of DVAW in our societies. This study analyses correlates of the acceptability of DVAW in the European Union (EU). DESIGN: Three level ordinal logistic regression of 13 457 people nested within 212 localities (cities), nested within 15 countries of the EU. Sampling is multistage with random probability. All interviews were face to face in people's homes. The outcome variable was acceptability of DVAW. Multiple correlates at the individual, locality, and country level were analysed. SETTING: European Union, 1999. PARTICIPANTS: National data were used of residents 15 years old and above of all member states in 1999 (n = 13 457). Average response rate was 47%, although it varied across countries (23%-73%). MAIN RESULTS: Higher levels of acceptability were reported by those who perceived DVAW as less severe and less frequent. Acceptability is higher among men who know a perpetrator and lower among men who know a victim. Victim blaming attitude is associated with higher levels of acceptability. In countries with higher gender empowerment measure the difference in acceptability among those who blame and those who do not blame the victim is greater. CONCLUSIONS: There are still widespread attitudes in the EU such as victim blaming that condone DVAW, contributing to a climate of social acceptability of DVAW. Further efforts to reduce the acceptability of DVAW are still needed.

Adolescent↗
PubMed2003

Preterm birth among African American and white women: a multilevel analysis of socioeconomic characteristics and cigarette smoking.

STUDY OBJECTIVE: Research shows that neighbourhood socioeconomic factors are associated with preterm delivery. This study examined whether cigarette smoking and individual socioeconomic factors modify the effects of neighbourhood factors on preterm delivery. DESIGN: Case-control study. SETTING: Moffit Hospital in San Francisco, California. PARTICIPANTS: 417 African American and 1244 white women, including all preterm and a random selection of term deliveries 1980-1990, excluding non-singleton pregnancies, congenital anomolies, induced deliveries, and women transported for special care. US census data from 1980 and 1990 were used to characterise the women's neighbourhoods, defined as census tracts. RESULTS: Cigarette smoking increased the risk of preterm delivery among both African American (OR=1.77, 95% confidence intervals (CI) (1.12 to 2.79)) and white women (OR=1.25, 95% CI (1.01 to 1.55)). However, cigarette smoking did not attenuate or modify the association of neighbourhood factors with preterm delivery. Among African American women, having public insurance modified the relation between neighbourhood unemployment and preterm delivery; among women without public insurance, the risk of preterm delivery was low in areas with low unemployment and high in areas with high unemployment, while among women with public insurance the risk of preterm delivery was highest at low levels of neighbourhood unemployment. CONCLUSIONS: Cigarette smoking was associated with preterm delivery, especially among African Americans. Adverse neighbourhood conditions had an influence on preterm delivery beyond that of cigarette smoking. The effects of some neighbourhood characteristics were different depending on individual socioeconomic status. Examining socioeconomic and behavioural/biological risk factors together may increase understanding of the complex causes of preterm delivery.

Black or African American↗
PubMed1998

An illustration of item homogeneity scaling and multilevel analysis techniques in the evaluation of drug prevention programs.

Many reports of successful school-based intervention programs can be criticized for their choice of a unit of analysis and for the neglect of measurement errors. This article is an illustration of how different conclusions can be reached from different choices of units of analyses and/or of different treatment of the data. This is done by a reanalysis of a well-reported data set. The data is thoroughly taken apart, using different statistical techniques. The result of the analyses shows that earlier reported effects of a normative school-based drug prevention program were not found. The subsequent search for moderator effects of the same program, such as a lowering effect on the relationship between the pre- and posttest or on the relationship between respondents' use and the use of their friends, was not successful either. It is concluded that the null hypothesis of zero effects should be retained. More successful was a search for individual characteristics that show significant relationships with respondents' alcohol use. Among them was the abuse of alcohol by adults in respondents' direct social environment and the use of friends.

Adolescent↗
PubMed2006

Effect of a ban on extracurricular sports activities by secondary school teachers on physical activity levels of adolescents: a multilevel analysis.

To study the effect of a teachers' ban on supervising sports-related extracurricular physical activities (ECAs), levels of physical activity among 979 grade 7 students (mean age=12.7 [0.5] years at baseline) were compared during and after the ban in seven schools that fully implemented the ban, and three schools that did not implement the ban fully. On average, schools offered 18.0 (SD=5.1) ECAs during a no-ban school year. Students attending full implementation schools were significantly more likely than students in nonimplementation schools to be active after the ban ended (odds ratio for being active=1.89 [95% confidence interval: 1.39, 2.58]). They also increased the number of physical activities in which they participated (coefficient=4.04; SE=1.01). Ending a teachers' ban on sports-related ECAs was associated with increased involvement in physical activity among secondary school students.

Adolescent↗
PubMed2002

Outlets, drinking and driving: a multilevel analysis of availability.

OBJECTIVE: This study examines the degree to which the physical availability of alcohol as measured by outlet densities is related to self-reported individual drinking patterns, preferred drinking location, as well as both driving after drinking (DAD) and driving while intoxicated (DWI). METHOD: Data from 7,826 drinkers were obtained from a general-population telephone survey of 1,353 zip code areas in California. Measures of individual alcohol consumption included drinking frequency, drinks per occasion and variance in quantities consumed per occasion. Preferred drinking locations included bars, restaurants and the homes of drinkers and of their friends. DAD was defined as driving a motor vehicle within 4 hours of having one or more alcoholic drinks, and DWI was defined as driving after having too much to drink and drive safely. Geographic measures of outlet densities were obtained for bars, restaurants and off-premises establishments, using zip codes as geographic units of analysis. Hierarchical linear modeling was used to relate outlet densities within and surrounding respondents' area of residence to respondents' drinking and to respondents' drinking and driving. RESULTS: Whereas restaurant densities were directly related to greater drinking frequencies and DAD, bar densities were inversely related to DAD. There were no direct effects of drinking patterns on drinking and driving. Drinking and driving was strongly related to drinking location preference (e.g., bars and restaurants) only when considered simultaneously with individual drinking patterns, particularly drinking frequency. CONCLUSIONS: Increased restaurant density is strongly related to higher rates of both self-reported driving after drinking and drinking frequency. The strongest influence on both driving after drinking and driving while intoxicated is preferred drinking location considered together with individual drinking patterns. Outlet density and preferred drinking location when considered together with individual drinking patterns support driving after drinking and thereby increase the potential for alcohol-related accidents.

Adult↗
PubMed2002

For all these reasons, I do...drink: a multilevel analysis of contextual reasons for drinking among Canadian undergraduates.

OBJECTIVE: Student drinking is largely related to the setting of the drinking occasion as well as to individual psychosocial characteristics. This article assesses the effect of the reasons for drinking on situational alcohol use above and beyond other environmental and individual factors. METHOD: The data were drawn from the Canadian Campus Survey, a national mail survey conducted in 1998 with a sample of 8,864 students in 18 universities. Each student provided information on up to five drinking occasions, resulting in 25,347 drinking occasions among 6,598 drinkers. At the individual level, this study focused on the university life experience. At the situational level, information about alcohol intake was recorded relative to why, when, where and with whom drinking occurred and the reasons for drinking. RESULTS: Our results show that the reasons for drinking explain 8.3% of the variance in individual alcohol intake per occasion at the individual level and 8.1% at the drinking occasion level. CONCLUSIONS: Reasons for drinking and the drinking setting together influence consumption. Moreover, reasons are context specific, because students drink for different reasons in different contexts. Thus, contextual motivational models may be more effective in helping one understand the various pathways to alcohol use and misuse.

Alcohol Drinking↗
PubMed2006

To what extent is intoxication associated with aggression in bars? A multilevel analysis.

OBJECTIVE: The purpose of this study was to assess the relationship between level of intoxication and the frequency and severity of aggression at the person, incident, visit, and bar level for aggressive incidents observed in bars or clubs. METHOD: Hierarchical Linear Modeling (HLM) analysis of 1025 incidents of aggression documented by trained observers during 1334 nights of observation in 118 bars and clubs in Toronto, Canada, was conducted. RESULTS: Both level of intoxication of the crowd during the visit as well as mean level of intoxication at the bar level significantly predicted frequency of aggression. There was a positive association between level of intoxication and severity of aggression at both the incident and person level except for the highest level of intoxication at the person level, where severity of aggression was less than for moderate intoxication. A person-incident level interaction between intoxication and severity of aggression was also found. CONCLUSIONS: These results suggest that prevention efforts should focus on both identifying bars that typically have more intoxicated patrons and reducing the intoxication levels of patrons across bars generally. The results also showed a strong positive relationship between level of intoxication and severity of aggression (except at the highest levels), indicating that intoxication increases risk in terms of both frequency and severity of aggression. The significant interaction between person- and visit-level intoxication suggests that greater attention needs to be paid to group dynamics in alcohol-related aggression.

Aggression↗
PubMed2007

[Socioeconomic context and perceived oral health in an adult population in Rio de Janeiro, Brazil: A multilevel analysis].

One of the main limitations of the ecological design is aggregation bias. The problem can be addressed by designs that use small contextual hierarchical units with internal homogeneity. This study aimed to compare different geographic data structures in order to explore contextual effects of socioeconomic variables on the odds of perceived poor oral health. The analyses were based on a mixed design that included data from civil servants participating in the "Pró-Saúde" censuses I and II and residing in the city of Rio de Janeiro, Brazil, in 1999 (n = 2,426); contextual data were obtained from the National Demographic Census conducted in 1991. We used the variance partition coefficient to compare different geographic hierarchical structures. In this study, the smallest geographic partition (census tract) was the best unit for contextual analysis for studying variability in perceived poor oral health. Individual socioeconomic attributes explained more of the variance in perceived oral health than did the socioeconomic contexts of residential areas.

Adult↗
PubMedDate not supplied

[Individual and ecological factors related to the geographical variability of psychological distress among small areas: a multilevel analysis in Catalonia, Spain].

BACKGROUND: To know the geographic distribution of the prevalence of psychological distress is important for mental health services planning. This study is aimed at identifying the individual factors and those related to the area of residence which may explain the geographic variability of psychological distress (by healthcare districts) in Catalonia. METHODS: The data was taken from the 1994 Catalonian Health Survey and from the 1996 Catalonia population statistics. The prevalence of psychological distress is estimated by age and sex and by healthcare district. In a two-level logic regression model, a study is made of the relationship between the individual variables (first level: health survey n = 12,455) and those of the area of residence (second level: the healthcare district, n = 46) to the geographic distribution of the prevalence of psychological distress. RESULTS: The significant variables at individual level are in men: age (45-64 years OR: 0.63 y > 64 years OR: 0.22), working status (no work OR: 1.60), number of chronic diseases (CD) (CD = 1 OR: 1.75 CD = 2 OR: 2.06 CD = 3-5 OR: 3.36 and CD > 5 OR: 8.9). In women: age (25-44 years OR: 0.63 45-64 years OR: 0.45 and > 64 years OR: 0.32), working status (no work OR: 1.30), number of chronic diseases (CD = 1 OR: 1.75 CD = 2 OR: 2.44 CD = 3-5 OR: 4.09 and CD > 5 OR: 11.14), and also the kind of parental household in women (single-parental OR: 1.42). The variables at the level of the area of residence are in migration (men OR: 1.55 and women OR: 1.68) and unemployment (men OR: 1.07 and women OR: 1.06). CONCLUSIONS: The individual factors do not suffice to explain the geographical variability of the prevalence of psychological distress, but the characteristics of the area of residence are also important.

Adolescent↗