Date of Degree


Document Type


Degree Name



Public Health


Katarzyna Wyka

Committee Members

Lorna Thorpe

Denis Nash

Joseph Egger

Subject Categories

Community Health and Preventive Medicine | Education | Epidemiology | Health and Physical Education | Public Health | Sports Sciences


Fitness, Physical Activity, School Absenteeism, Attendance, Chronic Absenteeism, Academics, Repeated Measures, Multilevel Model


Background: Only 42% of youth ages 6-11 in the United States meet the World Health Organization’s recommendation for ≥60 minutes of daily moderate to vigorous physical activity. Estimates for adolescents ages 12-19 are even lower, ranging from 8-17%. Literature suggests low levels of youth health-related fitness (fitness) may negatively impact attendance, potentially due to reduced physical and psychosocial wellness. Nationally, 10-15% of (5-7.5 million) students are chronically absent, meaning that they miss ≥10% of the school year (or ≥20 days of school per year). Moreover, 20-30% of students in high-poverty, urban school districts do not attend school regularly (≥6 days absent per year). To the author’s knowledge, this is the first multi-year study on the longitudinal fitness-attendance relationship, taking into account multilevel student and school factors.

Objectives: To describe differences in student- and school-level attendance, examine the longitudinal causal effects of change in fitness on attendance (days absent and chronic absenteeism), and assess gender effect measure modification in the fitness-attendance relationship in a large and diverse sample of New York City (NYC) middle school students.

Methods: Data were drawn from the NYC Fitnessgram. Six cohorts of NYC public middle school students were followed over 4 years during a seven-year study period (2006/7-2012/13; n=349,381). A 2-level cross-sectional Generalized Linear Mixed Model (GLMM) was used to assess the extent of school-level clustering in attendance, and determine the variation in student attendance accounted for by school-area poverty. A 3-level longitudinal adjusted GLMM, overall and stratified by gender, was used to test the effects of change in fitness (cardiorespiratory, muscular strength and endurance composite score) on attendance (1-year lagged student-specific days absent). Lastly, a 3-level logistic GLMM was used to test the effects of change in fitness on chronic absenteeism (1-year lagged student-specific chronic absenteeism, defined as ≥20 days absent per year).

Results: The Intraclass Correlation Coefficient (ICC) estimate demonstrated a large degree of variance in student attendance explained by schools (11%; p20% increase) demonstrated 0.66 fewer days absent per year (95%CI: 0.56, 0.75; p20% decrease in fitness based on change in composite percentile scores from the year prior). The difference in days absent for the most compared with least improved fitness groups was greatest in both genders for those attending schools in high- and very high- compared with mid- and low-poverty areas (0.567 vs. 0.249, and 0.522 vs. 0.275 in girls and boys, respectively). Lastly, 3-level logistic GLMM showed that a large increase in fitness (>20%) was associated with a 19% lower probability of chronic absenteeism the following year (95%CI: 17.82, 20.69), relative to students with a >20% decrease in fitness composite percentile scores (p

Conclusions: School-level clustering in attendance was sizeable and statistically significant. School-area poverty accounted for a large proportion of the variance in student attendance at the school-level. An inverse dose-response relationship was found between student change in fitness level and days absent in both genders, with slightly stronger effects in girls, and youth attending schools in high-poverty areas. Further, an inverse dose-response relationship was found between student fitness and 1-year lagged school chronic absenteeism. Study findings suggest cumulative effects of fitness improvement could have a significant impact on child attendance over time. Given over 200,000 NYC students are chronically absent each year, this work suggests that fitness interventions should be examined as a method to promote attendance at the population level.