EFFECTS OF PHYSICAL AND TRAINING CHARACTERISTICS ON MARATHON PERFORMANCE
Mark A. Sarzynski, Jocelyn C. Frey, Christopher J. Womack, FACSM, and James
M. Pivarnik, FACSM
Previous research on predictors of marathon performance has utilized
laboratory testing with small sample sizes of elite runners. Few studies
have involved a large number of participants with variable running
abilities.
PURPOSE: To evaluate physical and training characteristics of
all registered runners immediately after they completed a marathon and
determine which variables predict finish time.
METHODS: An online survey
was offered to all runners who registered for the 2005 Detroit Marathon.
Nearly 40% of race participants (N=1371) responded to the survey. To
determine which variables were associated with race time, the following
variables were evaluated: gender, age, body mass index (BMI), weekly
training mileage (during the 3 months prior to the race), number of years
running, past-year running injuries, and race related physical symptoms.
After performing bivariate analyses with each potential predictor variable,
stepwise-regression was used to predict race time using variables that
showed a significant relationship with marathon performance.
RESULTS: Race
time averaged 4.2±0.7 hours (range=2.4-6.8 hours). Sixty-four percent of
respondents were male and average age was 38.5±10.4 years. BMI averaged
24.3±2.7 for males 22.0±2.7 for females. In the past year, ~37%
experienced a running injury. During the race, 32% of respondents
experienced at least one physical symptom such as dizziness or nausea. Mean
weekly training distance was 34.5±12.7 miles, and years running averaged
11.3±9.9 years. All predictor variables except for injury in the past year
and years of running were significantly related to race time. Stepwise
regression showed that weekly training mileage, age, BMI, gender, and
physical symptoms during the race accounted for 41% of the variance in race
times. The strongest predictors were weekly training miles (r2=0.18), BMI
(r2=0.07), and gender (r2=0.14) accounting for ~39% of the variance.
CONCLUSIONS: Our results showed training mileage, gender, and body size to
be major predictors of marathon race performance in a large sample of
runners with varying abilities. Training injuries, years of running
experience, and physical symptoms during the race had no apparent effect on
performance. Thus runners have control over most of the variables
contributing to improved marathon performance. Also, average runners are
affected by the same factors as elite runners; qualitatively similar,
quantitatively different.