DOI: 10.1249/tjx.0000000000000395 ISSN: 2379-2868

Gender and Age Differences in Accuracy of Predicted Half-Marathon and Marathon Finish Times

Bridget A. McFadden, Leno Benny, Gregory Klimaytis
Show PDF Cite

Introduction:

Predicted finish time (PT) reflects a runner’s expected race completion time. This study evaluated the accuracy of PT across gender, age, and performance level in half-marathons (13.1 miles) and marathons (26.2 miles).

Methods:

Prediction accuracy (PA) was determined as PT − finish times (FTs); relative prediction accuracy (REL) was determined as log(PT/FT). Performance level was categorized by FT (<20th, 20–60th, and >60th percentile). Linear regression models assessed gender, age, performance level, and interaction effects on PA and REL ( α = 0.05).

Results:

Data from 960 half-marathoners and 1496 marathoners were analyzed. Half-marathon models explained a small portion of the variance in PA ( R 2 = 0.071) and REL ( R 2 = 0.088; P < 0.001). Performance affected PA and REL, with the >60th percentile demonstrating greater underprediction than <20th percentile (PA: b = −6.73, P = 0.03; REL: b = −0.065, P = 0.002). Age was a predictor of PA ( b = −0.26, P < 0.001) and REL ( b = −0.002, P < 0.001). No gender or interaction effects were observed ( P > 0.05). The marathon models explained a modest portion of the variance in PA ( R 2 = 0.178) and REL ( R 2 = 0.161, P < 0.001). Performance affected PA and REL, with the >60th percentile demonstrating greater underprediction than <20th percentile (PA: b = −24.0, P < 0.001; REL: b = −0.09, P < 0.001). A main effect of gender was found. Women demonstrated more accurate PA ( b = 7.00, P < 0.001) and REL ( b = 0.03, P < 0.001) compared with men. Age was a predictor of PA ( b = −0.21; P = 0.003) and REL ( b = −0.001; P = 0.003). No interaction effects were found ( P > 0.05).

Conclusion:

Performance level influenced PA, with runners finishing >60th percentile showing the lowest PA. PA differed by gender in marathon runners only, with women demonstrating greater accuracy than men. Increasing age was consistently associated with underestimation of FT. Age-specific considerations are relevant when developing strategies to improve performance prediction.

More from our Archive