Fatigue cracking is one of the major distresses in asphalt pavements. Accurate prediction of fatigue life of asphalt pavements can be extremely important both during the design stage and for prediction of remaining service life of in-service pavements. Traditional fatigue life predictions based on bending beam tests can be costly and time-consuming. The uniaxial push-pull (tension-compression) tests run on cylindrical samples have been a novel alternative. However, the traditional sample size for the push-pull tests may prevent its use for thin in-service pavements. This paper presents the results of a study investigating the possibility of using smaller sample sizes for push-pull tests. The viscoelastic continuum damage (VECD) characteristics of regular and small-size samples are compared, and the difference is observed to be negligible. In addition, a practical fatigue life formulation is derived on the basis of VECD theory. Uniqueness of the derived fatigue life (N sub f) equation, differing from previously derived VECD-based N sub f equations, stems from the fact that it does not force a certain form of equation to fit the damage characteristic curve. Finally, the differences in fatigue lives of different layers of the field sections at FHWA's accelerated loading facility are investigated.