The railroad tank car industry is moving toward damage-tolerance-based inspection intervals to maintain the safe and efficient transport of hazardous materials by rail. One aspect of this effort is predicting fatigue crack growth rates that result from railroad tank car service conditions. A fracture mechanics parameter, H, is proposed to relate the crack-tip driving force under a railroad tank car spectrum to the resulting crack growth rates. This approach will allow the direct application of fatigue crack growth rate data resulting from laboratory tests to models for crack growth under railroad service conditions. However, for this approach to be successful, it is essential that similitude between lab tests and the structural application is maintained, so that fatigue crack growth rates are predicted correctly. Two crack-closure-based models are used to predict the effects of various laboratory test design parameters on fatigue crack growth rates for the loaded vertical coupler force (LVCF) spectrum, which describes the vertical coupler forces for a loaded railroad tank car. Random ordered load blocks are predicted to result in growth rates approximately twice as great as those resulting from either low-high or high-low ordered blocks. Truncating even the smallest amplitudes from the spectrum was predicted to affect crack growth rates. The most significant effect predicted was that of the far-field stress level. Although changes in stress level are predicted to have only a minor affect (approximately 20%) on constant-amplitude fatigue crack growth rates, corresponding changes in stress level are predicted to affect variable-amplitude fatigue crack growth rates by up to a factor of six. These results suggest that tests designed to determine fatigue crack growth rates for the LVCF spectrum must be considered carefully, and that crack closure based models for fatigue crack growth are a valuable tool for designing such tests.