Analysis of symmetrically predicted endogenous subgroups (ASPES) is an approach to assessing heterogeneity in an ITT effect from a randomized experiment when an intermediate variable (one that is measured after random assignment and before outcomes) is hypothesized to be related to the ITT effect, but is only measured in one group. For example, researchers may feel that implementation quality was related to the effect of an intervention, but it was only measured within the treatment group. Research has been done on the discrete version of ASPES, which is used when the intermediate variable is discrete, but there is also a continuous version of ASPES, used when the intermediate variable is continuous, that has had very little research devoted to it. Continuous ASPES typically uses a linear model to estimate the relationship between the intermediate variable and the ITT effects, which assumes that the effect of the intermediate variable on the ITT effects is constant, which may be an unrealistic assumption in certain settings. This presentation will report results from my dissertation research that evaluated the performance of continuous ASPES across a variety of conditions in a Monte Carlo simulation study. This presentation will specifically focus on the results that provided evidence about how continuous ASPES can accommodate a quadratic relationship between the intermediate variable and the endogenous subgroup ITT effects.