Roll call scaling techniques are empirical standards for studies of voting behavior within legislative bodies. Though ideal point estimation techniques are frequently used, the theoretical implications of assumptions made in order to empirically estimate ideal points provide cause for concern. Current scaling techniques ignore the role of group-level dependencies within the data. Assumptions about independence of observations in the scaling model ignore the possibility that members of the voting body have shared incentives to vote as a group. In turn, this leads to potential biases in the estimated values of the ideal points and underestimation of the number of dimensions needed to model the ideal point space. In this paper, I propose a new ideal point model that explicitly allows for group contributions in the underlying spatial model of voting. I derive a corresponding empirical model that utilizes flexible Bayesian nonparametric priors to estimate group ideological effects in ideal points and the corresponding dimensionality of the ideal points. I apply this model to the 114th U.S. House and show how grouped ideological effects can be uncovered using only a set of roll call votes. This model provides insights into open questions related to group dynamics in legislative voting and has important implications for literature that utilizes ideal point estimates.