The authors present a method for estimating the ideology of political YouTube videos using cross-platform links between Reddit and YouTube. They start by applying correspondence analysis to a matrix of political Reddit posts linking to YouTube videos, placing these videos in an ideological space. This is followed by training a language model with these estimated ideologies as labels to predict the ideologies of videos not posted on Reddit. The predicted ideologies are validated against human labels. The method is validated through various tests, including aggregation to the channel level, human coding, and structural topic modeling. The authors demonstrate the utility of this method by analyzing the ideological homogeneity of individuals' video diets and the association between video ideology and viewer engagement. The findings suggest that individuals tend to favor ideologically congruent content and that ideologically extreme videos may receive more engagement. The method is scalable, fast, and can be easily adjusted to reflect changes in the ideological landscape.The authors present a method for estimating the ideology of political YouTube videos using cross-platform links between Reddit and YouTube. They start by applying correspondence analysis to a matrix of political Reddit posts linking to YouTube videos, placing these videos in an ideological space. This is followed by training a language model with these estimated ideologies as labels to predict the ideologies of videos not posted on Reddit. The predicted ideologies are validated against human labels. The method is validated through various tests, including aggregation to the channel level, human coding, and structural topic modeling. The authors demonstrate the utility of this method by analyzing the ideological homogeneity of individuals' video diets and the association between video ideology and viewer engagement. The findings suggest that individuals tend to favor ideologically congruent content and that ideologically extreme videos may receive more engagement. The method is scalable, fast, and can be easily adjusted to reflect changes in the ideological landscape.