This dissertation investigates the impact of different context factors such as domain, identity, and measurement on political ideology and its psychological underpinnings. Taking on an ecological perspective along the lines of Brunswik (1955, 1956), a special focus is given to representative sampling of experimental stimuli. The first project tested the ideology-symmetry-hypothesis with regard to domain-specificity of need for cognitive closure. Across three operationalizations, conservatives demonstrated more need for cognitive closure than liberals regarding a conservative domain, and vice versa, liberals exhibited more need for cognitive closure regarding a liberal domain. The second project examined the entanglement of political attitudes and identity dependent on one's political orientation. In two studies, individuals endorsed counter-attitudinal issues stronger when their personal identity was salient, and pro-attitudinal issues more strongly when their political identity was salient. The third project investigated the effect of exposure to in- or out-group fake news on political polarization. Exposure to in- and out-group fake news was associated with high affective polarization while higher levels in attitude polarization were found after exposure to out-group fake news. Informing participants about the fake news nature of the post attenuated affective polarization for those perceiving the sender to be an out-group member only. The fourth and final project aimed at providing a systematic review of political ideology measurement with a special focus on replicability and validity. Results suggested substantial heterogeneity in and oftentimes insufficient documentation of measurement practices. Taken together, the four lines of research highlight the necessity of representative sampling and context sensitivity when assessing the psychological foundations of political attitudes.