Do individuals conceal their political preferences under social pressure? While theories of social desirability bias predict self-censorship and preference falsification when under social pressure, observational evidence remains limited due to challenges in manipulating privacy or social expectations. This study addresses this gap by analyzing how privacy violations during face-to-face surveys — in the form of the presence of other household members — affect respondents’ expressed opinions on politically sensitive topics. In contrast to previous research, this study uses longitudinal panel data in the United States and the UK to control for individual-level changes over time and reduce selection effects. The findings reveal minimal evidence of preference falsification or self-censorship. Respondents largely express consistent views, even when observed by others. Where effects emerge, they are inconsistent with theories of systematic self-censorship or preference falsification. This challenges assumptions about the prevalence of self-censorship and preference falsification, suggesting that there is limited self-censorship within families. These findings highlight that the influence of social pressure depends on the nature of the social reference group and the specific social conditions in which political expression occurs.