Prior research has estimated intergenerational time preferences by asking respondents to choose between hypothetical life saving programs. From such choices, researchers have concluded that the public heavily discounts the lives of people in future generations. However, using a multiversion survey involving 401 respondents, I show that imputed intergenerational time preferences can be dramatically affected by the specific question that is asked. Different elicitation procedures can yield widely varying results by evoking or suppressing various relevant considerations (such as uncertainty). Many formats revealed no preference for current generations over future generations.