Study participants were presented with a menu in one of the following four formats: high-emission defaults with or without carbon labels, and low-emission defaults, with or without emissions information. Both these tweaks to the menu led to participants making lower-emission choices.