In a survey about health risk reductions, we also collect data about individual time preferences using a choice about payout options for hypothetical lottery winnings.We model individual discount rates as a function of age, other socio-demographic variables and variables to capture health expectations. For older subjects, undesirable current health behaviors are better predicted by their back-casted discount rates at age 21. Changing time preferences as people mature may lead to the development of health habits while they are young that are likely to be inconsistent with the preferences of their older future selves.
We assess whether samples of respondents drawn from large internet consumer panels are representative of the underlying population. We model the attrition/selection process for a major consumer panel maintained by Knowledge Networks, Inc (KN). Starting from KN’s over 525,000 random-digitdialed (RDD) panel-recruitment telephone contact attempts, and ending with a sample of respondents to an actual online survey, we span all junctures at which systematic selection could occur.