## An important note about data inclusion. In the preregistration, we specified that our inclusion criteria -- how we chose to include participants in these analyses -- was that participants must have purpose-in-life data at at least one of two time points (2002 or 2014 or both), have work satisfaction at least once during that time, and must have scores for burnout and work engagement concurrent with the measure of work satisfaction. We have added a new inclusion criteria post-registration: participants must have item-level data for purpose in life in 2002. Upon analyzing the data, we became aware that item-level data for 2014 was not available. In addition, we were unable to locate the 2002 item-level data for 101 participants. This was especially concerning given the unusually high reliabilities for this scale. To examine the robustness of our results, we ran our analytic models with and without the 101 participants missing item data. Results did not change, and coefficient estimates were nearly always identical. We include the results of the analyses with these 101 participants included on the OSF page, with the [full results][1] and [descriptives][2] labeled full sample. In the manuscript, we have excluded these participants in order to report the most conservative test possible. [1]: https://osf.io/zdktw/ [2]: https://osf.io/ren7g/