**Analysis** This folder contains the data, scripts, and output from the analyses. The knitted analysis out is [friendship-analysis.pdf][1], which is all reported values are taken from. *Data* This folder contains the [data][2] used in the analyses and a [data dictionary][3] explaining the data. *Figures* This folder contains all the figures in the main manuscript and the supplementary materials. *Models* This folder contains the compiled and computed models. *Scripts* The [data file][4] used in the analysis was produced using the [friendship cleaning data.r][5] script to clean the extracted from ODK. To protect participant's information, the original data file is not available outside the research team. Models were estimated using *brms* and *cmdstanr* package. The models specified in the [preregistration][10] are in [preregistered models.r][11]. The script for estimating the models in the paper are in the [model.r][12] file. Note that compiling and sampling the models can take significant time. Users who are interested in reproducing the analyses without running the models can download the models from the *models* folder. Code for analyzing the model and creating figures are in the [friendship analysis.r][13] file. **Other files** Included in this repository is the [preregistration][14] for the analyses and [xls file][15] used to create the survey on Open Data Kit. [1]: https://osf.io/nxj5t/ [2]: https://osf.io/jcxp7/ [3]: https://osf.io/gwbsm/ [4]: https://osf.io/jcxp7/ [5]: https://osf.io/ys8n5/ [6]: https://osf.io/g59rb/ [7]: https://osf.io/pwfjv/ [8]: https://osf.io/guw47/ [9]: https://osf.io/g59rb/ [10]: https://osf.io/97s5m/ [11]: https://osf.io/ea9wx/ [12]: https://osf.io/w7r83/ [13]: https://osf.io/r5fys/ [14]: https://osf.io/97s5m/ [15]: https://osf.io/zs4qu/