This is the home of project "In search of the preference reversal zone". The aim of this project is to examine, in a precise quantitative way, the predictions of hyperbolic and exponential delay discounting models for preference reversals. Preference reversals are an important explanatory construct for understanding various behavioural problems (e.g. relapse in addiction) as well as being of theoretical interest. The files [GlautierEisenbarthMacaskillEPRevisedShort2.pdf][1] and [GlautierEisenbarthMacaskillEP_ESM2.pdf][2] are manuscript and extra supplementary material (respectively) which were accepted as a stage 1 pre-registration study in journal Experimental Psychology. The files [GlautierEisenbarthMacaskillEPResults.pdf][3] and [GlautierEisenbarthMacaskillEP_ESMResults.pdf][4] contain the finally accepted (on 17/02/2022) preprint manuscript and the extra supplementary material. The finally published version is [here][5]. The data files are in three groups. 1. Participant data 2. r-code 3. Latex files The participant data gives the data obtained from a pilot study and from the main experiment. The r-code analyses the data, it is called by [knitr][6] commands embedded in the latex (.Rnw) files. The r-code itself contains only limited documentation. However, the 'literate' programming model of knitr is designed to facilitate understanding by linking the code, data, and the human readable report. For example, starting with the k values reported in the Stage 1 summary results in GlautierEisenbarthMacaskillEPResults.pdf review the code in GlautierEisenbarthMacaskillEPResults.Rnw. From there the r-code which produced the latex in GlautierEisenbarthMacaskillEPResults.tex can be located and inspected. There are four .r files to check. generic.r contains some generic utility functions (e.g. fndp to round values for formatted printing). e178.r and ddprz.r contain code used for the pilot data, construction of figures in the introductory material, and preparatory simulation/optimisation work, whereas dataAnalysis.r contains the code for analysis of the experimental data. So to continue the example of how to link the report, data and code, in GlautierEisenbarthMacaskillEPResults.Rnw we can see that the maximum likelihood hyperbolic k value for the small reward (0.26) was obtained from an indexed r-variable meanS1Params, which is in turn produced by the r-function colMeans called on a subset of columns from dataframe s1dv. dataframe s1dv comes from r data file s1dataValues.RData. The four r code files can be searched manually or a utility such as grep can be used to find relevant strings used in the Rnw files. The c++ code used in the experiment is available on [github][7]. [1]: https://osf.io/km2jr/ "GlautierEisenbarthMacaskillEPRevisedShort2.pdf" [2]: https://osf.io/qpkhg/ "GlautierEisenbarthMacaskillEP_ESM2.pdf" [3]: https://osf.io/ckudj/ [4]: https://osf.io/xpn3u/ [5]: https://econtent.hogrefe.com/doi/full/10.1027/1618-3169/a000542 [6]: https://yihui.org/knitr/ [7]: https://github.com/stegzzz/ddprz