Re-analysis of data from the ManyBabies1: Infant-directed Speech Preference project.
- For more information about the ManyBabies project, see http://manybabies.stanford.edu/
- The data and main analysis can be found at https://github.com/manybabies/mb1-analysis-public, stimuli, protocol, and further documentation is at https://osf.io/re95x/
- This is the paper (which we ask to cite if you do anything with this dataset) The ManyBabies Consortium. (2019). "Quantifying sources of variability in infancy research using the infant-directed speech preference." In press at Advances in Methods and Practices in Psychological Science (AMPPS). Preprint
The repository contains:
-
MB1_analysis.jmd- The main script, which- Reads in ManyBabies 1 data
- Shapes it as needed
- Reproduces the main analysis from the paper at https://github.com/manybabies/mb1-analysis-public
- Fits the preregistered maximal model
- Simplifies the random effects by inspecting the output of rePCA and the variance components
- Shows what can go wrong with multi-lab data (subid vs subid_unique)
- Re-processes more messy data to apply different cleaning criteria to deal with censoring (which you can adjust)
- Re-runs the analysis
-
MB1_analysis.ipynb- The corresponding Jupyter notebook that you can run in your browser, but which differs from the converted version of the .jmd (split code blocks to see all output, converted R code cell) -
MB1_minimal_lmer.R- The R code needed to reproduce the main analysis of the paper, extracted from https://github.com/manybabies/mb1-analysis-public/blob/master/paper/mb1-paper.Rmd -
intendend_complex_LMM.txt- Output for the preregistered model (which still takes some time to fit) -
Project.tomlandManifest.toml- See https://repsychling.github.io/pkg.html
For instructions how to run code in .jmd and .ipynb files, see https://repsychling.github.io/intro.html
This work was supported by the Center for Interdisciplinary Research, Bielefeld (ZiF) Cooperation Group "Statistical models for psychological and linguistic data".