• from_data_team has the Dunst datasets that were received from the data-collection team. The corresponding MB datasets are on Git and are automatically pulled in by the script 2_get_tidy_MB_data.R.

• Meta-analysis codebook: data/codebook.xlsx

• Main-analysis dataset: mb_ma_combined_prepped_0.125.csv

• Fractional suffixes on dataset names (e.g., "0.125") correspond to inclusion criteria. The "0.75" dataset (i.e., with the more stringent subject inclusion criterion) has fewer estimates than the main dataset (equivalent to the "0.125" dataset, corresponding to one of 8 trials per condition, the same as the published analyses) because some age groups were dropped completely if N<10.

To prep the main-analysis dataset and the various datasets used for sensitivity analyses, simply run 0_master_prep.R. This file sets different combinations of global variables regarding the inclusion criteria and whether the dataset should be age-matched and then calls a sequence of other data-prep files (1_get_tidy_MA_data.R, 2_get_tidy_MB_data.R, 3_merge_MA_MB_data.R, and 4_prep_for_analysis.R).

The analyses with the original Dunst, corrected Dunst, and the community-augmented ("augmented_ma_extended") datasets are set up in a parallel directory structure in both the data-prep and analysis directories. The reproducible manuscript on Overleaf has subdirectories called original_dunst, corrected_dunst, and community_meta for the results.

The analysis script conducts the main analyses as well as the sensitivity analyses. It uses only the datasets containing "combined_prepped" in the names.

• In tables_to_prettify, titles with uppercase model names represent main analyses