9 Structuring projects
9.1 psych-DS
I have contributed a very small amount to the debate around psych-DS, a data standard for psychological data. ‘Standards’ are specifications for both what must be done or not done, and what may be done to provide flexibility. By having a standard, and an (automated) method of checking compliance with it
I really like the concept, but also still have to be convinced about some of the choices made in psych-DS. For the moment, I recommend partial compliance with psych-DS: there are high level principles that make for a very well structured project that I try to employ everywhere; but there are also other aspects of psych-DS that are (currently) high-effort-low-reward.
github_repository_name/
├── .gitattributes
├── .gitignore
├── code/
│ ├── analysis.html
│ ├── analysis.qmd
│ ├── processing_study_1.html
│ ├── processing_study_2.html
│ └── ...
├── data/
│ ├── processed/
│ │ ├── study_1_processed_data.csv
│ │ ├── study_1_processed_codebook.xlsx
│ │ ├── study_2_processed_data.csv
│ │ ├── study_2_processed_codebook.xlsx
│ │ └── ...
│ ├── raw/
│ │ ├── study_1_raw_behavioraltask_data.csv
│ │ ├── study_2_raw_behavioraltask_data.csv
│ │ ├── study_1_raw_codebook.xlsx
│ │ ├── study_2_raw_codebook.xlsx
│ │ ├── study_1_raw_demographics_data.csv
│ │ ├── study_2_raw_demographics_data.csv
│ │ ├── study_1_raw_selfreports_data.csv
│ │ ├── study_2_raw_selfreports_data.csv
│ │ └── ...
│ └── outputs/
│ ├── plots/
│ │ ├── plot_1_self_reports.png
│ │ ├── plot_2_behavioral_task.png
│ │ └── ...
│ ├── fitted_models/
│ │ ├── fit_model_1.rds
│ │ ├── fit_model_2.rds
│ │ └── ...
│ ├── results/
│ │ ├── cor_matrix_study_1.csv
│ │ ├── cor_matrix_study_2.csv
│ │ └── ...
│ └── ...
├── LICENSE
├── methods/
│ ├── study_1/
│ │ ├── replication_instructions.docx
│ │ ├── study_1_labjs.json
│ │ ├── study_1_measures_and_procedure.docx
│ │ └── ...
│ └── study_2/
│ ├── replication_instructions.docx
│ ├── study_2_labjs.json
│ ├── study_2_measures_and_procedure.docx
│ └── ...
├── preregistration/
│ └── preregistration.docx
├── readme.md
├── reports/
│ ├── preprint/
│ │ ├── preprint.docx
│ │ └── ...
│ ├── presentations/
│ │ ├── conference_presentation.pptx
│ │ └── ...
│ └── ...
└── tools/
├── project_creator.qmd
├── project_validator.qmd
└── ...Note that psych-DS specifies that the data dictionary/codebook must be a json file. I’m not yet convinced it is worth the effort of creating these for many users, as there are few R workflows that make use of them.
psych-DS principles
- The earliest form of data must be preserved
- Original data should never be modified
- Different versions of the data should be kept separate
- All transformations should be documented
9.3 OSF
Linking github repo to OSF