Alfred3 is a package for Python 3 offering an easy way to create computer experiments that meet the highest standards of Open Science. Specifically, experiments created with alfred3 are transparent, accessible, reproducible, and adhere to the FAIR principles for research software. In its core version, alfred3 comes well-equipped for the creation of dynamic content that can be delivered online via a webserver or offline running on local machines. In addition, the alfred3-interact plugin enables users to create interactive group experiments with features such as automated group forming, quick access to members' experiment data, and a prepacked chat functionality.
Further advantages include:
- All alfred3 scripts are written in Python 3, a very popular open source programming language that is easy to learn and fast to develop with, as it focuses on code readability. Thus, even minimal programming skills are sufficient to create experiments with alfred3 (see the requirement section for more details and suggestions on beginner tutorials).
- Alfred3 uses the principle of Object-oriented programming (OOP) to maximize code reusability. By simply copying and pasting elements between scripts, users can instantly integrate content from previous experiments with their current project.
- Experimenters can share experiments created with alfred3 the same way they share code from their data analysis. In addition to highly reusable code, transparency and ease of sharing are two key advantages of using script-based experimental software.
- Using open source software is a core principle of Open Science and both Python 3 and alfred3 are published under permissive open source licenses. Researchers will not face the hurdles associated with closed source software when trying to reproduce an alfred3 experiment (see The Open Science Training Handbook for more information on the importance of using open source research software for your experiments and data analyses).
- Online experiments written in alfred3 support all types of mobile devices through a responsive interface, making them suitable for a wide range of applications and settings (e.g., laboratory experiments where users are invited to bring their own devices, or surveying passers-py with tablets in public spaces)
- Alfred3 is optimized for the collection of personal data in full compliance with both the GDPR and official German guidelines on data management in psychological science (English version | German version). The core version of alfred3 already includes data encryption and decryption methods as well as unlinked storage options for personal or sensitive data (meaning that you can store personal data separately without the possibility of linking it back to experimental data).
If you have Python 3.7 or newer installed, just install alfred3 via pip
$ pip3 install alfred3
Documentation and tutorials for alfred3's most important features is available here: Link to docs
We use GitHub Discussions. You can ask questions, share ideas, and showcase your work there. Do not hesitate to ask!
Creating a hello-world experiment is as easy as writing this script.py
file. You can even do it in a simple text editor. Note that the file
must be named script.py
import alfred3 as al
exp = al.Experiment()
exp += al.Page("Hello, world!", name="hello_world")To run the script, open a terminal and change the working directory to your experiment directory:
$ cd path/to/experiment
Next, simply execute the following command in the terminal::
$ alfred3 run
If you have Google Chrome installed on your machine, a browser window
with the experiment opens automatically. Otherwise, open any webbrowser
and visit http://127.0.0.1:5000/start to start the experiment.
Of course, this "Hello, world" experiment does not contain much content: Only a single page with a heading. To learn how to add content to an experiment, visit our tutorials in the alfred3 documentation.
If you are publishing research conducted using alfred3, the following citation is required:
Treffenstaedt, C., Brachem, J., & Wiemann, P. (2021). Alfred3 - A library for rapid experiment development (Version x.x.x). Göttingen, Germany: https://doi.org/10.5281/zenodo.1437219
If you want to use alfred3 and need more information, don't hesitate to contact us via alfred@psych.uni-goettingen.de.
If you want to stay up to date with current developments, you can join our mailing list. We use this list to announce new releases and spread important information concerning the use of Alfred. You can expect to receive at most one mail per month.