🎲 Digital In-Context Experiments (DICE)

DICE (formerly called oFeeds) offers an assortment of oTree (Chen, Schonger, and Wickens 2016) applications designed to simulate news, shopping, and social media feeds. This suite of tools is tailored for marketing researchers, enabling them to develop more realistic stimuli for their studies.

A key feature of DICE is its user-friendly design, which allows researchers to effortlessly manipulate a feed without any coding requirement. Functioning similarly to content management systems, DICE is structured so that alterations to a single *.csv file are sufficient to modify the content and appearance of the feed. This streamlined process facilitates quick and efficient changes, enhancing the flexibility and ease of use.

As of today, this page focuses on DICE's capabilities for mimicking social media feeds. It provides guidance on how to utilize the tool for creating and customizing dynamic, interactive social media environments for research purposes.

Features

• Stable and Fully Randomized Group Assignment: Each participant in a study is assigned to an experimental condition through a randomization process. This contrasts with the non-randomized approaches often found in observational and platform studies.
• Exposure to Rich, Interactive Context: Unlike static vignette studies, participants engage with a dynamic and immersive environment. Additionally, unlike observational and platform studies, researchers maintain full control over the context of the study.
• Advanced Metrics for Researchers: DICE provides novel metrics, including latent measures like per-item dwell time. These metrics can be integrated with direct responses such as likes or comments, and supplemented with post-experiment survey data, offering a more comprehensive analysis than traditional vignette, observational, and platform studies.
• Facilitation of Replicability: The researcher exercises complete control over the content and its delivery, ensuring thorough documentation. This feature simplifies the process of replicating stimuli presented in the study, a notable advantage over platform studies where such control and documentation may be limited.

Usage

1. Create and upload a *.csv containing all the content you want to display as well as information about potential experimental conditions. You can download a template here.
2. If desired, create a post-experiment survey (e.g. in Qualtrics) with embedded data (to accept user IDs and completion codes as URL parameters).
3. Visit the DICE App and provide the required information (such as a link to the *.csv file you uploaded and the Qualtrics survey).
4. The DICE App provides URLs for the researcher to monitor the study and to distribute to prospective participants (e.g. via Prolific).

Prerequisites

1. Prolific: DICE is optimized for Prolific. Hence, if you intend to run an online experiment with DICE, you should recruit participants via Prolific.
2. Qualtrics: Even though you can also use other tools or program a questionnaire directly in oTree, we recommend to use Qualtrics to append a questionnaire to the DICE stimulus.
3. Github: To display images or gifs in your feed, you need to provide URLs (within your *.csv file) directing to publicly available images. To create URLs for your own visuals as well as the above-mentioned *.csv file, you need to upload these files somewhere. Github is a versatile, transparent and easy-to-use platform to do so. You can create an account in a matter of a few clicks.

Resources & Links

• OSF
• DICE App Repository
• Code licensed under MIT License with added citation requirement:
• Please cite as: Roggenkamp, H. C., Boegershausen, J., & Hildebrand, C. (2024, January 5). Digital In-Context Experiments (DICE). Retrieved from osf.io/jcxvk
• Shiny app to encode DICE's input *.csv file.
• Shiny app to pre-process DICE's data data.