2  Launch a Study

We developed the DICE web-app for researchers to launch a study without any coding and without hosting the oTree app yourself. This section guides you through the process step by step.

2.1 Meta Information

We ask you to provide some information about yourself and the study you intend to run. The DICE app uses these information to provide participants with contact information if any questions occur. In addition, the DICE app displays an external study name. The internal name only serves internal documentation purposes for yourself.

2.1.1 Configuration Files

When you successfully create a session, DICE automatically generates and downloads a JSON configuration file (e.g., abc123_dice_config.json). This file contains all your study settings and can be reused to quickly set up similar studies.

To load a saved configuration file:

  1. Upload a previously downloaded JSON configuration file
  2. Click “Load Configuration”
  3. All form fields will be automatically populated with the saved settings from that session
  4. You can then modify any settings as needed before creating a new session

This is useful for running follow-up studies or quickly reproducing a previous study’s exact configuration.

2.2 Participant Recruitment

DICE supports multiple recruitment strategies to fit your research needs:

2.2.1 Study Type Options

You can choose from the following recruitment platforms:

Online Recruitment Platforms:

  • Prolific — Optimized for Prolific’s participant pool with automatic parameter handling for easy data linking
  • Connect (Cloud Research) — Integration with Cloud Research’s Connect platform for access to diverse participant populations

Lab Studies:

  • Lab Study (Single-Use Links) — Run studies in a controlled lab environment where each participant receives a unique, single-use link

2.2.2 Study Slots (Participant Capacity)

You need to specify the number of participant slots available for your study. Each slot can be used by one participant. If there is attrition—that is, if a participant enters the study but leaves before finishing it—their slot in the database will be consumed and cannot be reused by other participants. Hence, you should account for expected attrition when specifying the number of study slots.

For example, if you expect a 20% dropout rate and want 100 completed responses, you should allocate approximately 125 slots.

Important considerations: - Once all slots are filled, no new participants can join the study - We enforce a maximum of 400 slots per study to maintain performance - If you need larger sample sizes, please contact us to discuss options - Alternatively, consider deploying a Digital In-Context Experiment yourself

2.2.3 Automatic Participant ID Parameter Handling

The participant ID parameter is automatically configured based on your selected recruitment platform:

  • Prolific → Uses PROLIFIC_PID parameter
  • Connect → Uses participant_label parameter
  • Lab Study → Uses participant_code parameter

You don’t need to manually configure these—they’re set automatically when you select your study type. When integrating with survey platforms like Qualtrics, create an Embedded Data field that matches your platform’s parameter name (e.g., PROLIFIC_PID, participant_label, or participant_code).

In addition to the participant ID, DICE also passes the participant’s experimental condition as a URL parameter named condition. This allows you to:

  • Create condition-specific survey questions in your survey platform
  • Track which condition each participant experienced
  • Analyze responses by experimental condition

For example, when redirecting to Qualtrics, participants will be sent to a URL like:

https://unisg.qualtrics.com/jfe/form/SV_0DnMoLpM0VxjhrM?PROLIFIC_PID=abc123&condition=A

To capture the condition in Qualtrics, create an Embedded Data field named condition, and it will be automatically populated with the participant’s assigned condition.

Important Methodological Consideration: The condition label is visible in the redirect URL. If participants notice their condition assignment in the URL, they may infer details about your experimental design and hypothesis, which could create demand effects or bias their survey responses. To mitigate this risk, we recommend using abstract or non-descriptive condition labels in your CSV file (for example, “A” and “B” instead of “control” and “treatment”, or acronyms like “FV” and “SV”). This way, even if participants see the URL parameter, they won’t understand what condition they’re in.

2.3 Stimulus Design

2.3.1 Social Media Platform

2.3.2 URL

In this section, you will provide the csv file you created to configure your stimuli as described in Chapter 1. The DICE app “ingests” these configurations if you provide raw URL to the file. This requires you to upload the csv file in a publicly available storage (we recommend GitHub as it is free and easy-to-use).

The DICE app needs access to your CSV file through what’s called a “raw” URL. A raw URL points directly to the file’s content, allowing DICE to read your configuration. Think of it as a direct link to the raw data, rather than a webpage showing the file.

2.3.2.1 What is a Raw URL?

When you share files through platforms like GitHub, the standard URL typically leads to a webpage that displays your file. However, DICE needs a special type of URL - a raw URL - that leads directly to the file’s content. A raw URL usually follows this pattern:

https://raw.githubusercontent.com/username/repository/main/filename.csv

The key difference is that a raw URL starts with raw. in the domain name. This tells the server to send the pure file content rather than a webpage displaying the file.

2.3.2.2 Getting a Raw URL

If you’re using GitHub, you can find the raw URL by navigating to your file and clicking the “Raw” button. The URL in your browser will then show the correct format that DICE needs. If you prefer using Google Drive, you can share your CSV file and get a direct link. First, share the file by setting it to “Anyone with the link can view.” Then, instead of using the standard sharing link, modify it by replacing “file/d/” with “uc?export=download&id=” and removing everything after the file ID. Your Google Drive URL should look like this:

https://drive.google.com/uc?export=download&id=YOUR_FILE_ID

2.3.2.3 Using Your Raw URL

Once you have the raw URL, simply paste it into DICE where requested. The app will automatically fetch your configuration file and use it to set up your study.

2.3.3 CSV Delimiter

Your CSV file uses a special character to separate different pieces of data (like columns). While commas are the most common delimiter, some files use other characters. DICE can auto-detect the correct delimiter, but you can also manually select the character that matches your CSV file format—for example, if your data is separated by semicolons, choose “Semicolon” from the dropdown menu.

2.3.4 Search Term

Enter the term you want participants to see in the search bar of your simulated social media interface. Using a hashtag or a username creates a more authentic social media experience, as it mimics how users typically search for content on platforms like Twitter. This search term will appear pre-filled in the interface’s search bar when participants start the study. For example, you might use #myHashtag, @username, or a specific topic like climatechange.

2.4 Participant Briefing

This section provides two rich text editors that allow you to create formatted content for your study participants.

2.4.2 Study Instructions

The second editor is where you provide instructions and context for participants before they interact with the social media feed. These instructions could explain what participants will see, what they should do, and any specific guidance about how to engage with the content. Accordingly, your study instructions might include welcome text, task instructions, expectations about timing, and any special considerations for interacting with the social media posts.

If you leave the study instructions editor empty, participants will skip directly from the consent form to the social media feed. If both editors are empty, participants will go directly to the social media content without any preliminary pages.

The content you create in both editors is stored as part of your study configuration and will be displayed exactly as you format it to all participants in your study.

2.5 Measurement

2.5.1 Survey Integration

After participants interact with the social media feed, you can redirect them to a survey. Simply provide the base URL of your survey (for example, a Qualtrics survey URL) without any additional parameters. DICE will automatically:

  • Add the participant’s ID to track responses
  • Handle the redirect process

For example, if you’re using Qualtrics, your URL might look like this: https://unisg.qualtrics.com/jfe/form/SV_0DnMoLpM0VxjhrM.

If you leave this field empty, participants will see a closing screen with your contact information instead of being redirected to a survey.

2.5.1.1 Completion Codes

Completion codes are used to verify that participants have completed your study and are eligible for compensation or credit. Different recruitment platforms provide completion codes in different ways:

  • Prolific: Generates a unique completion code that participants must submit at the end of the study to confirm completion and receive payment
  • Connect (Cloud Research): May use completion codes or require a redirect back to their platform with specific parameters
  • Lab Studies: You typically manage completion verification directly (e.g., via course credit systems or manual verification)

Best Practice for Data Integrity: The most secure approaches to handle completion codes are: 1. Code-based: Provide participants with a unique completion code at the end of your survey that they submit back to the recruitment platform 2. Redirect-based: Have your survey redirect participants back to the recruitment platform with a completion confirmation, which automatically verifies their participation

This ensures that participants who complete your DICE study and survey can properly confirm their participation without relying on self-reporting. Check your recruitment platform’s documentation for their specific completion code requirements and implementation methods.

2.5.2 Dwell Time Tracking

The dwell time tracking threshold determines when DICE considers a social media post as “viewed” by the participant. This setting is expressed as a percentage of the post’s visibility in the participant’s browser window:

  • A threshold of 75% (default) means a post must be 75% visible to be considered “in view”
  • Lower values (closer to 1%) will count posts as viewed even when barely visible
  • Higher values (closer to 100%) only count posts that are almost fully visible

Choose your threshold based on how strictly you want to measure post exposure. For most studies, the default value of 75% provides a good balance between capturing meaningful exposure while allowing for natural scrolling behavior.

2.6 Create Session

After clicking “Create Session,” you’ll see a confirmation screen that contains all essential information for running your study and collecting data.

2.6.1 Saving Your Session Information

The success screen displays your unique session code (such as h4qfjieu) and a corresponding session URL. These identifiers are crucial as they provide the only way to monitor your study’s progress and download your data later. Make sure to save both the code and URL in a secure location for future reference.

2.6.2 Prolific Integration

This section is relevant only if you’re using Prolific for participant recruitment.

The success screen provides a specially formatted URL that includes Prolific’s parameter structure. It looks similar to this:

https://ibt-hsg.herokuapp.com/join/pifararu/?participant_label={{%PROLIFIC_PID%}}&prolific_study_id={{%STUDY_ID%}}&prolific_session_id={{%SESSION_ID%}}

Steps to integrate with Prolific:

  1. Copy the entire URL from the success screen
  2. Paste it (as is) into your Prolific study details
  3. In Qualtrics, create an Embedded Data field named PROLIFIC_PID to capture the Prolific participant ID

The URL contains three important placeholders: - {%PROLIFIC_PID%} captures each participant’s unique identifier - {%STUDY_ID%} tracks your specific study - {%SESSION_ID%} records individual participant sessions

You’ll also find a field to enter your Prolific completion code. This code needs to be entered and submitted to allow participants to confirm their participation after completing your study. The system automatically handles the completion code distribution to eligible participants, ensuring proper compensation through Prolific’s system.

2.6.3 Connect (Cloud Research) Integration

This section is relevant only if you’re using Connect (Cloud Research) for participant recruitment.

After creating your session, you’ll see information about how to set up your study in Connect:

  1. Copy the session URL from the DICE success screen
  2. In Connect, set the “Collecting Connect IDs” field to use participant_label as the parameter name
  3. This ensures Connect will append ?participant_label=XXX when redirecting participants to your study
  4. In Qualtrics, create an Embedded Data field named participant_label to capture the Connect participant ID

This integration allows you to automatically track which Connect participants completed your study and link their DICE responses with their survey data.