Human Interaction
Research Lab
The LearningByDoingXR lab is an innovative research system that enables the creation of AI Avatars to conduct repeatable and reproducible research studies.
Stimolo Persona
The persona stimulus is a structured and intentional representation of a hypothetical individual, designed to activate processes of interpretation, projection, and decision-making in research participants.
Unlike a simple demographic description, a persona is conceived as a narrative and cognitive stimulus: it integrates behavioral, motivational, and contextual elements that enable participants to “engage” with the presented case.
From a methodological perspective, the persona stimulus serves three main functions:
- Activation: it facilitates the emergence of mental schemas, biases, and implicit decision-making patterns, making otherwise latent processes observable.
- Standardization: it provides a shared reference point that allows responses to be compared across participants while keeping the initial context constant.
- Controlled projection: it enables participants to express judgments, preferences, or interpretations indirectly, reducing the effects of social desirability and self-censorship.
Designing a persona stimulus requires a balance between realism and functionality: it must be sufficiently plausible to be perceived as credible, yet sufficiently focused to activate the specific dimensions relevant to the research.
In this sense, the persona is not a faithful simulation of reality, but an epistemic tool, intentionally crafted to make cognitive, relational, or decision-making dynamics observable in a systematic and replicable way.
Research Studies
A research study within this framework is designed as a structured environment in which participants are exposed to one or more stimuli in order to observe, compare, and interpret their cognitive, interpretative, or decision-making responses.
The study is built to ensure consistency and comparability across participants, while allowing controlled variation in the elements under investigation. Each participant interacts with predefined stimuli under the same conditions, enabling the systematic collection of data that can be analyzed both qualitatively and quantitatively.
A key feature of this approach is the possibility to introduce multiple stimuli within the same study design, enabling comparative analysis. In particular, researchers can implement A/B testing configurations, where two alternative stimuli (e.g., two persona stimuli, scenarios, or narratives) are presented to different participant groups.
This structure allows researchers to:
- Isolate the effect of specific variables by modifying only selected elements between stimuli
- Compare response patterns across groups exposed to different conditions
- Test hypotheses about how variations in framing, context, or content influence interpretation and decision-making
- Increase internal validity by controlling for extraneous factors while varying key dimensions
Participants are typically assigned to conditions in a controlled manner (e.g., randomization), ensuring that differences in outcomes can be more reliably attributed to the stimulus rather than to participant characteristics.
Importantly, the study design remains flexible: researchers may choose between single-stimulus exploration (to deepen understanding of a specific construct) or multi-stimulus comparison (to evaluate differences and causal effects).
In this sense, the research study is not only a data collection procedure, but a controlled experimental setting, designed to systematically explore how specific inputs generate distinct cognitive and behavioral outputs.
Conversational Data
The framework allows researchers to access the raw conversational data generated during the study, providing full visibility into participant responses beyond aggregated outputs or summary metrics.
Raw data includes the complete sequence of interactions, preserving the original language, structure, and progression of each response. This level of detail enables researchers to conduct in-depth analyses that go beyond predefined indicators, supporting both exploratory and hypothesis-driven approaches.
Access to raw conversational data enables:
- Qualitative analysis: examining language use, framing, narratives, and implicit meanings
- Discourse and thematic analysis: identifying patterns, recurring themes, and interpretative structures across participants
- Custom quantitative processing: applying coding schemes, annotation frameworks, or NLP techniques tailored to specific research needs
- Auditability and transparency: ensuring that findings can be traced back to the original data, supporting replicability and methodological rigor
Importantly, raw data access allows researchers to revisit and reinterpret responses over time, making it possible to refine analytical models, test new hypotheses, or apply alternative theoretical lenses without needing to rerun the study.
In this sense, the availability of raw conversational data transforms the study from a fixed-output process into a flexible analytical resource, enabling deeper insight generation and long-term research value.
Realistic Expression Avatars
Avatars are designed to display facial expressions, adding a non-verbal layer to the interaction.
These expressions provide additional cues that help interpret tone, reactions, and engagement, enriching the overall understanding of participant responses.
Research Project Proposal
This framework can be adopted as a foundation for research project proposals, including master’s theses, PhD dissertations, and applied research initiatives. It provides a structured yet flexible environment for designing, implementing, and analyzing studies focused on cognitive, interpretative, and decision-making processes.
Researchers can leverage the system to formulate and test hypotheses, define experimental conditions, and collect both qualitative and quantitative data in a controlled and replicable way. The modular nature of the framework allows it to be adapted to different research questions, disciplines, and methodological approaches.
For doctoral and advanced research projects, it offers particular value by enabling:
- Rigorous experimental design, including multi-condition studies (e.g., A/B testing)
- Access to rich, high-resolution data, including raw conversational outputs
- Integration with analytical methods, from qualitative coding to computational analysis
- Iterative research development, supporting refinement of hypotheses and models over time
Moreover, the framework supports the development of original research contributions, making it suitable for academic contexts where methodological innovation, reproducibility, and theoretical grounding are essential.
In this sense, it serves not only as a research tool, but as a platform for designing and advancing structured inquiry, bridging experimental rigor with exploratory depth.
We invite researchers, doctoral candidates, and students to submit their research project proposals and explore how this framework can support their study design and analysis.