Personality & assessment methods
Alternatives to multiple-choice personality tests
Multiple-choice questionnaires are not the only way to explore personality. Interviews, behavioural observation and language analysis draw on a different kind of material: what a person says, does or expresses in their own words.
Closed-ended questionnaire
When I work with other people, I prefer to…
…to free expression
“I enjoy working with others when I can first understand where the project is heading. I need to connect the ideas before I can take a clear position.”
01 — Moving beyond predefined answers
Changing the method means changing the evidence
A personality questionnaire presents a series of statements and asks the user to choose the answer that seems closest to them. This format is quick, standardised and easy to compare. It becomes less natural when the answer depends on context, when none of the options quite matches the person’s experience, or when they would rather explain than select.
A genuine alternative does more than replace one set of questions with another. It changes the material being examined. Instead of a predefined choice, it may use an extended response, behaviour observed in a situation, or freely produced language.
These methods do not answer exactly the same question. An interview shows how a person tells and interprets their experience. Observation shows how they act in a particular context. Language analysis examines how they select, connect and transform what they express.
02 — Three broad alternatives
Telling, acting or writing
Three broad approaches are enough to understand the essential shift: personality can be narrated, observed in action or studied through language.
Personality as narrative
Interviews, open-ended questions and personal narratives allow people to build their own response. They choose the situations they consider important, clarify the circumstances, explain their reasoning and make sense of what they experienced. The analysis then focuses on the examples selected, the connections made and the way the experience is organised into a narrative.
Personality in action
Behavioural observation starts with what a person actually does in a situation: solving a problem, working with others, making a decision, adapting or responding to a constraint. It may involve a simulation, a diary study or observations gathered across several contexts. It captures personality in action, but only within the situations that were actually observed.
Personality in language
Language analysis uses an open response, a narrative or a piece of free writing. It examines not only the topics discussed, but also how elements are selected, connected and prioritised. A text may turn an experience into meaning, facts into an explanation, a difficulty into action, or an idea into a structure. That movement is itself a source of analysable information.
Why free text is distinctive
Content matters, but so does the movement of the text
Two people can write about the same subject and produce very different texts. One may first try to understand what the experience means. Another may reconstruct the underlying mechanism. A third may focus on what needed to be done, while a fourth may organise the material around a future direction.
The information therefore does not lie only in the words themselves. It also lies in the details selected, the weight given to facts, emotions or ideas, the causal links that are created, the progression of the reasoning and the form the text ultimately takes.
Free writing therefore occupies a useful middle ground: it allows people to express themselves in their own terms while still producing stable material that can be examined using explicit criteria.
03 — What the research shows
Language contains measurable information about personality
The relationship between language and personality is empirically established: language production contains a statistically detectable personality signal, whose magnitude depends on the criterion measure, the text collected and the analytic method.1
Participants across 31 samples in the first meta-analysis of Big Five traits and LIWC language categories.
Language categories examined simultaneously, including affect, pronouns, causation and several classes of vocabulary.
Average proportion of variance explained by the full set of categories for observer-rated traits in that meta-analysis.
Quantified associations
For self-rated traits, the 2022 meta-analysis reports maximum correlations ranging from |ρ| = 0.08 to 0.14; the 52 language categories explained an average of 5.1% of the variance. For observer-rated traits, correlations reached |ρ| = 0.18 to 0.39, and the complete set of categories explained 38.5% of the variance.1
A pattern, not a revealing word
Of 260 possible associations between five traits and 52 language categories, twenty were found for both self-ratings and observer ratings. The authors describe this overlap as a “kernel of truth” in linguistic personality markers. The measure therefore rests on a pattern of regularities, not on interpreting a single word.1
The text collected determines the quality of the signal
Text length, the task used and sample characteristics all affect effect sizes. Longer texts, in particular, yield stronger associations. A 2026 study also found that targeted prompts elicit more trait-relevant narrative units than generic prompts.12
Specialised models improve validity
In the same 2026 study, convergence with target scores increased as models became more tailored and methodologically sophisticated. Fine-tuned models performed best; when trained to reproduce direct ratings of narrative responses, they achieved an average correlation of 0.83 with the target scores in the study sample.2
What these findings establish
Language can be treated as measurable psychological data. Open-ended responses and naturally occurring texts can produce scores that converge with independent measures, provided that the prompt, criteria and model are designed for the construct being assessed. Recent methodological reviews therefore describe language-based assessment as a complementary method to traditional scales, applicable both to prompted responses and to texts collected in natural settings.34
What they do not automatically validate
These findings demonstrate the general feasibility of psychological assessment through language. They do not, by themselves, validate every individual model, every typology or the identification of an MBTI type from a text. A dedicated validation study of TypeInteractions would need to examine its reliability, reproducibility and convergence with independent criteria.
Write freely
Develop an experience, a relationship, an idea or a situation that matters to you.
04 — The TypeInteractions alternative
A structured analysis based on free writing
TypeInteractions begins with a piece of freely written text. Users do not have to choose the adjectives that describe them: they develop an experience, a relationship, an idea or a situation in their own words.
Write something personally meaningful
The text should contain an experience, line of reasoning, tension, choice or relationship developed in enough depth for an underlying dynamic to emerge.
Identify the dominant transformation
The analysis combines several levels of observation and looks for the recurring way in which the text selects, connects and transforms its material.
Receive a structured interpretation
The result presents a dominant dynamic, an overall portrait, its main strengths, its points of caution and a secondary MBTI indication.
An analysis of the text’s dynamic, not a keyword search
TypeInteractions is not based on the idea that a single word reveals a personality. The method examines the overall movement of the text: what becomes salient, the kind of material it starts from, the transformation it performs and the form it ultimately produces.
The result does not claim to summarise the whole person. It describes the dynamic that appears most clearly in the submitted text. Its relevance therefore depends on the richness of that material and it should be used as a tool for personal reflection.
TypeInteractions does not provide a medical or psychological diagnosis. The MBTI indication is for guidance only and does not replace an official MBTI assessment.
Discover TypeInteractions
Your personality may not fit neatly into a box. Start by writing.
Develop an experience or reflection in your own words. TypeInteractions analyses the dynamic that organises your text and offers a structured interpretation of your personality.
Indicative, non-medical analysis. Results depend on the text submitted.Key references
- Koutsoumpis, A., Oostrom, J. K., Holtrop, D., van Breda, W., Ghassemi, S., & de Vries, R. E. (2022). The kernel of truth in text-based personality assessment: A meta-analysis of the relations between the Big Five and the Linguistic Inquiry and Word Count (LIWC). Psychological Bulletin, 148(11–12), 843–868. View DOI
- Speer, A. B., Delacruz, A. Y., Chawota, T. A., et al. (2026). Unpacking the Validity of Open-Ended Personality Assessments Using Fine-Tuned Large Language Models. Organizational Research Methods. View DOI
- Brickman, J., Gupta, M., & Oltmanns, J. R. (2025). Large Language Models for Psychological Assessment: A Comprehensive Overview. Advances in Methods and Practices in Psychological Science. View DOI
- Nilsson, A. H., Eijsbroek, V. C., Gu, Z., et al. (2026). The Language-Based Assessment Model Library: Open Model Sharing for Independent Validation and Broader Applications. Advances in Methods and Practices in Psychological Science. View DOI