Methodology
SAMRUM builds on well-established psychological research frameworks, but is not a clinical test or diagnostic tool.
Scientific foundation
Internally, SAMRUM builds on two well-established research frameworks:
OCEAN (Big Five)
The most validated model for personality traits, describing five core dimensions: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism.
IPC (Interpersonal Circumplex)
A model for relational behavior based on two axes: dominance/submission and warmth/hostility. Used to understand dynamics between people.
In the app, we translate this into a family language with 10 themes from everyday life – not abstract personality types.
The 10 family themes
Reports use these themes to describe patterns. They're continuous tendencies – not categories or types. There's no "right" or "wrong" end.
Need for structure
Need for fixed routines and predictability in everyday life.
Predictability ↔ Change
Thriving on repetition versus energy from variation and new experiences.
Social energy
Energy from being together versus need for alone time and quiet.
Warmth & contact style
Need for closeness, expressing care, and tolerance for distance.
Boundaries & direction
Tendency to take charge in decisions versus following others' initiative.
Emotional reactivity
Intensity of reactions and time to calm down again.
Repair after conflict
Speed and willingness to restore after disagreement.
Fairness filter
Sensitivity to inequality and favoritism in the family.
Everyday persistence
Tendency to finish tasks and keep to agreements.
Calm under pressure
Ability to keep overview and tone when things get tense.
What we deliberately avoid
SAMRUM is designed with clear boundaries for what the tool does and doesn't do.
- ✕No diagnoses
We don't diagnose ADHD, autism, anxiety, or other conditions. That's for professionals.
- ✕Differences are patterns
Reports describe differences as patterns – not as flaws or shortcomings.
- ✕No therapeutic guidance
Reports don't provide therapy. For serious challenges, we recommend professional help.
- ✕No "right/wrong"
Our language avoids morally loaded words. Differences are differences, not faults.
- ✕No numbers or percentages
Users never see scores, percentiles, or numerical comparisons.
How we calculate and present results
The themes are designed to be neutral: neither end is "better" – both high and low scores have strengths and challenges depending on context.
Scoring
Each theme is measured with multiple questions. Reverse-worded questions are flipped before calculation to ensure consistency and catch random response patterns. The theme score is the average of all questions on that theme. Sub-areas (subfacets) are used to explain nuances but don't change the theme score itself. This makes results stable and predictable.
Age adaptation
The test exists in three versions adapted for adults, teenagers, and children – with different numbers of questions, scaling, and age-appropriate language.
Data quality
Each theme gets a confidence level based on coverage and balance in the questions. When confidence is low, the wording becomes more cautious, and if overall data quality is too low, we don't generate a report.
Zones and norming
Internally, we compare results to expected levels per age group and translate them into five broad zones (e.g., lower/middle/higher tendency). We show zones rather than numbers to avoid over-precision. In v1, we use conservative standard assumptions until we have enough data to establish more precise norms per theme. The zones are therefore best understood as tendencies for reflection and dialogue – not as a precise placement in a reference population.
Snapshot
Results are a snapshot, not a permanent truth. Life situations, stress, and age affect the answers. A test taken six months later may look different.
Psychometric quality
We assess internal consistency (Cronbach's alpha) across many responses when the data foundation is large enough. This is a property of the questionnaire as a whole – not of one person's response – and therefore isn't shown in individual reports.
Privacy
Raw test answers are always private. No family members can see each other's answers – not even in reports.
Report texts
AI transforms the results into readable text. Reports describe patterns and suggest conversation starters and small experiments. The focus is shared language and dialogue – a starting point for conversation.