Research Methodology
Last updated: March 2026 — The statistical engine behind personalised performance.
01 — The Correlation Principle
Most wellness applications apply population-level benchmarks to individual users. This is a fundamental flaw. What works on average is rarely optimal for any single person.
HABITIQ is built on N-of-1 methodology — a research framework borrowed from clinical science where each individual serves as their own control group. Every insight HABITIQ surfaces is derived exclusively from your data, your patterns, and your physiology. We do not average you against anyone else.
"General health advice is often right on average, but wrong for the individual. True optimisation requires individual-specific correlation analysis."
— HABITIQ Engineering Team
02 — Lag Analysis
Biological systems do not respond instantly. Our engine accounts for physiological lag by testing correlations across multiple time offsets (12h, 24h, 48h) — ensuring we capture the true relationship between behaviour and recovery.
03 — Statistical Confidence
Not every pattern is a signal. HABITIQ applies minimum data thresholds and significance testing before surfacing any insight. We would rather show you nothing than show you something misleading.
04 — Data Modelling
HABITIQ uses two primary correlation methods, selected based on the nature of the metric being analysed:
Pearson Correlation
Applied to continuous numerical metrics (e.g. water intake, workout duration) to measure the linear relationship with recovery scores.
Spearman Correlation
Applied to scale-based and ordinal metrics (e.g. stress level, mood) where the relationship may not be strictly linear.
Both methods are supported by two additional analytical layers:
- Bucket Analysis Behaviours are grouped into optimal ranges rather than treated as raw numbers, revealing performance zones specific to your data distribution.
- Delta Mapping Recovery variation is measured against shifts in baseline behaviour, isolating the impact of change rather than just absolute values.
05 — What We Do Not Do
HABITIQ does not use generalised machine learning models trained on other users' data. We do not apply external health norms or population averages to your results. Every correlation is computed locally from your own logged history. Your insights belong to you — and they are built entirely from you.