TCM Constitution Questionnaires: What Scores Can and Cannot Show

Identify the questionnaire before interpreting its result

“TCM body type quiz” can describe very different tools. The original Constitution in Chinese Medicine Questionnaire, or CCMQ, has 60 items covering nine constitution scales. There are adapted language versions and researched short forms. A website can also create its own simplified questions and show a ranking without using any of those instruments.

Those tools are not interchangeable merely because they use labels such as Qi deficiency, Yang deficiency or balanced constitution. The question wording, response options, scoring and intended use matter. Find the full instrument name and version before looking for an explanation of the number.

The 2013 Hong Kong validation study adapted the CCMQ for Cantonese-speaking Chinese people and tested its measurement properties. A 2023 short-form study developed and compared specific abbreviated versions, selecting a 27-item version named SF-WQ9CCMQ-A. It did not validate an arbitrary set of fewer questions chosen by another website.

On this site, the fourteen-question quiz is a simplified educational reading-order sorter. Its page states that it is not an assessment or validated personal result. It is neither the original 60-item CCMQ nor the studied 27-item short form. References to those studies explain the background; they do not transfer the studies' performance figures to this quiz.

Ask what the displayed number actually measures

Look for a scoring explanation rather than guessing from the shape of a number. A display such as “65 out of 100” may be a transformed questionnaire score. It does not, by itself, mean a 65% chance of having a disease, a 65% certainty that a constitution label is correct, or 65% of an organ's function.

Check whether the tool reports a raw sum, a transformed scale, a category threshold or simply a ranking. The 2023 paper describes scores transformed to a 0–100 scale. A scale's numerical range alone does not make it a probability or a clinical measurement.

Also check the direction. In the Hong Kong study, the gentleness scale and the other constitution scales had different directions of association with health-related quality-of-life scores. “Higher” cannot be read as “worse” across every scale. Reverse-scored items and category-specific rules can also make a do-it-yourself sum wrong.

A small score change can produce a category change

Scores and categories are different outputs. A category rule draws a boundary through a numerical scale. Two nearly equal scores can fall on opposite sides of that boundary even when the underlying answers changed very little.

For an original hypothetical example, imagine a tool that assigns a label at 40 or above. A score of 39.9 would not receive that label, while 40.1 would. This is a demonstration of a threshold rule, not a CCMQ scoring instruction or a claim that these decimal scores are achievable in a particular version. The changed label alone does not show a major health change.

A rank has a similar limitation. A fictional sorter might place scores of 18, 16 and 14 in its first three positions. They are still the top three even if none meets a separately defined threshold. This example explains ranking only; it does not describe this site's scoring algorithm or assign a threshold to its quiz.

Keep the original score and the classification rule beside the label. If the rules are not available, record “scoring rule unknown” rather than inventing a cutoff from a different article. The two research papers below use specified rules for their own analyses; copying one number without its version and context can mislead.

Read “validated” as a specific research claim

Validation is a collection of questions about a measurement, not one stamp that proves every possible use. A paper may examine whether the wording is understood, whether related items behave together, whether scores repeat, or whether the result agrees with another measure. Identify which question was actually tested.

Ask “validated for whom, in which language, against what, and for which use?” A reliability coefficient is not the chance that your particular label is correct. Likewise, a study of measurement properties does not automatically show that a diet, herb or treatment selected from the score improves outcomes.

Two published examples show why the details matter

Hong Kong CCMQ adaptation, published in 2013

The study recruited clinic patients, with 1,084 participating in the cross-sectional validation study. The questionnaire was administered by an interviewer. The authors found support for several measurement properties, including the reliability of continuous scale scores.

The two-week scale-score reliability coefficients were 0.71–0.88. Yet agreement on returning the same constitution category on retest was lower, with kappa values of 0.318–0.531. These are different statistics answering different questions; they are not percentages of people correctly diagnosed. The authors explicitly identified limited reproducibility of specific classifications.

The study also compared classifications with Chinese medicine practitioners' assessments. That reference concerned TCM constitution categories, not a laboratory test for a disease. Its limitations included convenience patient samples, interviewer administration and a lack of standardization of practitioner diagnosis. Those details constrain what can be assumed about an unsupervised online quiz.

Short-form development, published in 2023

The authors developed three named short forms and selected the 27-item SF-WQ9CCMQ-A. They assessed reliability and validity, and compared classification performance with the original CCMQ. For that comparison, the short-form item responses were extracted from data collected using the original questionnaire.

The reference for the reported receiver operating characteristic analysis was the original CCMQ. The study's “diagnostic validity” wording therefore needs that context: performance against the parent questionnaire is not proof that an online score independently diagnoses a disease. The authors also called the research preliminary and noted convenience sampling and a need for further cross-cultural research.

These are selected examples, not a systematic review of every constitution instrument. They demonstrate why the version, administration method and comparison measure must stay beside any headline about accuracy.

Use this ten-point result-reading checklist

1. Find the instrument name. Is it CCMQ, a named short form, a translated version or a website's own quiz?
2. Match the actual questions. Does the question count and wording correspond to the cited version? A common title is insufficient.
3. Check the language. Look for a described translation or adaptation rather than assuming a literal translation preserves meaning.
4. Follow the stated recall period. Answers about today and answers about several months are not the same record. Note if the instructions are unclear.
5. Mark missing or misunderstood items. Do not silently replace an unanswered question with zero or “never” unless the instrument's instructions say so.
6. Find the scoring rules. Note the scale direction, any reverse coding, transformation and category thresholds. Do not combine rules from different versions.
7. Keep scores beside labels. Record close scores, multiple categories and whether the output is only a relative ranking.
8. Read the study population. Check age, setting, language, exclusions and how the questionnaire was administered.
9. Identify the reference standard. Another questionnaire, practitioner classification and an independently assessed medical condition answer different research questions.
10. Choose a proportionate next step. Use an educational result to understand terms or prepare questions. A personal health decision requires appropriate care and evidence beyond the quiz.

If an important detail is unavailable, write “not provided.” You do not have to complete the missing method yourself. A clear uncertainty is more useful than a confident interpretation built from unrelated scoring instructions.

Before comparing two results, check what changed

A different result can reflect a change in answers, question wording, recall period, scoring rules, translation or administration. It might also reflect a real change in the experiences being reported. The label alone cannot tell you which explanation applies.

For a useful comparison, keep the exact version, dates, original scores and completion conditions. Note a question you understood differently or a change from interviewer assistance to self-completion. If the tool changed its questions, start a new record rather than presenting the difference as a measured improvement.

Do not repeatedly take the quiz to obtain a preferred category. A lower score after starting a product does not establish that the product caused a benefit. Timing alone leaves other explanations open, and a questionnaire result is not a substitute for evaluating symptoms or treatment outcomes appropriately.

Descriptions such as “balanced” or “gentleness” also do not clear a person of medical concerns. If you have a health concern, discuss the actual experiences and timing with an appropriate health professional rather than presenting the category as the explanation.

Download a printable result-reading record

Download the editable questionnaire result-reading record. It has space for the version, scoring instructions, study population, comparison measure and unanswered questions. It is an original reading aid, not a reproduction of the CCMQ or a newly validated questionnaire.

Instrument: exact name and version
Language and question count: record what you used
Score: record the value and its stated meaning
Category rule: record the source, or “unknown”
Research: identify the population and comparison
Limit: state what the result cannot establish
Next step: a reading question or appropriate professional discussion

Keep a personal copy if it is useful. This page does not ask you to submit answers or medical details. When preparing for a consultation, describe relevant symptoms in ordinary language and bring the questionnaire source if you want to discuss it.

NCCIH's TCM information advises discussing complementary approaches with health care providers and avoiding replacement or delay of conventional care. A constitution score does not choose a herb, supplement, medication change, restrictive diet or safe waiting period. That boundary matters especially when an app places product suggestions directly beside the result.

Five checks before relying on a quiz headline

These original cases test interpretation, not your constitution.

Read the evidence behind a claim

• WHO TCM Terminology: Cite the 2007 and 2022 Books Accurately
• WHO Body Constitution and the CCMQ Nine Types: Different Frameworks