> ## Documentation Index
> Fetch the complete documentation index at: https://docs.liveclin.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Choose scales
> Choose the right response mode in LiveClin: single choice, linear scale, emojis, metric, number and photos, and how many points each scale needs.
LiveClin offers seven response modes. Each one measures a different kind of thing, and swapping one for another is the most frequent technical mistake in questionnaire building.
If you don't know the modes in practice yet, see the [Create questions](/en/tutoriais/questionarios/criar-perguntas) tutorial.
## Decision map
| Mode | Measures well | Use when | Pitfall |
| ------------------------------ | --------------------------------------- | -------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- |
| Single choice | Behavior with real, discrete categories | The answer fits in up to 5 clear options and you want to rank from terrible to great | Options that overlap or don't cover every case |
| Linear scale (0–5, 0–10, 1–10) | Continuous subjective perception | You want a trend across weeks and a readable chart | Using it without labeling the endpoints; changing the scale mid-follow-up |
| Emoji scale (Likert) | Perception with an affective charge | Mood, satisfaction, stress, perceived exertion, patients who aren't comfortable with numbers | Using it for quantity. An emoji is not a count |
| Metric (kg, km, steps, g, min) | Objective data with a unit | The patient can actually measure it | Requiring a measurement they don't take; making it required |
| Number | A count without a unit | Number of workouts, meals out, episodes | Asking for precision that memory can't support ("how many grams of protein") |
| Open answer | Context, cause, reflection | You need the why, or you want to prompt self-analysis | More than one required per questionnaire |
| Photo upload | Visual progress | Fine aesthetic change, a lesion, a plate, a lab result | Making it required, which sends the answering cost through the roof |
## How many points a scale should have
Preston and Colman tested scales from 2 to 11 points in a study published in Acta Psychologica in 2000:
* Two-, three-, and four-point scales performed worse on reliability, validity, and discriminating power.
* The indices rose up to around seven points.
* Test-retest reliability dropped on scales with more than ten categories.
* Respondent preference was highest on the ten-point scale, closely followed by the seven- and nine-point ones.
In other words: what the respondent prefers and what holds up under repetition are not the same scale. The recommendation is to accept that conflict and split the difference:
* **0–10** on one or two global anchors, such as the week's rating and confidence for the week ahead. It's the scale the patient already knows and the one that produces the most expressive chart.
* **1–5, linear or emoji**, for the rest of the core, where week-to-week stability matters more than preference.
* **Single choice** whenever the answer has real behavioral categories, because a well-written option carries more clinical information than a rating.
## Three scale rules that aren't worth breaking
"0 = worst possible sleep, 10 = best possible sleep". Without a verbal anchor, the patient recalibrates the ruler every week and the series turns into noise.
Higher is always better. If a question inverts that, such as pain level, rewrite it: instead of "how much pain did you feel", ask "how was your physical comfort".
A question that was 0–10 and became 1–5 destroys the comparability of the history. That's why LiveClin locks question editing once the questionnaire is saved: the platform is protecting your chart.
## Next steps
The twelve writing rules that protect data quality.
Tested questions with the mode and scale already defined.
* Preston, C. C., & Colman, A. M. (2000). Optimal number of response categories in rating scales: reliability, validity, discriminating power, and respondent preferences. *Acta Psychologica*.
* Lozano, L. M., García-Cueto, E., & Muñiz, J. (2008). Effect of the number of response categories on the reliability and validity of rating scales. *Methodology*.