In short
A blinded fragrance test compares versions under the same conditions, in a chance-assigned order, with coded samples, so that neither the assessor's expectations nor yours can decide the result. You've used versions of this informally since Level 2. Here you set it up properly, because the difference between "it seemed better to me" and "it came out better in a fair test" is the entire point of this track.
What you will learn
- Choose suitable controls for a fragrance test
- Randomize and balance sample order
- Set up blinding with a sealed key
Before this lesson: Reading Fragrance Research Critically
Controls
| Control | Purpose | Example |
|---|---|---|
| Comparator | The version you're measuring against | The same accord, without the change |
| Solvent blank | Checks what people report with nothing there | Strip with carrier alone |
| Positive control | Shows your test can detect a known difference | A pair known to be clearly different |
| Identical pair | Shows the rate of false differences | The same sample given twice |
Make the control alongside the test sample, from a single batch where possible, and adjust so that total weight and solvent are equal. Age them together. Handle them identically. A difference in strip dipping or storage is a difference in the test.
Randomizing and balancing
People often judge the first sample differently, and they adapt to what they've just smelled. So order must be managed. For two samples, A and B, half the people get AB and half BA. For more samples, use a balanced set of orders, such as a Latin square, in which each sample appears in each position equally often. Assign people to orders by chance, using a random number tool or by drawing slips of paper.
| Sequence | First | Second | Third |
|---|---|---|---|
| 1 | A | B | C |
| 2 | A | C | B |
| 3 | B | A | C |
| 4 | B | C | A |
| 5 | C | A | B |
| 6 | C | B | A |
Blinding in practice
- Codes. Use three-digit random numbers, rather than A/B or 1/2, because people read meaning into labels.
- Key. A helper prepares the codes and seals the key. You open it only after scoring.
- Appearance. Same vials, same volume, and strips dipped to the same depth. Color differences can reveal a sample. A tinted vial may help.
- Instructions. Written, identical and neutral, so you don't lead anyone. Don't say what the study hopes to find.
- Your presence. If you can't be blind, stay out of the room, or at least silent.
- Analysis. Score the data before you unblind it, with the rule already written.
Controlling the environment
- A well-ventilated, odor-free space, without perfume on anyone or nearby.
- Consistent time of day, temperature and dose.
- Breaks between samples, as you found with olfactory fatigue. State the interval.
- Instructions on how to smell the strips, and how long to do it for.
- No talking between assessors while testing.
Common design errors
| Error | Fix |
|---|---|
| Coding and testing yourself | Ask someone else to hold the key |
| Always smelling the new one last | Randomize the order |
| Comparing fresh against old samples | Make and age them together |
| Testing until you see what you want | Fix the number before you start |
| Excluding someone after seeing their answers | Set the exclusion rules beforehand, and report any exclusions |
Practical exercise: A complete test design with a dry run
- Take the question from your earlier file.
- Write a design sheet: versions, controls, medium, dose, order, codes, key-holder, instructions, environment and timing.
- Generate your allocation table with a stated random method.
- Do a dry run with one person, using strips only. Do not include their results in your final data. Note what confused them.
- Fix the design, and date the final version.
You have done this well if you could leave the sheet with a stranger and they could run it, you would not know which sample was which, and the dry run changed something.
Common mistakes
- Calling a test blind while you know the code. It isn't.
- Using A and B as labels. Use neutral numbers.
- Telling participants your hypothesis. Save it for afterward.
Key terms
- Randomization
- Assigning samples or orders by chance, so that unknown influences are not systematically linked to any one version.
- Blinding
- Keeping those taking part, and ideally those scoring, from knowing which sample is which.
- Carryover effect
- An influence of an earlier sample on the judgment of a later one, such as adaptation from the first strip smelled.
Knowledge check
Written and reviewed by the RUŌOD Lab team. This is general education, not legal, regulatory, medical or safety advice. Example dilutions and exercises are for learning, are not validated products, and have not been safety-assessed. Completing a lesson confers no qualification. How we check these lessons.