Lesson 2 of 2 in this stage
Read longevity results: averages, ranges and test limits
Analyze repeated observations without selecting the best run or making a universal promise.
A number needs a definition
'Eight-hour longevity' is incomplete without surface, amount, detection criterion and number of observers. Persistence on paper does not establish the same skin duration or constant projection. State the method, then the result and the limits of generalization.
An average can hide variation
Teaching example using the same criterion: A records 4, 5 and 6 hours; B records 2, 5 and 8. Both average 5 hours, but A is more consistent in these observations and B more variable. Selecting only B's 8-hour run does not prove it always lasts longer. Three trials are too few to establish a general difference for all users.
The endpoint may remain unobserved
If observation ends at six hours while scent remains detectable, report 'at least six hours in this test'. Do not declare it ended at six or assume it continues into the next day. If a shower ends the trial, record it rather than attribute disappearance solely to the formula. Good records state why observation ended.
A stronger comparison reduces bias
Use A and B codes instead of names where possible, vary order across sessions and standardize detection criteria and breaks. Familiar smells may still reveal identity, so blinding is imperfect. The aim is to reduce expectation, fatigue and brand preference effects, not claim a home trial is a complete laboratory study.
Longevity: a distribution is better than the best trial
If five wear trials return different last-detectable times, selecting the longest and advertising “12-hour longevity” is not a fair reading. Report all observations or a median with range and conditions, especially when tests are subjective and sample sizes small. The endpoint also depends on the observer's detection threshold and how they were asked.
What repeated testing can establish
- Keep dose, surface and environmental conditions as similar as possible.
- Define “end of longevity” before testing: no longer detectable at a distance or even close to skin?
- Repeat trials and retain between-person variation rather than hiding it.
A small convenience sample does not support a universal promise across all skin types and weather conditions.
Your turn
Use what you learned
A scent remains detectable when testing stops at five hours. Report this accurately.
See the answer and why
'It remained detectable for at least five hours under these conditions; I did not observe its disappearance.' No unobserved endpoint is invented.
Check your understanding
Answer four questions, then review the reason for each answer. You can retry.
Interpret the evidence
A reproducible protocol specifies substrate, dose, temperature, checkpoints and a detection endpoint. Report both average persistence and variability, since a single trial cannot predict all wearers.
SESH exercise: choose two scents and record results before viewing their marketing descriptions.
