AtelierLe journal · French, understood
English

Language

The course

Updated

Phrasebook

Explain limits in a French environmental survey

Edited by Chris Lee · Updated · About the library

Explain survey limits in French: response rate, self-selection, a missing group and changed questions. Four practical scenes preserve useful evidence.

In short

Name the specific limit and the conclusion it affects. Le taux de réponse est faible concerns participation; les répondants se sont portés volontaires concerns selection. Neither phrase means every answer is false.

Yellow the idea to rememberBlue French to learn

A visitor fills in a printed questionnaire about Wikipedia at the YOU fair in Berlin.
A paper questionnaire at the YOU fair, Berlin, 2008. Offering another response method can reach people an online-only survey misses. © Raimond Spekking / CC BY-SA 4.0 (via Wikimedia Commons) · resized and converted to WebP

A survey on a recycling scheme receives encouraging responses. A committee needs to understand what those answers can represent. The four scenes below describe separate limits and offer a useful next step. Practise the precise question rather than replacing every limitation with a vague “the survey is biased”.

1

Response rate

Committee colleagueNous avons reçu cent réponses. C’est suffisant ?We received a hundred responses. Is that enough?
YouSur combien de personnes sollicitées ? Le taux de réponse doit être précisé.Out of how many people invited? The response rate needs to be specified.
YouLes non-répondants pourraient avoir des pratiques différentes.Non-respondents might have different practices.
2

Self-selection

Committee colleagueLe questionnaire était ouvert sur notre site.The questionnaire was open on our website.
YouLes répondants se sont donc portés volontaires.So respondents volunteered to take part.
YouCes réponses décrivent ce groupe ; sa représentativité reste à examiner.These answers describe that group; its representativeness remains to be examined.
3

Missing group

Committee colleagueLes habitants semblent tous satisfaits.Residents all seem satisfied.
YouLes personnes sans accès à Internet n’ont pas été interrogées.People without internet access were not surveyed.
YouIl faudrait proposer un autre mode de réponse pour inclure ce public.Another response method would be needed to include that group.
4

Changed wording

Committee colleagueLa satisfaction a augmenté depuis l’an dernier.Satisfaction has risen since last year.
YouLa question n’était pas formulée de la même manière.The question was not worded in the same way.
YouComparons d’abord les formulations avant d’attribuer l’écart à une évolution des opinions.Let’s compare the wording before attributing the gap to a change in opinions.

A response count is not a response rate

One hundred answers from a hundred invitations and one hundred answers from ten thousand invitations have different participation rates. The count alone cannot give that rate. It also does not establish representativeness: who was invited and who answered matter. Describe each element separately instead of treating a single impressive number as a full method.

Le sondage apporte des témoignages utiles, mais ne permet pas de généraliser à l’ensemble des habitants.

The survey provides useful testimonies, but does not allow generalisation to all residents.

Question 1No account needed
What is needed to calculate the response rate?

Choose one answer.

Answer and explanation

The invited population count as well as responses. The rate divides responses by invitations or the eligible invited base, so both counts must be known.

Question 2No account needed
A question changes between years. What should you examine first?

Choose one answer.

Answer and explanation

Whether the wording changes what is measured. Different question wording can produce different answers even without a change in underlying opinions.

A useful recommendation follows the limit: broaden contact methods, describe non-response, preserve common questions or report subgroups separately. The point is to fit the conclusion to the evidence that exists, while identifying what a better follow-up would need.

Keep going