Sample adjustment is a technique that consists of modifying the sample of respondents in a survey to ensure that it is representative. Adjustment is necessary when bias, possibly due to the method of administration, has resulted in the over- or under-representation of certain categories of respondents in the final sample.
Adjustment can be made by extracting over-represented profiles or by counting under-represented profiles several times.
In the parameters of your indicator, under the heading "Statistics" you will find the option "Adjustment" :
The adjustment applies only to Questions, Segments and Attributes.
When you click on the adjustment button, the following fields appear:
- Form field value: which corresponds to the different ways of answering the chosen question, the different segments or the different attributes.
- % original: which represents the quantity of results obtained.
- % target: which corresponds to the adjustment you wish to apply for representativeness reasons.
For example, the breakdown of respondents by country is as follows: 23.08% Canadian, 38.46% French, 21.54% Japanese, 16.92% English. This does not reflect the actual proportion of each country. We therefore apply an adjustment so that Canadians represent 20%, French 30%, Japanese 10%, English 40%
The "Apply Profile" function allows you to select a adjustment profile from those already saved on your account. If you wish to add new profiles that you can use again later, click on "Manage your profiles".
You can also access the list of your adjustment profiles by clicking on "Settings" in the side menu and then "Adjustment Profiles".
Then click on the green "Add profile" button. You can now configure your adjustment profile. For example, to respect the proportion of men and women in the answers, we create a "Gender " profile in which men represent 40% and women 60%. We save this profile so that we can reuse it:
Don't forget to save it !