Data Weighting allows researchers to rebalance their sample results to more evenly distribute responses across a population. When you use this feature in your Results, weighted data will be reflected across platform charts and crosstab exports.
1. Preparing to construct your custom schema
You will need:
- your desired metric or variable for weighting
- the exact target-percentage for each option or value
- the sampled percentage for each option
- a raw data file from your survey with individual values to be weighted (Excel with Coded Map is recommended)
2. Constructing the Schema
- Download a raw data file from your survey (Excel with Coded Map is recommended).
- Delete all columns EXCEPT the Response ID column and the original variable column (ex: gender).
- Add a column for the custom weight with a title that starts with aytm: followed by either wt or weight, and another word or phrase to help you remember what the variable is.
Ex: aytm:wt_gender or aytm:age_weight
- Calculate the value of each group by dividing the target percentage by the sampled percentage.
Note: You must use exact, non-rounded percentages in calculations. Weighting should not create extreme changes within the data set - values should be greater than 0.2, and less than 5.
- Populate the aytm:wt___ column with the appropriate weighted value for each respondent
Tip: use a filter and copy+paste the values, or consider an if/then formula.
- Save the file as a .CSV or in Excel format.
weighted value = target % / actual(sample) %
3. Data Weighting your results
- On your results page, click the Filters tab, scroll to the bottom, and click Upload custom variable.
- Select the weighted variable file from your hard drive.
- Click the Data Weighting icon in the left menu.
- Click the Choose Weighting Variable drop-down, and select Weight by custom variable.
- click the drop-down to select the custom variable you uploaded.
- Toggle Weighting is OFF to Weighting is ON.
The data on your page will reload with the weighted variable data. From here you can save as a new report, or reset the data to the Original - as fielded report.