Why import previously coded data?
If you have transitioned from another coding tool or previously did your coding manually in-house, that 'bank' of coded data can be used to train the codeit AI so it can auto-code future verbatims. The investment in time spent importing coded data can bring huge time savings as codeit automatically applies codes based on its learning from this historical data.
What is the format of previously coded data?
The examples below show the two different formats you may encounter when importing coded data into codeit.
Concatenated
RID | Q1 Verbatim | Q1_C |
1 | I love the pizza | 103 |
2 | The pizza arrived in a timely manner, it was hot, and the ingredients were as requested. | 301;104;202 |
codeit only supports semicolon ";" as the delimiter for codes in the Concatenated format.
Expanded
RID | Q1 Verbatim | Q1_1_C | Q1_2_C | Q1_3_C |
1 | I love the pizza | 103 | ||
2 | The pizza arrived in a timely manner, it was hot, and the ingredients were as requested. | 301 | 104 | 202 |
How to import previously coded data
You can follow the steps below to import pre-coded data, this guide often refers to the Survey Data Import Wizard documentation.
First, select your project and files similarly to step 1 to 3 in the documentation above.
When selecting the file step 3, in the Options tab, you have to toggle "Contains Coded Data" on:
Click "Next" after toggling the "Contains Coded Data" toggle.
In the "Select Fields" step, you have to do different steps depending on if your data is expanded or concatenated.
If the data is concatenated, you have to select "Target Variable" as Field Type for each coded field:
You have to select "Target Variable" as Field Role for the code data to be correctly imported. Please check also that your codes in the concatenated fields are separated by a semicolon. For example, for verbatim "price and service", the concatenated codes might be "1;2".
If your data is expanded, you first might need to collapse the fields together into 1 single field:
You can find more information about collapsing on this page. Once the coded fields have been collapsed to a single field, you need to select both "Multicoded" as Nesting type, and "Target Variable" as Field Role:
You have to set those 2 fields for the concatenated coded data to be correctly imported.
Once the fields have been selected, you can proceed to the next page by clicking "Next" and finish the import as outlined in the Survey Data Import Wizard documentation.
How to import looped coded data?
If your data is also looped, the "Select Fields" step is more complicated. You first need to select and set Nesting Type to Multicoded and Field Role to Target Variable for each coded field within the loop, as described above.
For example, if you have a loop with q1_1_c, q1_2_c and q1_3_c, you have to select "Multicoded" and "Target Variable" for each iteration:
You then have to collapse the iterations for the coded data together and choose the same Nesting Type as the source field:
Leave the Field Role as "Target Variable" for the coded field.
Once all fields are collapsed, you can click "Next" and finish the import.