Analyze open-ended survey responses without reading them one by one
You upload the spreadsheet, pick the column, and the answers end up grouped into categories with their verbatim quotes. Each category comes back with the case number, so it pastes into your dataset as one more column and crosses with the rest of the survey.
- Finds the open-ended column on its own
- You review categories, not individual answers
- Comes back to your dataset as one more variable
From spreadsheet to new variable
Four steps, and none of them requires preparing the file first.
The spreadsheet as it came out of the form
Excel or CSV, without trimming anything. HILO finds which columns are open-ended and offers them to you. The criterion isn't the name: it checks that the answers are almost all different and contain real text, so it doesn't mistake the case ID for an answer, or the wording of a scale for an open-ended question.
Categories come from the material, not from a preset list
The answers are read and categories with proper names are proposed, not loose words. A single answer can fall into two, which is normal: someone complaining is complaining about three things at once.
You decide by category, not answer by answer
The list of categories with how many answers each one has. You accept, rename, merge with another, or discard. Inside a category you remove the answers that don't belong. With four thousand answers the work is almost the same as with a hundred, because the number of categories saturates.
A new column for your dataset
The category table, the quotes with their case, and the column to paste into the spreadsheet by ID.
How you got to those categories is on record
Recoding an open-ended question by hand leaves a result with no history: nobody can reconstruct why those categories and not others. Here it is recorded how many assignments the machine proposed, how many you confirmed, and how many you rejected, category by category. When a reviewer, a client, or an accountant asks how the classification was built, the answer is a file, not a verbal explanation.
Only one column of your dataset is saved
The open-ended column you chose is stored, and nothing else. The rest of the spreadsheet is never even saved, so the names, emails, and other data in the other columns never leave your computer. To export the dataset with the new column you pick the file again and the pasting happens in your browser. In a survey with informed consent behind it, that is not a technical detail.
And if you’re going to share the whole dataset, you can first remove its identifiers, including ID numbers and emails written inside open-ended answers.
The spreadsheet as your form exports it: Excel or CSV. You don't have to prepare anything or trim columns. HILO finds which ones are the open-ended questions and offers you the edit.
None. What you review is the category, not the answer. You see the list of proposed categories and decide on each one: you accept it, rename it, merge it with another, or discard it. Only inside a category do you look at the assigned answers, and there you only remove the ones that don't belong.
Three files. The category table with how many answers each one has. The verbatim quotes with the case number of each. And a new column to paste into your dataset, which turns the category into one more variable and lets you cross it with the rest of the survey.
Edit mode is turned off when the material is a survey. A respondent's answer is the data: if it is touched, traceability is lost.
No. Only the open-ended column you chose is saved, which weighs a few hundred kilobytes. The rest of the spreadsheet is never even stored, so the personal data in the other columns never leaves your computer. When you want to export the dataset with the new column, you pick the file again and the pasting happens in your browser.
It records how many assignments the machine proposed, how many you confirmed, and how many you rejected, per category. That is what makes the analysis reproducible when someone asks how you arrived at those categories.
Try it with your own survey
Upload the spreadsheet, pick the open-ended column, and look at the first categories. In ten minutes you'll know whether it's useful to you.