5 easy steps for importing SurveyMonkey results into NVivo
Update: NVivo 10 Service Pack 3 now lets you import your SurveyMonkey data directly into NVivo—read Jason's new post Connect to SurveyMonkey and bring your data into NVivo.
SurveyMonkey (www.surveymonkey.com) has become a very popular tool for conducting online surveys. It’s really easy to use, and once you’ve collected your data, SurveyMonkey offers some great features for analyzing the responses to your multiple choice and numeric survey questions. But what if you have open-ended, free text questions on your survey and you want to analyze the responses to those? That’s where a qualitative data analysis tool like NVivo can help.
In this post, I'll give you a few tips for importing SurveyMonkey results into NVivo 10. In my next post, I'll look at techniques for analyzing these results using NVivo 10.
1. Export from SurveyMonkey to Excel
First, you’ll need to download your survey results from SurveyMonkey as a spreadsheet. To do this:
When SurveyMonkey notifies you that “Your export file is complete”, click Download to download the results as a .zip file. Then, using a tool such as WinZip, extract the .zip file to a folder. Your survey results are in a file called CSV\Sheet_1.csv.
2. Cater for accented and non-Western characters
The next step is to open your survey results in a spreadsheet package such as Microsoft Excel, but there’s a pitfall to avoid here if your survey results contain accented or non-Western (Unicode) characters. If you just click on the Sheet_1.csv file to open it, Excel will interpret the .csv file as a non-Unicode file, and you’ll get strange symbols in place of your accented or non-Western characters.
Instead, follow the advice SurveyMonkey gives for opening the survey results in Microsoft Excel in “Why is the Excel file showing the data/characters incorrectly?”, and in the Text Import Wizard’s File origin field, be sure to select “Unicode (UTF-8)”.
3. Prepare your columns
Once you’ve got your survey results into Excel, you’re ready to start preparing them for import as an NVivo “dataset” source.
These preparation steps are important, because once your data is in NVivo as a dataset, it can’t be edited or updated. Don’t get halfway through your analysis and realise you should have prepared your data differently!
First, you’ll notice that the spreadsheet contains two header rows. You’ll need to edit these so that the first row containing what will become your field (column) names in NVivo. Then delete the second row.
- This is also a good opportunity to rename your columns. You might choose to replace a wordy question such as “What did you like about…?” with something simpler like “Likes” for NVivo to use.
- You can choose to delete columns from the spreadsheet if you don’t want to bring them into NVivo, but it’s safer not to unless you’re certain you’re not going to want to use them in your analysis. You can always hide dataset columns in NVivo, but you can’t bring extra columns in later without starting over. In addition to the free text columns you want to analyze, make sure you preserve SurveyMonkey’s RespondentID column (which uniquely identifies each row), plus any demographic data or multiple-choice questions that you might want to use later to slice and dice your free text responses (e.g. you might want to ask “How did the responses to this question differ by age group?”)
4. Prepare your data
Before you import your data into NVivo:
- Scan the free text questions in Excel looking for cells containing the text “#NAME?” This usually indicates cell text starting with a dash (-) or equals (=) symbol, which Excel thinks represents a mathematical formula. To correct this, in the Excel cell editor, add a quote (') symbol at the start of the cell text.
- If you want to spell-check your data, do this in Excel before importing into NVivo.
- Think about whether you’re going to analyze all the free text responses, or just a random sample. When I’ve conducted surveys with thousands of respondents, I’ve often analyzed the free text responses for a random sample of several hundred respondents. You can take a random sample by adding a column containing Excel’s RAND() function to your spreadsheet, then sorting the data on that column and deleting the rows you don’t want.
When you’re done, save the spreadsheet as an Excel (.xlsx) file, and you’re ready to import it into NVivo.
5. Import survey results into NVivo
In NVivo 10, to import your survey results as a dataset, create or open a project to work in, and then from the External Data ribbon tab, in the Import group, select Dataset.
The Import Dataset Wizard opens. Press the Browse button to select your Excel (.xlsx) file.
The NVivo help topic “Import data from spreadsheets and text files” contains detailed guidance for the options offered in the Import Dataset Wizard, but the most important choice you need to make in this import wizard is in Step 4 of the wizard where you specify which columns you want NVivo to treat as codable columns, and which you want NVivo to treat as classifying columns.
In NVivo, codable columns represent free text to be coded and analyzed, and classifying columns represent names or attributes that can be used to slice and dice the data:
If you need to change a column from codable to classifying or vice versa, select the column in the Data Preview by clicking on its header, and then select the option you want from the Analysis Type box.
At the end of the Import Dataset Wizard, choose a name for your dataset and press Finish to complete the import.
What happens next?
In my next post, I’ll cover how to analyze your data once you’ve imported it into NVivo.
Any experiences to share? I’d love to hear your thoughts – just add your comments below.