I'm grateful for everyone's help, and apologize if this has been asked/answered before.
I'm developing an annotation survey where a respondent sees one of 1000 randomly selected text snippets, and answers a few questions about the document. I can load the data through any number of formats directly to the survey (through CSV/Excel/JSON/etc).
The functionality I am trying to find is:
a) how to pull in those 1000 text items without making Text Blocks for each one manually
b) where each person can retake and see a new text snippet
c) each text snippet would appear to no more than 3 unique respondents
What I've found so far has been on the order of 4-20 items that can be manually imported, not 1000. I also have seen these answers.
https://www.qualtrics.com/community/discussion/4073/how-can-i-randomly-assign-images-to-different-conditions-without-replacement (great for sampling without replacement)
https://www.qualtrics.com/community/discussion/4662/piped-text-based-on-external-data (also helpful)
As I understand it, I need to load in each of these 1000 text snippets as a Contact List, load that in as Piped text for each person, then randomize the Contact list. Is this the best way to accomplish this, or have I missed an easier way to solve this problem? Thank you!
I'm developing an annotation survey where a respondent sees one of 1000 randomly selected text snippets, and answers a few questions about the document. I can load the data through any number of formats directly to the survey (through CSV/Excel/JSON/etc).
The functionality I am trying to find is:
a) how to pull in those 1000 text items without making Text Blocks for each one manually
b) where each person can retake and see a new text snippet
c) each text snippet would appear to no more than 3 unique respondents
What I've found so far has been on the order of 4-20 items that can be manually imported, not 1000. I also have seen these answers.
https://www.qualtrics.com/community/discussion/4073/how-can-i-randomly-assign-images-to-different-conditions-without-replacement (great for sampling without replacement)
https://www.qualtrics.com/community/discussion/4662/piped-text-based-on-external-data (also helpful)
As I understand it, I need to load in each of these 1000 text snippets as a Contact List, load that in as Piped text for each person, then randomize the Contact list. Is this the best way to accomplish this, or have I missed an easier way to solve this problem? Thank you!
Page 1 / 1
Hmm. I don't think the snippets as a contact list will work, as those are tied to email addresses (or other contact identifier), so the functionality listed in b won't apply.
You could, in theory, build this using branches and embedded data and 1000 different quotas (to solve the need for each snippet being used only 3 times) but even thinking about it makes me shudder. It's not very practical and I don't even know if it would work at that magnitude.
I hate to just answer to tell you what you can't do BUT I'm not a JS coder. I believe the best way to do this is with an array and a quick google found this: Getting a Random Value from an Array. Maybe others will be able to offer more substantial coding advice.
You could, in theory, build this using branches and embedded data and 1000 different quotas (to solve the need for each snippet being used only 3 times) but even thinking about it makes me shudder. It's not very practical and I don't even know if it would work at that magnitude.
I hate to just answer to tell you what you can't do BUT I'm not a JS coder. I believe the best way to do this is with an array and a quick google found this: Getting a Random Value from an Array. Maybe others will be able to offer more substantial coding advice.
There are a number of ways to do this. This one will have the minimum amount of coding:
1. Add all 1000 snippets as choices to a MC choice question.
2. Use Advanced Randomization to randomly display 1 of the 1000 snippets and check "Evenly Present"
3. With JS: hide the MC question, select the one and only displayed choice, and click the Next button.
4. Pipe the selected choice from the MC question where you want to show the snippet.
1. Add all 1000 snippets as choices to a MC choice question.
2. Use Advanced Randomization to randomly display 1 of the 1000 snippets and check "Evenly Present"
3. With JS: hide the MC question, select the one and only displayed choice, and click the Next button.
4. Pipe the selected choice from the MC question where you want to show the snippet.
Thank you both for your responses! The MC option suggested above is a potential solutions. Though....I'm not opposed to a coded solution, especially one that may preserve the formatting of the text snippets (some may have multiple paragraphs) and that also lets me import/track other information from the CSV/snippet (snippet ID).
I can upload the data to an external server as well if needed!
I can upload the data to an external server as well if needed!
> @schancellor said:
> Thank you both for your responses! The MC option suggested above is a potential solutions. Though....I'm not opposed to a coded solution, especially one that may preserve the formatting of the text snippets (some may have multiple paragraphs) and that also lets me import/track other information from the CSV/snippet (snippet ID).
>
> I can upload the data to an external server as well if needed!
You can have multiple html paragraphs in a choice. Just make sure the only carriage returns are at the end of the choices when you copy and paste them in. You can embed the id in an html tag. Each will also be assigned a unique choice id.
If you really need to keep the id as a separate field, I would use a web service script. Load the 1000 snippets in a database and have the script grab and random snippet and keep track of how many times it has been used in the DB. The script would return the id and the snippet to Qualtrics as embedded fields.
> Thank you both for your responses! The MC option suggested above is a potential solutions. Though....I'm not opposed to a coded solution, especially one that may preserve the formatting of the text snippets (some may have multiple paragraphs) and that also lets me import/track other information from the CSV/snippet (snippet ID).
>
> I can upload the data to an external server as well if needed!
You can have multiple html paragraphs in a choice. Just make sure the only carriage returns are at the end of the choices when you copy and paste them in. You can embed the id in an html tag. Each will also be assigned a unique choice id.
If you really need to keep the id as a separate field, I would use a web service script. Load the 1000 snippets in a database and have the script grab and random snippet and keep track of how many times it has been used in the DB. The script would return the id and the snippet to Qualtrics as embedded fields.
I know this discussion has been some time ago, and I'm currently facing a similar issue. schancellor, do you still know the solution you went for in the end? The suggestion by TomG sounds nice, but I could use some help in setting a web service script up. I don't mind the programming bit, but I've never wrote a web script before, so if you could point out some examples or helpful tutorials that would be really nice!
https://community.qualtrics.com/XMcommunity/discussion/comment/42985#Comment_42985Here is something I posted on StackOverflow that might help.
Leave a Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.