mprinzing Yes you can do this using the loop and merge feature in Qualtrics, using the exact methods you want.

You can get participants in survey 1 to provide text. You can then download the data from survey one in excel format. You then copy the cells with the text you want (needs to be vertically). You then paste them in the loop and merge fields for the correct block.

Loop and merge allows you to randomize the presentation of the texts, as well as decide if you want participants to answer for all parameters in the loop and text fields (e.g., each text) or just a random subset.

You can read more here -> https://www.qualtrics.com/support/survey-platform/survey-module/block-options/loop-and-merge/

If I have the second group evaluate a few randomly selected texts, will this method ensure that each text is seen once? There will be hundreds of texts, and I'd like to have about 50 or 100 judges. What I need to avoid is a situation where some texts have been evaluated repeatedly and others go un-seen.

https://community.qualtrics.com/XMcommunity/discussion/comment/45716#Comment_45716If you select to randomize the presentation of the messages, and only specify you want X/N shown (e.g., 10/100), then there is no guarantee regarding the distribution of presentation. you would just have to ensure you have a large enough sample size.

The only way that I believe this could work is if you put each message into its own block and then put each block (even if that is 100s) into a randomizer using survey flow. In the randomizer element you can then select to only display x/N of the blocks (statements) and then tick "evenly present elements" - qualtrics will then try to ensure that each element is presented evenly but randomly across all participants.

you could in theory take a hybrid approach. Let's say you have 100 statements you want to assess. You can place 10 statements in a loop and merge field, in 10 different blocks. You can then place those 10 blocks in the randomizer using survey flow. You can then have the randomizer set to randomly distribute 1 of the 10 blocks to each participant with even presentation. That way each participant would rate a subset of 10 statements, and qualtrics will work to keep it even across statements. However, it isn't truly random at this point because you manually decide which statements are grouped together. As a fellow researcher, I wouldn't get too caught up on this and would put each statement into a randomization software and group by 10 that way.