Randomization of multiple stimuli within the same trial?

MediaLab is not quite as adept as DirectRT when it comes to handling multiple random stimuli [I]within[/I] trials although DirectRT is not yet capable of displaying Word documents. MediaLab could do this by randomizing 200 trials, each containing a pre-determined Word document and then adding a simple HTML script to randomly pick an image to show along with it. However, that HTML code would somehow have to ensure that that the images were being picked randomly [I]without replacement[/I]. That would be hard to do since the HTML would be independent on each trial.

Another possibility--if the Word doc and the image could be sequential as opposed to being on the same screen: it would be easy to do this via the WGR parameter. You would have 400 trials of alternating Word docs and images. All the word docs would have a WGR value of 1 and all the images would have WGR value of 2. However, you wouldn't see them both at the same time. That would be super simple to execute though.

Yet another possibility--you could assign all the item types a constant image and then use a WGR value of 1 to randomize them. And then use a [I]responses.xls[/I] file (see Advanced Features in the user's guide) to generate the random question wording. In the responses.xls file, you could create 200 variables such as "word1", "word2" etc. and assign them your various word document content. Using Excel's randomization features, you could randomize the order of these associations--i.e., which wording is associated with which variable name. Then in your MediaLab item, you would simply insert <word1> as the question wording for your first item, <word2> for the second and so on up to 200. I think this is the way I would do it if I wanted both on the screen at once. The immediate drawbacks here I see are that you would be randomizing blocks of text as opposed to Word documents (unless you can somehow embed Word documents within cells in Excel!) and that you need to figure out how to randomize the order of the text cells in Excel (that part wouldn't be too-too hard--and I could help you with it if need be!).

Final yet another possibility: Try using DirectRT. If your text (i.e., non-Image stim) doesn't HAVE to be a word document, then DirectRT would make this a gloriously tasty piece of cake to execute. 5 minutes to generate the input file and you'd be done. It eats this type of design for breakfast. But then, I don't know if you have access to it or would want to venture out beyond the [I]usually[/I] comfy confines of ML. If you want to try let me know--if you don't have access I can get you a free copy to try for a few weeks to see if you want to go that way.

Anyways, those are my thoughts. Btw, good question--I was surprised there was not a simpler answer for this and will think about how ML can be modified in the future to more easily accommodate situations like this!