Background - my DCE has 5 attributes: 2 are scenario attributes (about the individual) and 3 are product attributes.
After pilot testing, it has become apparent that some respondents to my DCE are confusing scenario attributes with product attributes. As a result, I would like to split the table presented to respondents in two; with the first half presenting scenario attributes (with no header), and the second half presenting the product attributes (with the original header above them). I understand this might be an issues also faced by others in designing their experiment.
I have encountered a few issues when trying to work round this. From my understanding, you can only have one header per table and I do not want the same header for both parts of the table. Using the attribute also doesn’t help because it will hide both instances of the header (rather than just the top one). I have explored a few other workarounds like hiding the headers and adding a row to mimic a header at the start of the table split but these are not ideal.
Attached image shows what I am trying to achieve (but effectively unable to remove/hide top header).
What is the best way you would recommend presenting scenario and product attributes to respondents separately? How would be best to split the table so that the header only appears on top of the latter (product) attributes?
If you could also explain how to reinstate the border under SCENARIO ATTRIBUTE 2 that would be excellent!
This is very helpful and I will use the code in the tutorial to double-check my code.
However, one issue still persists that above the Scenario Attributes, there is still the Product A and Product B heading. Technically, even in your example, your scenarios are independent of your products and should be presented as such. When piloting my DCE (similar to above), this became an issue with respondents thinking that Scenario 1/2 was associated with either of the products.
I think a couple of options would be ideal:
Merging the top header so it reads “scenario” or “your disease characteristics” etc. and keeping Product A and B header on top of the product attributes.
If this is not possible (or for simplicity) removing the top header altogether might also work.
It look like the 3rd example (page 4 of the preview) in the tutorial shows you how to do this.
Having this an an option would be good in a very narrow case like this - but would make the ui very complex and then what about all the other cases?
What would be good would be to load up some sort of choice template - but the HTML editing is pretty easy IMHO its about a minute’s work.
Hi,
I also wish to include scenario attributes in my experiment, I have a question preluding the formatting: how do I add the scenario attributes to the experiment and make sure that I keep the levels consistent across both alternatives?
Vera,
If you just want to keep the scenario attributes levels the same, but have them both displayed under each alternative, then your experiment design should reflect this. i.e
if your scenario attribute is X1 and this is in columns 1,5 and 15 in the design, then columns 1,5 and 15 need to be the same. Your experimental design is nothing magical, it’s just a playbook for which levels to display together.
Otherwise can you elaborate on ‘keep the levels consistent across both alternatives’?
I want to recreate the experiment as displayed on page 4 of the tutorial.
(1) I don’t know how to set up the experiment as such, that the levels for the scenario attributes “pain” and “mobility” are the same across product A & B. I know the theory behind, but don’t know how to specifically enter this in survey engine. I have added constraints to my design, but this did not work.
(2) To remove the headers above scenario 1 & 2, the tutorial refers to “To see how this was done - look at the HTML template for the Merged Scenarios V2”. Where can I find this HTML template? I cannot find it in the zip files. I have gotten as far as this, but I cannot get in the appropriate text, how do I this?
Vera,
The Experiment renderer in SurveyEngine just replaces ‘tagged’ attributes with the current experiment scenario so the solution is purely a display template one. The renderer will render whatever set of attributes you have.
So the trick is to set up your experiment design to duplicate the ‘scenario attributes’ across the alternatives - this is not entirely necessary as you’ll only be displaying one of the alternatives merged across, but it is good practice.
Then in your experiment template its just repositioning and merging cells as per the tutorial.
The ‘Experiment Merged V2’ is under Experiment then select this from the dropdown as below.
Then selecting layout->template.
Looking at your attempt - it looks like the colspan is forgetting there is a very small spacer column in position 3.
Maybe post your html templat once you’ve found it and some kind soul will fix it up for you. (maybe me)
The ‘zip file’ is a red herring it’s just the format of the wrapped up project file.
