Create skip patterns based on previous responses (case 8698)

How do I create a skip to option based on several responses? I want a question to be asked only if participants answer "no" to a series of previous questions, and a different question to be asked if they answer "yes" to the same questions.

Hi lcava1!

You can accomplish this using the skip to feature in the questionnaire files or in the responses.xls file. I'm also going to assume that you are able to use scale response items for all of these questions, most likely with just two options: "yes" and "no."

The primary difference is whether your skips depend on the answer to each individual question or the answers as a set. If you want to or can direct people from one question to another based on their responses to the question that came right before it, you would just use the skip to feature. Just enter the variable name that you would like a response choice to lead to in the "skip to" field next to the "text label" field for that answer choice.

If you instead wanted to direct people to a particular question based on if their answers to a set of items, you could do this by using the responses.xls file. I won't go into all of the specific details about this option here; you can find detailed informaiton about it in the "Advanced Features" section of the MediaLab manual and an actual example in the folder named "Sample6 Advanced Features" in the Samples folder where you installed the MediaLab program files. In the responses.xls file you would include in your experiment folder you would need to create a calculated value using Excel conditional logic. Let's say that you have 10 questions named item1 through item10 in a MediaLab questionnaire file. You would put all ten of those item names in variable column of the responses.xls file. You would then make a new variable, let's call it "total," in the responses.xls file. In the value cell of the responses.xls file directly to the right of total, you can calculate the total number of yes or no responses. In the skipto field next to that value field, you could then create a formula of which variables in the MediaLab questionnaire files to direct participants to.

Please don't hesitate to ask if you have any other questions. Good luck!

Thanks for your help Jason! Unfortunately, I'm still having some trouble getting the complex skips to work. I tried to adapt the resonses.xls file, but it's not reading the series of "no" responses for some reason...Any chance you could help with it? Thanks!

I got the files that you attached; thanks for including them! Can you please provide me with more detail about exactly what you want to accomplish with these files (e.g., where the skips should be and what items should skip to which items) and, more importantly, exactly where the problems are?

Thanks!

I basically want participants to be asked a series of follow-up questions only if they respond "yes" to one or more questions about their drug use. So for instance, if they answer yes to question 15, 16, or 17, I'd like questions#18-20 to be asked, but if they answer "no" to all three (q15, 16, AND 17), I'd like questions 18-20 to be skipped. To do this, I assigned {0} values to all the "No" responses and had excel sum the results of q#15-17, such that if the sum is less than 1, it should skip the subsequent questions (18-20). It works fine if they answer yes to any of the three questions, but it won't read the skip pattern if the sum of the question responses is 0 (i.e., they answered no to 15, 16, and 17). This same process applies to the other skip patterns in the excel file, so I'm assuming it's something I'm doing wrong, but I just can't figure out what it is...

Hi lcava.

Sorry that it took me a while, but I looked at your file and found why the skip patterns you created in the responses file weren't working. The main reason is that, like with the skip to column in the MediaLab questionnaire files, the skip to formulas you use in the responses.xls file need to be in the row of the experimental variable to which the apply. Instead of creating new variable names and rows for them, you simply needed to move the formula to the proper row. If you take a look at the file that I attached, you can see that I put the formula you had for the variable "OtherUseProbs" in the row for the experiment variable "OthUse17." Now, right after participants answer "OthUse17," MediaLab looks at the formula totaling the responses to "OthUse15," "OthUse16," and "OthUse17" (called "OthUse15to17Total" in the attached file) and then uses the formula in the skipto column for OthUse17 in the attached responses.xls file.

You would also need to change how you calculate responses to your multiple response items, "OthUse12.1" and "OtheUse12.2." As the MediaLab manual points out in the Advanced Features > Tips section, the variable names used in the responses.xls file need to match the names of the variables recorded in the data files. Because "OthUse12.1" and "OtheUse12.2" are multiple response items, they are recorded in the data file with_xx following the item name where xx represents a number for each of the choice options. If you take a look at the attached responses.xls file, you can see that I put the variable names for all 7 options of each of the 2 multiple response items in rows B2 through B15. You can now use the 0 or 1 recorded for each of these choices to create skip patterns as needed.

I hope that this helps and gets you started on creating the skip patterns you were looking for. Please let me know if you have any other questions.