Empirisoft Support

    Welcome to Empirisoft Support
Results 1 to 30 of 30

Thread: helping with programming esp. randomization

Hybrid View

Previous Post Previous Post   Next Post Next Post
  1. #1
    Join Date
    Sep 2007
    Posts
    18

    helping with programming esp. randomization

    Hi there,

    I'm getting ready to program MediaLab for an upcoming experiment. I've actually tried a little - but it's not coming out how I'd like so I'm hoping that you can help me.

    In our original experiment, we had videotaped participants during a 10 minute interaction. In our upcoming experiment, we'd like raters to look at clips from these videos and rate them on a number of dimensions. Because we don't want to tire out our participants from watching too much - we've split the 10 minute videos into 30 second clips (so there will be 20 of them for each interaction). We'd like participants to receive one random 30 second clip from the first 2.5 minutes, one from the second 2.5 minutes, one from the third 2.5 minutes, and one from the final 2.5 minutes. After each 30 second clip, they'll rate the video. The trick here is that obviously these random clips need to come from
    the same 10 minute session.

    After that, they would go on to rate another set of 30 second clips from a different interaction. We also would want to make sure that they wouldn't rate an interaction that they've already rated. We have about 40 or so interactions in total - but participants will probably only rate 15 of them or so.

    Does this make sense? How would I go about setting this up? I was thinking that each interaction should probably be set up as a separate condition - but can you program medialab so that participants would randomly view 15 out of the 40 interactions? Also, what code would I use to randomize within each interaction?

    Thanks!
    Robin Freyberg
    Last edited by jarvis24; 09-19-2007 at 01:33 PM.

  2. #2
    Join Date
    Nov 2005
    Posts
    3,328
    I would create 4 questionnaires--each with 5 clips. Randomize the order of the 5 clips within each questionnaire and then play only the first clip in each by using an unconditional skip value of "+4" on all 5 clips--so no matter which one they get, MediaLab will skip over the other four in that questionnaire. I think this may require 2006.2.40 or later.

  3. #3
    Join Date
    Sep 2007
    Posts
    18
    Thanks so much for your response - it's definitely helpful.

    I've been able to get it to work and managed to do a sample run with sets of videos from two separate participants by assigning each participant the same order value but assigning each participant a different value for within-subjects randomization within an experiment file.

    But, the question remains about how to make sure that raters will not end up viewing the videos for all participants. As I mentioned before, there are about 40 different participant videos - there's no way that raters will be able to handle viewing all of them.

    So, what would I need to do so that raters would view a random selection of 20 of the participants out of the 40?

    Thanks for your help!

  4. #4
    Join Date
    Nov 2005
    Posts
    3,328
    Can you zip your folder (without the videos) and attach it so I can see where you are with it?

  5. #5
    Join Date
    Sep 2007
    Posts
    18
    Okay, so here's the sample that I worked on - you'll see that there are only two 30 second clips for each of two different participants.

    Also, because there are sometimes two people in the video, I brought in an image of an arrow to indicate who the rater should focus on. What this means is that I set up the video portion as a backgroundvideo and the image of the arrow is the central feature of the questionnaire. I hope that makes sense.

    So, I guess my question would be if there was a way that the rater would only end up randomnly rating 1 instead of both sets of video clips (but just realize that this will be expanded greatly for more participants, more clips, etc.).

    Thanks for your help!
    Attached Files Attached Files

  6. #6
    Join Date
    Nov 2005
    Posts
    3,328
    I missed that you were going to repeat the whole process 15 out of 40 possible times. That certainly makes it harder. Take a look at the simplified sample I've posted. The is the logic only--no filenames are mentioned.

    Basically, it illustrates a situation where there are movies from three interactions--and each subject will see clips from two of those three. Further, each of the three movies is broken down into three sets of two clips. Each subject will view one of the two clips from each set and then answer a question about that clip.

    See if the logic in the sample makes any sense. If so, you should be able to expand on the pattern to create your larger design. Let me know if you have any questions or if I've missed something.
    Attached Files Attached Files

  7. #7
    Join Date
    Sep 2007
    Posts
    18
    Hi there,

    Sorry it took so long to respond to your post - we were on break so I was away from the software. Regardless, I took a look at your attachment - and it looks like it will solve the problem. I haven't expanded the whole thing yet - but at least with the example you posted, it seems to work. So, thank you! I will let you know if any problems emerge as I expand the programming.

    Thanks again!

  8. #8
    Join Date
    Sep 2007
    Posts
    18

    questionnaire limits

    Hi there,

    If you remember, you were helping me with the randomization for all of those video files. So far, so good - except that it looks like I ran into the limit on a questionnaire file. I'm on the 2,000th row - and it looks like it won't go any further.

    Is there anything I can do to fix this situation? Is there a way to extend the limit? Or should I start with a new questionnaire file and randomize between the different questionnaires? What do you think?

    Thanks for your help!

  9. #9
    Join Date
    Nov 2005
    Posts
    3,328
    You can try opening the file in Microsoft Access and editing it there. Before you do all that work though, I would test what you have so far--things might start getting slow once you have 2000+ items in a single questionnaire!

    See also: http://www.empirisoft.com/support/showthread.php?t=663

  10. #10
    Join Date
    Sep 2007
    Posts
    18
    Thanks for your reply. I've tested it out a little and have no problems running the study. Things get a little slow when I'm copying and pasting in the questionnaire - but the actual running (so far) seems to be okay.

    I'm a little nervous about using access since I've never worked with it before. Would it be possible to create several questionnaire files (let's say 4) with the video files equally broken up among them and then participants randomnly get the videos from that questionnaire file. How would I set up the experiment to randomnly choose one of the four questionnaire files?

    Thanks for your help!

  11. #11
    Join Date
    Nov 2005
    Posts
    3,328
    That would be a bit tougher because, although you can randomly order the questionnaire files within the greater experiment file, you can't set up skips in the exp file like you can in the que file.

    With regard to Access, it really simple if all you are doing is editing the questionnaire information--it's like editing in Excel. Maybe make a copy of the file and just try it and see if you like it. I would definitely prefer it if I were working with so many items. But, hey, if it's working the way you're doing it, then maybe there is no need.

  12. #12
    Join Date
    Sep 2007
    Posts
    18
    Another problem - I tried working with Access but the formatting looked strange when I brought the questionnaire file over. So, I've been sticking with splitting the experiment into 4 questionnaire files.

    However, I think I may have run into another limit of the program. Do you know if there's a limit for the numbers that can be entered for WGR and BGR?

    This happened in one of my previous files when I ran a mini experiment to make sure that everything worked, I got the error message "9subscript out of range". It seemed to work again once I changed the values for WGR and BGR.

    Have you seen this before? Do you know what the limit on these files are for this value? Or is this referring to something else?

    Thanks!

Similar Threads

  1. two blocks of between group randomization
    By Michal in forum DirectRT Older Versions: How Do I...
    Replies: 4
    Last Post: 03-01-2010, 05:03 AM
  2. Complicated Block Randomization Issues
    By klgreen in forum DirectRT Older Versions: How Do I...
    Replies: 1
    Last Post: 09-19-2007, 02:35 PM
  3. Randomization and Skip Patterns
    By eop2 in forum MediaLab Older Versions: How Do I...
    Replies: 5
    Last Post: 06-28-2007, 12:05 PM
  4. Too much Randomization
    By L4lab in forum DirectRT Older Versions: Troubleshooting
    Replies: 2
    Last Post: 03-31-2007, 05:18 PM
  5. Presenting multiple text stimuli w/ randomization?
    By sangoh in forum DirectRT Older Versions: How Do I...
    Replies: 1
    Last Post: 03-26-2007, 09:44 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •