Generating and - saving - random values from responses.xls files

[jarvis24]

Try taking a look at the section on Advanced Features--especially the subsection on calculated values. You could create variables that would be "stand-ins" for the vignette values. For example, you could generate the three values for each the three items in your example by using Excel's conditional logic and randomization function (e.g., to select one from 15/35/55 and one from man/woman, etc) . Take a look at that section and the accompanying sample and see if you can get the gist of that functionality. If it seems in the ballpark let us know and we could elaborate on how to proceed.

[lpike]

Hi Jarvis,

Thanks for the advice, I've had a look at Excel's conditional logic and it does what I'm hoping for- I just need to find out how to transfer it to Medialab now. The formulas I've used are:

=RANDBETWEEN(1,3)

=LOOKUP(A45,{1,2,3},{"A PHYSICAL DISABILITY","A LEARNING DISABILITY","NO DISABILITY"})

=RANDBETWEEN(1,3)

=LOOKUP(A47,{1,2,3},{"PHYSICAL ABUSE","FINANCIAL ABUSE","INSTITUTIONAL ABUSE"})

etc, then

=CONCATENATE("You are working with a ",A50," with ",A46,". You suspect that ",A52," ",A48," is occurring. ")

which translates to a combination of the above, i.e.
You are working with a WOMAN with A PHYSICAL DISABILITY. You suspect that SEVERE PHYSICAL ABUSE is occurring.

Is there is way to link to the excel spreadsheet, or do you have to create a Medialab file with the above info in it?

Many thanks,

Lindsey

[Trevor Newell]

Hi Lindsey,

As Blair has noted, you'll want to use the 'advanced features' that MediaLab supports. See the following link for more information:
http://www.empirisoft.com/medialab/h...d_features.htm

MediaLab has the ability to harness the power of excel through 'responses.xls'. This workbook allows a user to modify their experiment to communicate with excel during each screen. When a response is created, it can be written and passed through excel's coniditional logic - which will then allow for complex variables and skip patterns to be written and occur, respectively.

Take a look at the above link and try out the advanced features sample. These items should be able to help point you in the right direction. Let me know how it goes!

Regards,
Trevor

[lpike]

Thanks for the prompt response!
I've had a stab at it and have managed to get the calculated values working with a few testers, but not on my actual experiment. The problem is that rather than getting the content of the vignette (named "vig" on the responses.xls worksheet) I just get <vig> displayed in the instructions. But I get reports etc so excel and Medialab are talking, just not quite enough...!
Ideas of what might be going wrong are:

-I've had to add in a few extra rows- all named as calculated variable- that aren't features in the experiment at all (i.e. aren't dependent on responses from participants). I need them there to carry out the conditional logic in excel in order to create the vignettes- I've looked but the formula can't be used in a 3D reference (i.e. from another sheet) so have to be on the same one as the experimental and calculated variables (i think!)

- The contents of the cell is quite a number of words, is there a restriction on how many it can translate?

- Do calculated values need to be dependent on past responses of participants or can they be independent stimuli?

Thanks again- have attached experiment file for your perusal..

Lindsey

[Trevor Newell]

Hm. After taking a chance to look at your files, I noticed that the file 'responses.xls' had null values on row 14. This may have been the issue. I've updated your excel spreadsheet and have attached it to this post - let me know if the modification solves the problem.

- The contents of the cell is quite a number of words, is there a restriction on how many it can translate?

I don't believe so.

- Do calculated values need to be dependent on past responses of participants or can they be independent stimuli?

They should be allowed to exist as independent stimuli.

[lpike]

hmm.. the instruction 1 and 2 page is just blank now; it doesn't even say "here is scenario1 <vig> " ...! I've tried moving the zip file out of the experiment file but no change- help! I've attached the zip file again. Thanks!

[lpike]

Hello again,

I'm back on the case after a little break, did you have a chance to look at the above spreadsheet to see what could be the problem?

Thanks,
Lindsey