By lovetolearn - 3/13/2016
Hi,
I am trying to write a study based on an adaptation of the Age IAT from the test library (http://www.millisecond.com/download/library/AgeIAT/). What I am trying to run is based on a 2 (Good_A vs. Good_B) X 2 (Young_A vs. Young_B) between-subject design. So all participants will still see same words related to "Bad" and same pictures related to "Old". But they will be randomly assigned to see just one of the four conditions:
1) words related to "Good_A" and pictures related to "Young_A";
2) words related to "Good_A" and pictures related to "Young_B";
3) words related to "Good_B" and pictures related to "Young_A";
4) words related to "Good_B" and pictures related to "Young_B".
I wonder how I may revise the coding of the Age IAT script, to achieve such a design.
Any advice would be greatly appreciated.
|
By Dave - 3/13/2016
You can use the <variables> element to implement the four conditions and substitute items as needed, here the <item> or <text> elements representing good A and B as well as young A and B respectively.
https://www.millisecond.com/support/docs/v4/html/language/elements/variables.htm
|
By lovetolearn - 3/14/2016
Thanks so much for your speedy response Dave - you are awesome!
I notice the following notice on the link you provided "This element is obsolete and is included for backwards compatibility only." So, if I am using Inquisit 5, may I still use this element or should I use https://www.millisecond.com/support/docs/v4/html/language/attributes/groupassignment.htm instead?
Thanks again! Sorry I just realized that I should have posted my question under Inquisit 5 rather than Inquisit 4.
|
By Dave - 3/14/2016
The use of <variables> is not in general recommended anymore -- but it still works and it is useful in select cases such as this one.
An alternative would be to use conditional <include> elements instead:
https://www.millisecond.com/forums/FindPost15640.aspx
|
By lovetolearn - 3/30/2016
Hi Dave,
I haven't got lucky by using <variables> to set up the design I want. However, I wonder if I may just use the AgeIAT file to develop each of the four conditions into: AgeIAT_condition 1.iqx, AgeIAT_condition 2.iqx, AgeIAT_condition 3.iqx, and AgeIAT_condition 4.iqx. Then generate a fifth file:
<batch>
/ file = "AgeIAT_condition 1.iqx"
/ subjects = (1 of 4)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 2.iqx"
/ subjects = (2 of 4)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 3.iqx"
/ subjects = (3 of 4)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 4.iqx"
/ subjects = (4 of 4)
/ groupassignment = groupnumber
</batch>
By running the fifth file on Inquisit Web, can I still get participants randomly assigned to one of the four conditions? The other question is I just ran test monkey on this, and it asks me to enter the subject id and the group number - do I need to turn off the subject and group codes dialog box in order to ensure the random assignment? How can I achieve that?
Thanks heaps.
|
By Dave - 3/30/2016
Yes, you can use <batch> elements as well. You should note, though, that each IAT *itself* contains two conditions (the block order is counterbalanced according to odd vs. even groupnumber). I.e., you need to make sure that each of your IATs is fed both odd and even groupnumbers. That is not the case with your current <batch> elements. You'll want to change them to something along the lines of
<batch>
/ file = "AgeIAT_condition 1.iqx"
/ subjects = (1,2 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 2.iqx"
/ subjects = (3,4 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 3.iqx"
/ subjects = (5,6 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 4.iqx"
/ subjects = (7,8 of 8)
/ groupassignment = groupnumber
</batch>
> By running the fifth file on Inquisit Web, can I still get participants randomly assigned to one of the four conditions?
Yes, of course.
> do I need to turn off the subject and group codes dialog box in order to ensure the random assignment?
No. You simply set your web experiment to generate random subject and group numbers (those are the default options anyway) when you upload the scripts and configure the web experiment's settings. There will be no box that prompts you / your participants to enter anything.
|
By lovetolearn - 3/30/2016
Thanks so much for your advice Dave. You are brilliant:)
May I ask a follow-up question on adding survey questions before and after the study where participants are randomly assigned to complete one of the four conditions?
If I have developed 4 survey questionnaires: pre_survey_male.iqx, pre_survey_female.iqx, post_survey_male.iqx, post_survey_female.iqx (i.e., participants need to answer one survey prior to the study, and a second survey after the study, and the contents of the surveys vary depending on their genders), may I organize these .iqx files using <batch> again, so that males only get questions on pre_survey_male.iqx and post_survey_male.iqx, and females only get questions on pre_survey_female.iqx and post_survey_female.iqx? Otherwise, is there any way to achieve this on Inquisit Web?
|
By Dave - 3/30/2016
You should set up two *.iqx files -- one for your pre-survey, one for your post-survey, with each containing the pages / questions for both male and female respondents.
At the very start of both the pre-survey and the post-survey, have the participant indicate their gender. Then use /skip attributes in the <surveypage> elements to only administer the pages / questions applicable to the gender indicated.
Run the pre- and post-survey scripts via your <batch> elements as usual.
|
By lovetolearn - 5/31/2016
Hi Dave,
Following on your advice on this post:
"
Yes, you can use <batch> elements as well. You should note, though, that each IAT *itself* contains two conditions (the block order is counterbalanced according to odd vs. even groupnumber). I.e., you need to make sure that each of your IATs is fed both odd and even groupnumbers. That is not the case with your current <batch> elements. You'll want to change them to something along the lines of
<batch>
/ file = "AgeIAT_condition 1.iqx"
/ subjects = (1,2 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 2.iqx"
/ subjects = (3,4 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 3.iqx"
/ subjects = (5,6 of 8)
/ groupassignment = groupnumber
</batch>
<batch>
/ file = "AgeIAT_condition 4.iqx"
/ subjects = (7,8 of 8)
/ groupassignment = groupnumber
</batch>
> By running the fifth file on Inquisit Web, can I still get participants randomly assigned to one of the four conditions?
Yes, of course.
> do I need to turn off the subject and group codes dialog box in order to ensure the random assignment?
No. You simply set your web experiment to generate random subject and group numbers (those are the default options anyway) when you upload the scripts and configure the web experiment's settings. There will be no box that prompts you / your participants to enter anything.
"
I am currently trying to integrate this Inquisit Web study into qualtrics for data collection. By reading http://www.millisecond.com/support/docs/v4/html/howto/interopsurveys.htm, I see such an instruction "To configure Inquisit to retrieve the subject id from the query parameter, simply run through the webscript registration wizard. When asked how to generate subject ids, select the "Query Parameter" option and specify the name of the parameter (in this case, the name is "respondentid"). That's it, Inquisit will now extract this subject number from the url and record it in the data file."
Given your advice "You simply set your web experiment to generate random subject and group numbers (those are the default options anyway) when you upload the scripts and configure the web experiment's settings. There will be no box that prompts you / your participants to enter anything", does that mean if I follow "When asked how to generate subject ids, select the "Query Parameter" option and specify the name of the parameter (in this case, the name is "respondentid")", it will ruin the intended study setup such as randomization.
Sorry for the long post - I look forward to your advice!
|
By Dave - 5/31/2016
> [...] does that mean if I follow "When asked how to generate subject ids, select the "Query Parameter" option and specify the name of the
> parameter (in this case, the name is "respondentid")", it will ruin the intended study setup such as randomization.
No, it does not mean that. The condition assignment in your scripts (the <batch>es as well as the IATs) is based on the *group id*. The group id has to be numerical for that to work, i.e., you should leave it at either random or sequential generation.
You can safely select the query parameter option for the *subject id*. The subject id doesn't play any role in the condition assignment and thus you can safely use the alphanumerical code supplied by Qualtrics. Nothing will be ruined.
|
By lovetolearn - 5/31/2016
Thanks so much for your speedy response Dave. You are always so brilliant!
|
By lovetolearn - 6/2/2016
Hi Dave,
When the Inquisit Web study is redirected into qualtrics for some post-Inquisit survey, does the Inquisit tell the participants the instructions on how to uninstall the Inquisit plugins before being redirected to qualtrics? If no, should I tell participants how to uninstall the Inquisit plugins at the end of the qualtrics survey? Are there standard instructions I may copy and paste there on the uninstallation?
Thanks once again!
|
By Dave - 6/2/2016
No. When you redirect back to (some other) Qualtrics survey, that's where participants will be sent directly -- they will not see the uninstall instructions before that. To provide them with those instructions you can link them directly to https://www.millisecond.com/support/uninstall.aspx at the end of your Qualtrics post-survey or copy the instructions on that page.
|
|