|
Ali
|
|
|
Group: Forum Members
Posts: 26,
Visits: 194
|
Hello,
1- Include element: When I finish the last page, the survey takes me back to the very first page as if I am going to repeat the whole survey again (so participants may not realize that they have finished). How can I shut the program after the survey is completed?
2- Progress status: My survey contains around 50 and some participants might get bored half the way through. Is there a way to show progress status, remaining percentage, or anything equivalent that will encourage the participants continue?
3- Cancelling output files: I prepared a "welcome" page in a separate block so that I can counterbalance the other blocks. Although it does as a question, there is no real information I need in this welcome block and so I used the / recorddata = false attribute. However, I still get a separate output file for this block. In order to avoid cluttering, can I do something so that I don't get any output for this block?
4- IAT reliability: Is the researcher expected to report some form of reliability for the IAT?
5- Scales: I need to recode negatively worded survey items and check the Cronbach reliability before the analysis. However, because the items are generated randomly with noreplace, I can't locate each through stimulusnumber (I have to look at the trialcode also). If I number my items sequentially in the Item elements so that the first item in the first scale is 1 and the last is, say, 45, can I preserve this numbering in the output file? Or is there any other way to locate each scale easily?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
#1 cannot be answered without you providing the actual code.
#2 There are many options to convey progress. They've been discussed before, e.g. here: https://www.millisecond.com/forums/FindPost8176.aspx
#3 also cannot be answered without actual code.
#4 is a methodological question unrelated to Inquisit. I'd recommend reviewing the literature on IAT reliability.
#5 is also unclear without any code. Survey questions usually involve neither <item> elements nor any form of noreplace, so it is not obvious what exactly you are referring to.
|
|
|
|
|
Ali
|
|
|
Group: Forum Members
Posts: 26,
Visits: 194
|
Thanks Dave for your reply. I attached the files. As for #5, You'll see in explicit.iqx that under <item> elements the survey questions are numbered sequentially. Is there a way to show this numbering in the output file? (This is only because it will allow me to locate items easily [to recode... etc.], and so please let me know if there are other options.)
Thanks.
Ali
PS. If you happen to spot something terribly wrong when you open the files, please let me know.
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
Re. #1:
> 1- Include element: When I finish the last page, the survey takes me back to the very first page as if I am going to repeat the whole survey again (so participants may not realize that they have finished). How can I shut the program after the survey is completed?
Which survey / last page are you referring to exactly? Note that <include> effectively pastes contents of other scripts into the "main" script containing the <include> element. Several of the scripts you <include> have their own <expt> elements which will be added (and possibly run in addition to) the <expt> elements already included in the main script. For example:
<expt> / subjects = (1 of 4) / blocks = [...; 2=explicitintro; 3=explicit; 4=explicitend; ... ] [...] </expt>
in 'combine.iqx' (the main script) runs the above three elements contained in 'explicit.iqx'. However, 'explicit.iqx' contains yet another <expt> element, which also runs the same three elements:
<expt> / blocks = [1= explicitintro; 2=explicit; 3=explicitend] </expt>
You sould remove the latter from 'explicit.iqx'. There is an analogous issue with respect to 'implicit.iqx'.
I also wonder why you use the <include> element to combine your various tasks / scripts. Using <batch> may be preferable and much easier to understand.
Re. #3: > 3- Cancelling output files: I prepared a "welcome" page in a separate block so that I can counterbalance the other blocks. Although it does as a question, there is no real information I need in this welcome block and so I used the / recorddata = false attribute. However, I still get a separate output file for this block. In order to avoid cluttering, can I do something so that I don't get any output for this block?
You can do that by running the respective <surveypage> via a <block> element instead of a <survey> element. A <surveypage> is a type of <trial>, just like a <survey> is a special type of <block>. See the attached version of 'welcome.iqx' for details.
Re. #5:
Since your scales have 7 items each, you can easily compute a "global" item number ranging from 1 to 42 along the lines of
<values> / scaleitemnumber = 0 </values>
[...]
<likert ATESP> / ontrialend = [values.scaleitemnumber = 0*7 + text.atespqs.currentitemnumber] / stimulusframes=[1=ATESPqs] / anchors=[1="Disagree"; 4="Unsure"; 7="Agree"] / numpoints=7 / position=(50, 80) </likert> <likert ETHNO> / ontrialend = [values.scaleitemnumber = 1*7 + text.ethnoqs.currentitemnumber] / stimulusframes=[1=ETHNOqs] / anchors=[1="Disagree"; 4="Unsure"; 7="Agree"] / numpoints=7 / position=(50, 80) </likert> <likert ASSIM> / ontrialend = [values.scaleitemnumber = 2*7 + text.assimqs.currentitemnumber] / stimulusframes=[1=ASSIMqs] / anchors=[1="Disagree"; 4="Unsure"; 7="Agree"] / numpoints=7 / position=(50, 80) </likert>
[...]
<likert ATLE> / ontrialend = [values.scaleitemnumber = 5*7 + text.atleqs.currentitemnumber] / stimulusframes=[1=ATLEqs] / anchors=[1="Disagree"; 4="Unsure"; 7="Agree"] / numpoints=7 / position=(50, 80) </likert> See the attached version of 'explicit.iqx' for details. You then ought make sure to log values.scaleitemnumber to the data file (i.e. add it to the <data> element in your main script):
<data> / file = "RAWDATA.iqdat" / columns = [date, time, subject, group, blockcode, blocknum, trialcode, trialnum, response, correct, latency, stimulusnumber, stimulusitem, expressions.da, expressions.db, expressions.d, values.scaleitemnumber] / separatefiles = false </data>
|
|
|
|
|
Ali
|
|
|
Group: Forum Members
Posts: 26,
Visits: 194
|
Hi Dave,
You suggested that I use batch instead of element. Does the batch allow me to counterbalance the compatible and incompatible blocks in, say, an IAT?
Also, in order to minimize cluttering, how can I remove latency responses from a demographic survey?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
> Does the batch allow me to counterbalance the compatible and incompatible blocks in, say, an IAT?
Has nothing to do with the <batch> in this particular case. If you look at any current IAT script, you will find that it's the groupnumber (odd vs even) that gets passed to the IAT which determines the block order.
Run
<batch> / file = "somescript.iqx" / file = "IAT.iqx" </batch>
supplying an odd groupnumber and you'll get one block order in the IAT. Supply an even groupnumber and you'll get the other order.
> Also, in order to minimize cluttering, how can I remove latency responses from a demographic survey?
In general: Edit the script's <data> element. For details see the language reference topics for the <data> element and its /columns attribute. Two important points to note:
#1: Data output for *<survey>* elements cannot be customized. #2: In your current setup using <include>, the one <data> element in the main script applies globally. I.e. you cannot selectively enable latency recording for, say, the IAT portion but disable it for the <likert> questions portion.
|
|
|
|
|
Ali
|
|
|
Group: Forum Members
Posts: 26,
Visits: 194
|
Thanks Dave. I attached the new files with the batch element as you suggested. As you can see, for some reason when you finish the bit (demographics) you are taken back to the beginning.
(Also, after replacing include with batch, I noticed that the transition between scripts is now less smooth. You see the program starting between scripts. Is this normal or I did something wrong?)
Finally, regarding counterbalancing, yes you're right. The IAT compatible vs. incompatible conditions are counterbalanced through groupassignment. I have four conditions, adding explicit before vs. after. I imagine it would be much easier for the analysis if I can assign these four conditions to four group numbers (e.g., 1=explicit first & compatible, 2=explicit first & incompatible, 3=explicit last & compatible, 4=explicit last & incompatible). I think in this case I will enter the subject number as usual and then manually enter the group number (1-4). Is this possible?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
> As you can see, for some reason when you finish the bit (demographics) you are taken back to the beginning.
That's what you set up your batch elements to do:
<batch> / file = "1 welcome.iqx" / file = "2 explicit.iqx" / file = "3 implicit.iqx" / file = "4 demographics.iqx" </batch> <batch> / file = "1 welcome.iqx" / file = "3 implicit.iqx" / file = "2 explicit.iqx" / file = "4 demographics.iqx" </batch>
will run the scripts in the first <batch> element and then those in the 2nd one. Specify proper /subjects and /groupassignment attributes to change that. That is something along the lines of:
<batch> / subjects = (1,2 of 4) / groupassignment = groupnumber
/ file = "1 welcome.iqx" / file = "2 explicit.iqx" / file = "3 implicit.iqx" / file = "4 demographics.iqx" </batch> <batch> / subjects = (3,4 of 4) / groupassignment = groupnumber
/ file = "1 welcome.iqx" / file = "3 implicit.iqx" / file = "2 explicit.iqx" / file = "4 demographics.iqx" </batch>
> You see the program starting between scripts. Is this normal or I did something wrong?
It's normal.
> I think in this case I will enter the subject number as usual and then manually enter the group number (1-4). Is this possible?
No. The subjectnumber and groupnumber you supply to the batch script apply to all the scripts run by it.
|
|
|
|