Millisecond Forums

Counterballanced Repeated Measure (separate sittings)

https://forums.millisecond.com/Topic5078.aspx

By IOPsyc - 10/27/2010

Hi, I've searched the documentation and forum posts, but I'm still not sure whether this is possible. I'd like to use the web version to conduct my research:


I need my participants to complete a 60 likert-item survey twice, a week apart, under separate instructions. I also need to counterballance, so I'd like to randomly assign them to do either ExpA then ExpB or ExpB then ExpA. I'll ask the participants for the first couple of letters of their first and middle names, to match their responses under each condition.


Ideally, it would be great if I could get inquisit to first ask for the identifying characters: if not previously tested, then assign them randomly (alternating participants would suffice) to do ExpA or ExpB; if previously tested, direct to them to the Exp they did not do the first time.


I assume that this is probably a bit of pipe dream, so...


Alternatively, can I control how participants are assigned to each group, by sending separate web links for ExpA and ExpB (and without needing 2 licences)?



Thanks very much for any support you can provide,



Ben

By Dave - 10/28/2010

Hmm, I'd try with properly configured <batch> elements and a matching URL parameter.


<batch>
/ file = "ExpA.exp"
/ subjects = (1 of 2)
</batch>

<batch>
/ file = "ExpB.exp"
/ subjects = (2 of 2)
</batch>


I'd set up the launch page to assign subjects based on an URL parameter. Then, for phase one,  I'd send the first half of participants a link to the experiment with a parameter like group=1, the other half gets group=2. For phase two, particpants get the other group assigned. Seems viable, because the numerical subject is not the primary identifier in your case.


Hope this helps,


~Dave