Psychology related question

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

Doctor_Strange

Full Member
7+ Year Member
Joined
Mar 28, 2015
Messages
955
Reaction score
604
Hi,

Is it appropriate to group Digit span forward and backward tests together, or do they need to be separated if I am doing a meta-analysis?

Thanks

Members don't see this ad.
 
Short answer, depends.

I got the answer I needed through some review of literature. But I now have a bit more broader question: is it possible to group more diverse tests together such as Digit Span and, say, N-back tasks? I ask because I see in some papers that a Hedges' g test was conducted which allowed for such grouping.
 
Members don't see this ad :)
It is possible, but limits generalizability and will muddy the interpretation. These tasks differ fairly greatly as working memory is far from a singular construct as measured by objective testing. Additionally, you will run into fairly significant norming differences depending on how diverse the populations under study are on a variety of demographic variables. So, yes, but there are a TON of caveats.
 
  • Like
Reactions: 1 user
It is possible, but limits generalizability and will muddy the interpretation. These tasks differ fairly greatly as working memory is far from a singular construct as measured by objective testing. Additionally, you will run into fairly significant norming differences depending on how diverse the populations under study are on a variety of demographic variables. So, yes, but there are a TON of caveats.

Interesting.

If you could enlighten me, take the following example (not my data, just for teaching purposes):

upload_2016-11-30_16-47-47.png


In my own data set, I now must distinguish between acute vs chronic exercise studies. As such, one of my subgroups which exclusively included N-back tasks as the measurement tool for working memory, now changes from 3 studies down to 1 (the one that is only chronic) which now, I believe, makes this subgroup inappropriate since it is now really just one study. So instead of having several subgroups similar to the image above, I think I may be forced to measure all chronic tests together even though they use different tools to measure WM. Thoughts? I am pretty unsure of what to do. I do not want to remove the meta-analysis from my review, but I do not know what is appropriate to do at this stage.
 
Tough call, no easy answer, either way you are running into a serious methodological limitation. Depends on what the reviewers say to some extent, as that is the barrier to publication at this point.
 
  • Like
Reactions: 1 user
Tough call, no easy answer, either way you are running into a serious methodological limitation. Depends on what the reviewers say to some extent, as that is the barrier to publication at this point.

Thank you for your insight. I am scheduled to have a phone call with my research adviser from my alma mater on Friday. It's tough since I already have graduated, I cannot just go to office hours and am relying on the internet to read about this topic. Also, neither of the 2 reviewer's commented explicitly on removing the meta-analysis. They just cited some concerns. I think, though, what you said is what I am feeling: I need to accept that my study now will suffer from less generalizability and methodological limitations, but hopefully not at the expense of not being published if I address and revise such concerns from the reviewers!
 
Top