WISC-V Interpretation Help

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

EKS9

Full Member
2+ Year Member
Joined
May 8, 2020
Messages
14
Reaction score
11
Question: Looking at pairwise comparisons, what does it mean if there is a significant difference between an individuals scores on the VCI and WMI, and also between the VSI and WMI?? What could this tell us about the client?

Thanks,
A Trying First Year lol

Members don't see this ad.
 
In the real world, we don’t really read into such differences unless they are germane to the referral question.
 
  • Like
Reactions: 1 users
Members don't see this ad :)
Question: Looking at pairwise comparisons, what does it mean if there is a significant difference between an individuals scores on the VCI and WMI, and also between the VSI and WMI?? What could this tell us about the client?

Thanks,
A Trying First Year lol

It tells me you are too test bound.
 
  • Like
Reactions: 1 users
It means there’s a difference between how the person performed on the tests comprising those indices. If the p value of the difference is low, the comparison is statistically unlikely to be due to chance (in a perfect world).

Im assuming this is a homework assignment. However, as others have said, it could mean nothing or a great deal. Tests only have meaning in context of the patient and their presenting issue. Same with the significance of the comparison, it only means something in the right situation.

There’s a saying in medicine: treat the person not the lab values. The same applies here: tests are useful to understand what’s going on, but they are not useful in isolation.
 
Last edited:
  • Like
Reactions: 2 users
It means there’s a difference between how the person performed on the tests comprising those indices. If the p value of the difference is low, it’s statistically unlikely to be due to chance (in a perfect world).
Oh dear. No, p values are not like an inverse of the probability of a difference being due to chance.

It's much worse than that.

The question we generally want answered is something like 'are these two (or more) things genuinely different or are they really the same?'

The question a p-value answers is 'Assuming these two things are underlyingly identical, and a certain set of parameters describes the distribution of their apparent values, and we repeated this experiment/procedure an arbitrarily large number of times, how many of those repetitions would give us differences at least as extreme as the ones we observe here?'

This reasoning is exactly as tenuous as it sounds.
 
  • Like
Reactions: 1 users
Oh dear. No, p values are not like an inverse of the probability of a difference being due to chance.

It's much worse than that.

The question we generally want answered is something like 'are these two (or more) things genuinely different or are they really the same?'

The question a p-value answers is 'Assuming these two things are underlyingly identical, and a certain set of parameters describes the distribution of their apparent values, and we repeated this experiment/procedure an arbitrarily large number of times, how many of those repetitions would give us differences at least as extreme as the ones we observe here?'

This reasoning is exactly as tenuous as it sounds.

I should clarify as i may have been imprecise - I’m speaking specifically about the test index comparisons that come from these intelligence tests, not general stats. When you compute the indices, there’s a significance test for pairwise index comparisons as well as a base rate estimation of the magnitude of the difference. Both pieces of info are needed to conclude whether it is a real finding or just “due to chance” as I said above. Just leaving it at “p value” was me trying to envision the computer printout rather than being clear.

So in this case, a small index-index comparison magnitude (say, a few points) is exceedingly common in the general population and is unlikely to meet statistical significance for the pairwise comparison. Thus, my shorthand explanation for this is that the difference can be better explained by chance (based on the reliability of the test, normal small variation in individual subtest performance across a battery of tests) than a true ability difference.

So, My bad. It’s more precise for me to say “if the index comparison is insignificant/common in the general population, it’s likely an insignificant finding which may be due to chance”. It has nothing to do with p values in general

Edit: took me like 5 tries to word it how I wanted. Infant sleep regressions are no joke.
 
Last edited:
This is even more complicated by findings about scatter. I used to be a huge pattern of strengths and weakness fan, but more and more im turning into a FSIQ fanboy.
 
  • Like
Reactions: 1 users
Top