Anybody else as excited as me for the polytomous scoring on the new Bayley-4? Anybody?
Maybe I need to get out more...
Maybe I need to get out more...
-
CycleTrack and SDN are teaming up to make medical school admissions more transparent and accessible! Read the announcement to learn how we’re supporting the future of CycleTrack’s free application-tracking tools .
Polytomous Scoring!
Started by ClinicalABA
Get help with your application
Use all the free resources available to you from SDN: articles, guides, expert advising, forums discussions, and school research.
There are some initiatives to use IRT in existing neuropsych measures within a larger overall battery. That's something I'm excited about. Potential to cut test admin time quite a bit, while maintaining sensitivity.
The behavior analyst in me got excited that you were talking about measuring interresponse intervals, with anticipation of a lively debate regarding the merits of logarithmic transformations of the data. Alas, I'm guessing you are talking about Item Response Theory, right?
We're such friggin' nerds.
Ha, yeah, item response. Honestly, something we should have embraced as a field about 2 decades ago when it comes to assessment.
Not something I know that much about. I'll have to do some research on it. I rely mainly on criterion-based measures for diagnoses with my population, and use the norm-referenced stuff very cautiously (what exactly is intelligence in a 2 year-old?) vs. performance on that specific day. Always looked at things like the Bayley as a measure of where the kiddo currently is in their development of language and cognitive skills vs. some measure of an underlying trait. Still, always looking for more info on how things can be done better.
There has been some great work coming out recently with personality assessment and variable/conditional test administration stuff as well. It's very exciting and I'm glad to see it happening.
Links?
Would also love to see some of the recommendations anyone has on solid IRT (Item Response Theory is what I mean @ClinicalABA 🙂 ) that y'all have!
This x100. CTT was misguided from the start. IRT isn't perfect, but a step in the right direction. It also applies beyond just <traditional> assessment. CTT shapes a lot of our lab science. Picture a paradigm using video clips to elicit emotions. What do we do? Show a dozen per category and average ratings across them. Might we be losing some (potentially very useful) information there? What about neuroimaging? We average across 100 stimuli and assume trial-by-trial variability is just noise...usually without even checking that assumption.
Not to change the topic, but this and bayesian approaches are things that have pissed me off about the field for years. And yet I rely almost exclusively on CTT and frequentist approaches...because journals hate it when I don't and we lack the tools to easily do otherwise without investing extensive time.
....I have literally never heard of the Bayley-4 before. I probably don't belong here.
They're infant development scales. I've never administered one, but have seen one. I tend to believe that people don't exist until they hit 18. Makes my clinical world easier.
It's for ages 16 days through 42 months, so very limited and specialized age range. Contains cognitive, language, and motor scales for direct testing (norm referenced) as well as some parent rating scales for social/emotional and adaptive behavior. 4th edition releases this month. I think its a lot of fun to administer, but it is pretty heavy on duck related items (seriously, ducks are used to assess imitating actions with objects,grouping by color, grouping by sizes, pattern completion, and discriminating objects by mass).
I do find myself frequently "explaining away" low cognitive scores. For kiddos in the 30+ month range, several cognitive scale items are related to pretend, imaginative, and multi-scheme play, and still more have a strong receptive language component. It shouldn't be a surprise that kiddos suspected to be on the autism spectrum don't always do to well on these types of tasks. When I look at the distribution of standard scores for my population, the curve has the same shape and variance as the normal distribution, but the mean is somewhere around 85-90.
ETA- the Bayley-3 cognitive scale has, imho, the best test item ever- using a spoon as a lever to launch a rubber duck into the stratosphere. I will admit that I usually stray a little from protocol and give the spoon a bit of a Uri Geller pre-bend to improve the launch angle for maximum effect! If the kiddo doesn't giggle at that one, I get concerned!
Last edited:
