search

IALSA / IALSA-2015-Portland

Hub to accompany IALSA 2015 workshop at Portland, OR, Feb 22-25, 2015
GNU General Public License v2.0
2 stars 0 forks source link

Unidentified measures in EAS #157

Open andkov opened 7 years ago

andkov commented 7 years ago

@andreazammit ,

The new models you have uploaded contain a few measure that either new or spelled differently. Please help us identify the following measures:

Thank you!

andreazammit commented 7 years ago

@andkov

Vocabulary I believe used to be called waisvocab? I think I put it in as such in the PEK-cognition models?

I included a new measure in the models in case you wanted more stats on memory. I never used this before and I don't think any other studies have it, so feel free if you want to use it or not.. It's an episodic memory test.

Thanks so much, Andrea

andkov commented 7 years ago

@andreazammit , thanks! sorry, do you remember any details about freerecall measure? is it words? characters? We do have some episodic memory tests, but you are right, not freerecall. Here are our episodic memory measures: image anything you recognize? @ampiccinin , should it be grouped under prose recall total?

andreazammit commented 7 years ago

@andkov the full name of the test is called the Free and Cued Selective Reminding Test. It is a list of words. So it is word recall of a list of about 15 words, but it not prose like Logical Memory.

andkov commented 7 years ago

@andreazammit , thanks a lot!

@ampiccinin , I don't think we have any similar tests in our https://github.com/IALSA/IALSA-2015-Portland/blob/master/manipulation/rename-classify-rules.csv . Any suggestions on how to classify it within our existing taxonomy? Right now i'm inclined to add a new measure called freerecall under episodic memory domain. Please suggest a better alternative if you see it.

ampiccinin commented 7 years ago

@andkov @andreazammit

I don't think there are any other tests with selective reminding, but MAP (CERAD), ILSE, ELSA and HRS all have immediate recall tests that would otherwise be most similar.

ampiccinin commented 7 years ago

@andkov - also - just noticed an error in the taxonomy. The following should all be "semantic memory":

cognitive knowledge episodic memory info information information info literal TRUE
cognitive knowledge episodic memory information information information info
cognitive knowledge episodic memory waisgeneralknowledge waisgeneral information waisg

... and the following "episodic":

cognitive memory semantic memory delayedrecall word_de word list delayed word_de alt to cognitive memory semantic memory delayedrecall word_de word list delayed word_de alt to cognitive memory semantic memory recalld word_de word list delayed word_de alt to
cognitive memory semantic memory wordlistdelay word_de word list delayed word_de cognitive memory semantic memory wordlistrecog word_rec word list recognition

I'm also a bit confused about how the prose memory is listed as "short-term memory" (immediate) and "episodic" (delayed).

also, this one could just be "recognition memory", though I think it is the only one cognitive memory semantic memory wordlistrecog word_rec word list recognition

After our recent conversation about what this one is, I would now categorize it as "working memory": cognitive memory short-term memory digitbackwardtotal digit_tot digit span total (and edit this to be digitB_tot and "digit span backward total"

I hope I got them all!

andkov commented 7 years ago

freerecall is now classified under short-term memory and will serve as an alternative to word list alternative test: image

andkov commented 7 years ago

After our recent conversation about what this one is, I would now categorize it as "working memory": cognitive memory short-term memory digitbackwardtotal digit_tot digit span total (and edit this to be digitB_tot and "digit span backward total"

digitspanbackward is now placed withing the working memory domain. image

andkov commented 7 years ago

also, this one could just be "recognition memory", though I think it is the only one cognitive memory semantic memory wordlistrecog word_rec word list recognition

That's right, we don't have a domain recognition memory. I am working on the updating the outcome space graph to which this classification is relevant. Let's examine the item wordlistrecog in its context to be convinced that another single-item domain is required (single-item domains look ugly and take up space, but i agree theoretical consideration are more importent). For the moment, i'm leaving it in episodic memory: image

andkov commented 7 years ago

I'm also a bit confused about how the prose memory is listed as "short-term memory" (immediate) and "episodic" (delayed).

Perhaps you've glanced at the wrong row? There are only two raw items dealing with prose : prose and proserecall, both classified under short-term memory. The one immediately below those, bostonstorydelay, was placed under episodic memory, because it was delayed as opposed to immediate. Do you think we should move bostonstorydelay to short-term memory? image

Here are other delayed measures that all got classified under episodic memory: image

andkov commented 7 years ago

The following should all be "semantic memory": cognitive knowledge episodic memory info information information info literal TRUE cognitive knowledge episodic memory information information information info cognitive knowledge episodic memory waisgeneralknowledge waisgeneral information waisg

image

andkov commented 7 years ago

and the following "episodic": cognitive memory semantic memory delayedrecall word_de word list delayed word_de alt to cognitive memory semantic memory delayedrecall word_de word list delayed word_de alt to cognitive memory semantic memory recalld word_de word list delayed word_de alt to cognitive memory semantic memory wordlistdelay word_de word list delayed word_de cognitive memory semantic memory wordlistrecog word_rec word list recognition

image

ampiccinin commented 7 years ago

Perhaps the best way to sort this out is to list everything except the word_de that you current have as "semantic" memory as "knowledge" instead. Then add in some of what you have in episodic to knowledge as well (bnt, nart, synon, vocabulary/waisvoc, info, waisg) (i.e., eliminate the semantic and episodic domains. Part of why this is problematic is because they are nomenclature from a different cognitive taxonomy.)

Then, the digits forward can remain as short-term memory and digits backward as working memory.

Finally, the word-list and story/prose measures can all be listed under "free-recall". Ideally we would distinguish between immediate and delayed, as well as between word-lists and prose (prose, bstory and logic), but we can drop those additional distinctions if necessary to simplify.

andkov commented 7 years ago

@ampiccinin , I have adjusted the outcome space graph to represent the statistical information from the models. I think this display will help out scout the results in search of a story. Here's an example print-domain-map-1 You can review the entire pulmonary domain map, the live version of the report is also available at #152.

I will implement your suggestions above in the next iteration of the graph (perhaps, you'll have more after you play with it).

ampiccinin commented 7 years ago

Interesting use of that graph!!

(just noticed in overall outcome space that ILSE seems to have piccomp listed twice. I think it goes with "line orientation" better than the fluid reasoning measures. thanks!)

andkov commented 7 years ago

(just noticed in overall outcome space that ILSE seems to have piccomp listed twice.

ah. thank you for catching this. I'll make a note of it.