Gains in fluid intelligence?

[moe]

After years of stagnant scores on tests of fluid intelligence I SEEM
to have finally made some improvement. The improvement may be a
result of education, single n back, juggling, etc. On the JCTI I was
previously stuck at 144 but today I scored 152!!. On the CFNSE I was
previously maxed out at 99.8ish and the other day I scored at 99.908
(balanced score of 8). I'm very excited, finally some
improvement :). I guess the results are confirmation of the
subjective changes that I've been noticing.

[Gwern Branwen]

On Thu, Dec 22, 2011 at 5:36 PM, moe <lindas...@bellsouth.net> wrote:
> After years of stagnant scores on tests of fluid intelligence I SEEM
> to have finally made some improvement.  The improvement may be a
> result of education, single n back, juggling, etc.  On the JCTI I was
> previously stuck at 144 but today I scored 152!!.

So you've been taking them repeatedly for years, and just now you
shifted 8 points.

I see...

--
gwern
http://www.gwern.net

[moe]

It seems as if power tests of fluid intelligence are a good way to
gauge cognitive improvement (the test-retest reliability seems high
with the good tests). My take on the situation is that I do not seem
to have become more intelligent logically (my logical abilities do not
seem to have improved) but I seem to have made nice improvements in
visual memory which has resulted in improvements in my pattern
recognition ability. It seems as if the combination of snb and
juggling is "working". Learning new stuff may have played a role (I
just finished up calculus II and General physics 1 this semester and
the majority of the topics were new material to me) but most of the
gains seem to be a result of juggling and snb. I'm pretty sure my
actual gf hasn't gone up 8 points but I think what's important here is
measurable improvement.
On Dec 22, 5:44 pm, Gwern Branwen <gwe...@gmail.com> wrote:

[ailambris]

http://www.nytimes.com/2011/12/23/science/pigeons-can-learn-higher-math-as-well-as-monkeys-study-suggests.html?_r=1&hp

[Gwern Branwen]

On Thu, Dec 22, 2011 at 7:19 PM, ailambris <aila...@gmail.com> wrote:
> http://www.nytimes.com/2011/12/23/science/pigeons-can-learn-higher-math-as-well-as-monkeys-study-suggests.html?_r=1&hp

Reminds me of "If a Lion Could Talk: Animal Intelligence and the
Evolution of Consciousness"
http://www.nytimes.com/books/first/b/budiansky-lion.html

--
gwern
http://www.gwern.net

[Mike]

Hi Moe, that's awesome! were the scores set with a standard deviation of 15? I will try those 2 tests. for comparison, how much do you get on iqtest.dk ? 

were you/are you/will you join a high IQ society at some point?

I always wondered how super high IQs look like.

btw does anyone know if there are good high IQ society journals/forums/websites still produced these days? I'm really curious. I read some old articles from high iq societies a while ago but I'd like to read more. I would be curious to "size" better the pattern behind high IQ writing style/subjects/ideas. --> ie what do they have in common when you read them a lot. maybe not much. some people say it's not interesting. I thought it was, on most occasions--but want to read more. I could also be stimulating, a good source of inspiration and information. in a way the TED talks are decent, they do that job--some extremely good ones pop up every now and then (some terrible ones too of course), but I'd like to read high iq insights more often.

snb? what is snb? --> sorry, lots of questions.

do you feel juggling works? it's worth it?

On Thu, Dec 22, 2011 at 5:36 PM, moe <lindas...@bellsouth.net> wrote:

--
You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.
To post to this group, send email to brain-t...@googlegroups.com.
To unsubscribe from this group, send email to brain-trainin...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/brain-training?hl=en.

[polar]

heh, it looks like this shift can be caused by flynn effect in your
case ;) but nevermind and enjoy!

[moe]

Yes the scores are set with a standard deviation of 15. I took
iqtest.dk a few years ago and scored in the low 130s (I don't do as
well on timed tests). I qualified for a few high iq societies after
doing well on the hoeflin power test and the get-y (144 sd 15 on these
two) but I'm not sure if I'll ever join one. I would feel like an
impostor if I joined and it would not feel too great to be around a
bunch of people that are much more intelligent than I am :). I prefer
groups like this one that are open to everyone that shares a common
interest. I'm also a member of physicsforums.com (which is a math/
physics forum that is also open to everyone). Overall I'm not really
super intelligent I'm just high in a few areas (gf, spatial, mental
arithmetic, math/science, etc). Verbally I'm above average according
to a few tests but not super high (probably around 120-130).

Snb is single n back. I tried dual n back last year and didn't gain
anything from it (if anything it may have hurt cognition). Single
position n back is the only variant that truly works for me. I'm
pretty sure that the juggling is working. Actually juggling and snb
may have a synergistic effect. Juggling basically has increased my
ability to deal with visual complexity. Deciphering complicated
graphs/diagrams in my math/science books has become considerably
easier after taking up juggling. I also feel more efficient when I
read (my brain seems to process the words with less effort than
before). The only downside to juggling (in my case) is that it seems
to slow my brain down a bit but I do not mind trading in a bit of
speed for more working memory capacity.

On Dec 23, 12:04 am, Mike <mikebk...@gmail.com> wrote:
> Hi Moe, that's awesome! were the scores set with a standard deviation of
> 15? I will try those 2 tests. for comparison, how much do you get on
> iqtest.dk ?
> were you/are you/will you join a high IQ society at some point?
> I always wondered how super high IQs look like.
>
> btw does anyone know if there are good high IQ society
> journals/forums/websites still produced these days? I'm really curious. I
> read some old articles from high iq societies a while ago but I'd like to
> read more. I would be curious to "size" better the pattern behind high IQ
> writing style/subjects/ideas. --> ie what do they have in common when you
> read them a lot. maybe not much. some people say it's not interesting. I
> thought it was, on most occasions--but want to read more. I could also be
> stimulating, a good source of inspiration and information. in a way the TED
> talks are decent, they do that job--some extremely good ones pop up every
> now and then (some terrible ones too of course), but I'd like to read high
> iq insights more often.
>
> snb? what is snb? --> sorry, lots of questions.
>
> do you feel juggling works? it's worth it?
>

[moe]

Well I'm definitely not "claiming" that score it's just convenient
that I improved on both of those tests right after I hit the snb/
juggling routine pretty hard. Maybe the actual shift in my case is
from the low 140's to the 145 or above category in gf. I think that's
a fair estimate but some of you may not agree, oh well.

[☉]

We've discussed quite a few of the findings in the literature to
death, and one of the points that I (as well as others) brought up is
that in order to establish a bona fide increase in a Gf measure that
excludes confounding variables (e.g. standard error, pretesting,
practice effects, and their various magnitudes), it is imperative that
one takes a battery of tests and not rely on a single performance
metric/test.

The JCTI, according to the claims of its author, is a very reliable
metric. The fact that your scores are so discrepant may put into some
doubt this touted reliability. Your CFNSE scores are so close to each
other that I'm doubtful of any marked improvement, as those
percentiles are only a few I.Q. points apart (in 15-point SD), and
such differences aren't uncommon. In general, your scores are in fact
remarkably close and comparable.

There are many ways to describe how someone gets better at a test
without actually improving in the area (not entirely) measured by the
test.

In the end, I'd answer your thread question with "mu".

Anyway, you might want to try iqtest.dk's Dansk test. (The Dansk
version is the hardest, and includes age-based norms - albeit somewhat
erroneously.)

argumzio

[Mike]

yes me too I find that test at iqtest.dk quite hard, thanks for sharing your score. last year I couldn't score higher than exactly 126 and 130 at it, and recently I actually scored only 122 :(. I've scored much higher before (in the 130s and up to 149, 156) (but also sometimes lower: 124, 130 I have to point out) on other tests that involved a more diversified set of questions, including "pattern matrices" but also vocab, spatial problem, mental math problems, etc. that type of stuff. I'll try the JCTI and CFNSE later.

I tried dual n back last year and didn't gain
anything from it (if anything it may have hurt cognition) 

what were your perceived losses in cognition after DNB?

how do you play SNB? strategies? by how much did you improve your SNB span (letters, numbers or sounds you're using?)

The only downside to juggling (in my case) is that it seems
to slow my brain down a bit but I do not mind trading in a bit of
speed for more working memory capacity. 

how did you measure that?

personally I noticed that I can't score in the 150ms and 160ms at visual reaction time like I used to do regularly on cog fun and other sites. my average is now in the 190ms 200ms (after months and months of DNB). I thought this was a coincidence and was due to either the cogfun website or my gear (new laptop and mouse).

could my reaction time really have slowed down? that would seem strange. a WM-processing speed tradeoff?

thanks for sharing your experiences!

JCTI

[The.Fourth.Deviation.]

If you dont train reaction speed, it will go down. mine does on
lumosity when i dont use the site for some time and only do DNB. Also,
some of that could be age related declines.

> JCTI- Hide quoted text -
>
> - Show quoted text -

[moe]

ha just figured out another item which brought my score up to 157 haha
(due to fatigue I didn't quite max out my reasoning on a few of the
items when I took the test yesterday). Anyway, I wish that the get-y
was still available so that I can use my second attempt but the author
took it down due to it being compromised. Any other untimed/power gf
tests that I could take?

[moe]

With dnb the reasoning aspect of my cognition may have taken a slight
hit (I'm very sensitive to subtle changes in my cognition). This may
have actually been due to playing dnb incessantly and ignoring other
intellectual activities.

With snb I basically attempt to hold all of the positions in my mind
simultaneously, and tend to use various patterns to my advantage. At
position I went from beating 10 and 11 back with intuition/guessing to
being able to beat those two with straight concentration (I must add
that beating 11 back with straight concentration feels great for some
reason). My max at position is 13 back.

The slowing down of my cognition with juggling is subjective.

[☉]

That is quite funny. I'd like to see you submit your final results and
join a few societies. (I should note that you've incorrectly estimated
what typically takes place in them; you have very little to be
concerned about as *generally* people there aren't motivated in
showing off how smart they are, since it kinda goes without saying.)

Anyway, some other power tests that come to mind that are fashioned in
the same spirit are (in no particular order):

Space, Time & Hyperspace (has to be taken with other subtests of TFG)
Logima Strictica
Foritensum

There are others, of course. (No, I haven't tried them yet.) The last
isn't easy to find unless you know how to navigate the interwebs
(which may serve as a kind of pretest, if you take it in the spirit of
puzzles, games, and occult ludism).

argumzio

[Brandon Woodson]

The Jouve-Cerebrals Test of Induction (JCTI) is another Gf saturated
test and offered at cerebrals.org. Out of the free psychometric tests
online, I believe it is one of the more valid test, since its results
correlate well with other popular standardized tests such as the
Performance Index of the WAIS. It supposedly measures accurately up to
around I.Q. 150. The link is:

http://www.cerebrals.org/wp/tests/jcti/

[Brandon Woodson]

I didn't read that this test was already included in the original
post. :) Anyway, I attempted to take the CFNSE a while ago, but was
unable to obtain a score - the test wouldn't score for some reason. I
suppose I'll try again some time soon.

[The.Fourth.Deviation.]

I'm uncertain that a test would be able to capture all nuances of
changes in cognition due to DNB training. I've subjectively noted many
changes in the way that I perceive the world after doing DNB, music
training, meditation etc., but I haven't taken a test to measure
changes. My PIQ was 100 when I did the WAIS at age 13.

If I were to take another test today and still scored about 100, I
would say with certainty that these tests are incomplete in their
ability to measure "gf" or "g". My ultimate point is that if, after WM
training, everyone around you notices that you seem more intelligent,
creative, humorous, etc, and you do as well, yet your scores remain
the same, which would you believe?

One of the most intelligent girls I know is poor at matrix tests. I
gave her one and had to explain some of the answers to her. Yet she
notices a numbing amount of abstract patterns in the real-world that
makes me feel blind in comparison. The level of awareness she has is
probably higher than anyone I've met. Ironically, she feels dumb in
comparison to me, since I could outperform her on a matrix test,
despite the fact that aspects of her brain function are probably on a
much higher plane than mine. Even though my IQ test alerted me to my
innate abilities, matrix tests in particular fail to do justice to her
brilliance.

So, the bottom line is that IQ tests may fail to detect real gains you
have made, and you should be cautious of this when interpreting
results.

[Brandon Woodson]

While I am sure that it is true that no psychometric test (I.Q. or
otherwise) is capable of measuring the all of the innumerable facets
of
cognition we call 'intellect', I'm not sure that we should completely
disregard the useful applications of I.Q. testing. I.Q. tests don't
attempt to measure intelligence, per se. They do, however, attempt to
measure 'g', the factor common to all mental tasks. While taking water
and fructose and placing them together, alone, in a vessel might not
make a very 'complete' and appealing juice, it's hard to deny that
they are constituent to what we can label as fruit juice. They are
what we might consider the essence of fruit juice itself. The same
concept is the one that 'g', and by extension I.Q., is built on. Sure,
we can not measure intellect in its entirety (even if we knew exactly
what parts made up intelligence, it would be impractical).
Nevertheless, we can attempt to determine how important mental tasks
correlate to each other and isolate the common factor, the essence of
intelligence.

While I.Q. tests may fail to completely capture the nuances
accompanying DNB progress, the absence of either should not completely
invalidate the value of the present other.

[The.Fourth.Deviation.]

IQ tests have value, but the girl i mentioned demonstrates that matrix
testing can apparently fail to detect profound intellectual
horsepower. That is a severe failure, and means that the results must
not have as much iron-clad credibility as we might think. It appears
that her mind works so differently that it completely escapes the
realm of what matrices were designed for. if she can slip through the
cracks, I wonder who else can, and has.

I am therefore suggesting that the OP, and others, use additional
measures beyond matrix tests to try and measure their gains,
especially if the matrix results are flat despite significant and
noticeable subjective improvements across several domains.

[Brandon Woodson]

This is why I said the 'the absence of either should not completely
invalidate the value of the present other.' - to de-emphasize any
exclusionary value that one has over the other. Again, I do not (and
did not) mean that I.Q. measures intelligence in its entirety. Is it
possible that the girl you write of possesses superior talents in
specific domains while having an average 'Gf'? To make further use of
my earlier analogy, one can continue add apple flavor to our water/
fructose concoction, but does that make that concoction any
'juicier' (i.e, having more fructose and water) or simply add more of
a useful but extraneous flavor (i.e., apple). In other words, one can
develop brain skills independent of increases of the brainpower
common
to all activities. Since improvements of Gf (and specific Gf subsets)
are the
ONLY scientific (empirical) basis for the claims made in the
published
n-back studies, it IS the scientific purpose of brain training; it's
what the resurgent interest in brain training built on. Any other
means alone and we cannot be entirely sure of what it is we are
measuring anymore. Of course, if one is experiencing significant
subjective improvements in absence of further proof, the purpose
should be, at that point, completely autotelic.

[The.Fourth.Deviation.]

"Does it seem [she] has talents in domains while having average Gf?"

Not quite, since humor, creativity, ability to ask theoretically
interesting questions, ability to notice abstract patterns, reason
abstractly, are supposedly real-world expressions of Gf, i.e. skills
predicted by Gf, from what I read. In other words, she seems to have
the real-world benefits of high Gf, but does only average on matrix
tests.

http://changelog.ca/quote/2011/10/16/aspie_children_have_superior_fluid_intelligence
this paper says abstract reasoning ability is a core manifestation of
Gf; she is overwhelmingly abstract for most people. So again, she
seems to have real world traits that indicate high Gf. So, if a
person is, in real life, 6 foot 2, but on paper is 5 foot 5, which
would you believe?

And why doesn't the test predict her real world abilities?
The only answer is that something must be lacking from these, or that
it is only accurate for certain individuals, or finally, that it only
measures a particular brand of logic, while failing to detect novel
brands of logic that produce the same real world outcome or benefits.
Finally, it could mean that these supposed benefits of Gf (humor,
creativity, abstract thinking and pattern recognition) are not
determined by Gf, but something else. For if this girl has average Gf,
she should not be above average in these intellectual domains, logic
would have us think. If she has average Gf, but real world traits of a
high Gf, apparently we must redefine how to identify real-world Gf.

I think, then, that for certain individuals with novel thought
processes, there is a disconnect between real-world Gf and theoretical
Gf as predicted by matrix tests. Thus, to re-iterate, some here should
not be hung up on stagnant scores on matrix tests, if they notice real
world benefits of fluid intelligence, which is what we seek, I assume.
I think it would be better to score the same on a test and notice
benefit, than to score higher on a matrix test and notice no real
world effect (if thats possible).

[Millicent Bliesener]

Regarding the high-iq societies, it's actually kinda nice to be around
people who don't go into fits if you speak using your vocabulary. Or
if you mention that you read the paper. or have a hobby outside
watching tv. There are a few people who pontificate, but it's not the
norm. :-) I don't recall who asked about high-IQ societies having
publications . . . Mensa has a few that are pretty good, but you need
to be a member to receive them. Their research journal is excellent.
They also have a more artistic/philosophical publication.

Millicent

--
If spider webs unite, they can tie up a lion.
- Ethiopian proverb

http//:10-Cent-Lifestyle.blogspot.com
https://millicentb.scentsy.us/Home

[Brandon Woodson]

:) All studies, including the above, uses tests to establish high
Gf, with which to correlate traits which stem from high 'Gf'. Without
the tests (I never specified matrix, only Gf or Gf subset), there
would be no characteristics to prove this person possessed high Gf
traits; and no way to redesign the high Gf measuring tests which
determine high Gf traits which determine the person, who in-turn,
redefines the high Gf test which is responsible for determining his/
her high Gf status. The last sentence is partially a joke. :)

[moe]

What do you guys think about iqcomparison.com test of fluid intelligence? valid?  I took it in an attempt to gauge cognitive improvement and the score is actually in agreement with the increase in score on the JCTI.  I scored 156 sd 15 on this test (1st attempt)  http://www.iqcomparisonsite.com/Results.aspx?UserID=b6052d29-9920-4ecf-9f28-10ad40862e00

[moe]

here are two more results

Tests by Paul Laurent (sd 15)

Simplex 148
ASIT 146

What's odd is the fact that I overestimated my score by about 10 points with Laurent's tests.  With most tests I tend to agree with my score. 

From the last 3 tests that I've taken (tests of Gf) it looks as if I'm averaging about 150.  I'm counting just the last 3 simply because of the fact that they are the only one's that I have not taken a bunch of times.  This is probably a bad/dumb question for many reasons but I'll ask anyway, do you guys feel that 150 level Gf is enough to be an ABOVE AVERAGE physics student?  Is this a field where IQ actually matters all the way to the top (would the 160+ iq guy be considered a better student than the 145-150 guy with comparable amounts of effort put in)?

[jotaro]

my friend got 138 on online test like you, but when he did official one he got 118. 20 points difference.

--

You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.

To unsubscribe from this group and stop receiving emails from it, send an email to brain-trainin...@googlegroups.com.

To post to this group, send email to brain-t...@googlegroups.com.

Visit this group at http://groups.google.com/group/brain-training.

For more options, visit https://groups.google.com/groups/opt_out.

[jotaro]

by the way... if you really like physics who cares about iq anyway.

[Brandon Woodson]

In my opinion, it's possible that higher I.Q. will help you absorb information more quickly as a student or neophyte in some specialty. And as a credentialed specialist established in your field, intelligence seems helpful, but expertise might be equally (or more) important - so I wouldn't get too disillusioned if I were you.

Terence Tao, possessing a psychologist-estimated I.Q. of >200, is heralded as the "Mozart of Math". And though nowadays any foremost eminent authority in a given field can earn an identical epithet (e.g., Magnus Carlsen as the "Mozart of Chess"), Terence Tao has garnered a reputation such that professional mathematicians often vie for collaborations with the former wunderkind by attempting to interest him with their problems. In his blog, Tao addresses his perception of the role of intelligence in success in modern mathematics (and later, academia as the topic meanders into a discussion over I.Q. in the comments section).

http://terrytao.wordpress.com/career-advice/does-one-have-to-be-a-genius-to-do-maths/


"It's not that I'm so smart. It's just that I stay with problems longer."

~~Albert Einstein, physicist


--Brandon

[Gwern Branwen]

On Fri, Sep 20, 2013 at 1:32 PM, moe <lindas...@bellsouth.net> wrote:
> From the last 3 tests that I've taken (tests of Gf) it looks as if I'm
> averaging about 150. I'm counting just the last 3 simply because of the
> fact that they are the only one's that I have not taken a bunch of times.
> This is probably a bad/dumb question for many reasons but I'll ask anyway,
> do you guys feel that 150 level Gf is enough to be an ABOVE AVERAGE physics
> student? Is this a field where IQ actually matters all the way to the top
> (would the 160+ iq guy be considered a better student than the 145-150 guy
> with comparable amounts of effort put in)?

Holding all else equal, probably, yes. See for example the Termites or
the SMPY longitudinal study or Roe's study
http://infoproc.blogspot.com/2008/07/annals-of-psychometry-iqs-of-eminent.html

That said, I must point out:

1. this is a hard question to study: all else is rarely equal, it's
very hard to get a sample of people with IQs in the 140s and 160s+,
it's increasingly unclear what IQ tests mean that far out, otherwise
obscure points like 'range restriction' become increasingly important
when dealing with extremes, etc.
2. why do you care? Your IQ is not going to increase another 20 points.

--
gwern
http://www.gwern.net

[Pontus Granström]

There are no tests valid above 150 since the it's impossible to norm results that high as it would require millions of test takers.

[Gwern Branwen]

On Sat, Sep 21, 2013 at 9:01 AM, Pontus Granström <lepo...@gmail.com> wrote:
> There are no tests valid above 150 since the it's impossible to norm results
> that high as it would require millions of test takers.

More difficult than impossible. The difficulty explains why the Roe
link I gave above follows a curious strategy for devising her test:

> Roe devised her own high-end intelligence tests as follows: she obtained difficult problems in verbal, spatial and mathematical reasoning from the Educational Testing Service, which administers the SAT, but also performs bespoke testing research for, e.g., the US military. Using these problems, she created three tests (V, S and M), which were administered to the 64 scientists, and also to a cohort of PhD students at Columbia Teacher's College. The PhD students also took standard IQ tests and the results were used to norm the high-end VSM tests using an SD = 15. Most IQ tests are not good indicators of true high level ability (e.g., beyond +3 SD or so).

The US military and the ETS test millions of people per year and have
data going back many decades, so if anyone could norm that high, it'd
be organizations like them.

--
gwern
http://www.gwern.net

[Pontus Granström]

I wonder how high iq that create these "high end" tests have. Would be better of measuring working memory since it's easier to quantify.....

--
gwern
http://www.gwern.net

[Pontus Granström]

Besides if you give "school type" of problems to experienced scientist they will probably score WELL above the average due to their training and experience, but I doubt that you will find many people with for example working memory or mental speed 6-7 SDs above average.