Re: Cambridge Brain Science averages?

[jttoto2]

If it isn't representative of the entire population it could very well be close.  The cambridge science bell curve for the digit span peaks at 7, a perfect population average.  Some of the other games might be considered entertaining enough to some might have multiple playthroughs.  If I were optimistic, I would put less stock on the planning games, as these might get played more than once do to their gamelike structure, and thus skewing the average score.  Then again, maybe not.

On Sunday, July 22, 2012 8:53:44 AM UTC-4, Carth wrote:

There probably is a sampling bias on there. The general public would not be using that website at all. Generally the only people using Cambridge Brain Sciences will be those who are interested in their intelligence or improving it. Hardly any member of the general public would bother with it (or encounter it), and I strongly suspect the average person going on it will be of above average intelligence.

Therefore you can take a score of 50% to mean you are average in a group of above average people.

Like you, I also found there were a few I struggled to get above 50% on at first. I suggest you go back to them after practice and you'll find a lot of them become easier or more intuitive. Whether or not scores you get after practice are still indicative of your performance is another matter. But given that a lot of people use the website in just that way (repeating the tests) it's safe to say that a more valid comparison to them would be your scores after you have also practised.

For some of them you only really get the hang of what the task is after doing it a few times and before that some of your brain power is taken up trying to work out what exactly it is you are supposed to be doing. e.g. For some like "odd one out" the kind of patterns that occur are very quickly learned, and while at first you find yourself staring for ages looking for an odd one out, after a few tries you find yourself just looking for patterns you've seen before and they jump out at you.

Finally, for some of them the possible scores you can get are integers in a small range (e.g. 2-9) and getting one more point would push you into a new boundary, for example if you got 6 and nearly everyone got 7 you can see a big jump in your percentage just by getting one extra. I believe "digit span" is one of these. Most people usually get to a certain number easily and then simply cannot remember an extra number without losing the others. This is the limit of your short-term memory (commonly 7 ± 2 items).

Some people use strategies like digit grouping on that one to increase their score beyond their short-term memory capacity. Some of the non-timed ones can be perfectly solved without error, and nearly all of them can be quit at any time and the score doesn't get counted. This means you can inflate your average if you only record your better scores. So the averages will be skewed a bit from this.

On Thursday, July 19, 2012 9:55:05 AM UTC+1, brainslug wrote:

I am wondering if the averages on http://www.cambridgebrainsciences.com are representative of the general population's average on these tests.

Normally, I only score a bit above average, 50-70 percent, and, on a few, I score terribly. Specifically, on "digit span" and "polygons", my scores are 28% and 27% respectively. I am quite aware that these are my weak spots, but 28% is absolutely horrible if these averages represent the general population.

I expect more from myself, and I certainly know that my number span is not bordering on a mentally challenged person's. I was thinking that maybe there is a sampling bias of more intelligent people being more interested in these games, but I don't really know if this is popularly linked on Facebook or something similar which would bring it closer to the general average, or if maybe Cambridge has collected random samples that make up the majority of the results. I also see that the sample sizes are not giant, but most of them seem to form nice curves anyway(although some of them are deceptive in that they round off the edges to create an appearance of a bell curve when there are not even enough different possible scores to form a real curve.)

[jttoto2]

Also, a good portion of people above-average just aren't going to be above-average on every score.  Don't beat yourself up over it.  

[brainslug]

Well, that is a bit sobering, but I guess it just means there is more room for improvement. Funnily enough, the more the task utilizes numbers, the worse I do on it. I have 70% and 79%(score of 8) on the pairing and spacial respectively, but I have low scores on the monkey ladder and digit span.

I guess it is okay that my concentration scores and number scores are the lowest. I know I am not too good at calculations, and I don't attach too much importance to concentration anyway. My reasoning scores are high(except for double trouble), and so I guess I am happy. I am good at what I want to be good at, I guess.

I think it just shocked me to get below average on something when, for all my life, I have been at the top of my class and everything. I guess it just shows that there is more to education than raw intelligence. Accepting your weaknesses is the first step to fixing them.

On Tuesday, 24 July 2012 00:40:01 UTC-4, jttoto2 wrote:

If it isn't representative of the entire population it could very well be close.  The cambridge science bell curve for the digit span peaks at 7, a perfect population average.  Some of the other games might be considered entertaining enough to some might have multiple playthroughs.  If I were optimistic, I would put less stock on the planning games, as these might get played more than once do to their gamelike structure, and thus skewing the average score.  Then again, maybe not.

Also, Carth, thanks. I will try practicing them more. I did each about 2 or 3 times and am using my scores from that. I also make a point not to do grouping because I feel like it skews my data if I do. Maybe I can get my digit span up by using nback too. Thanks for the encouragement.

[jttoto2]

Actually, there is some evidence digit and spatial span tests overestimate capacity because most people use grouping (chunking).  You are probably skewing your results by not chunking!  Some estimate that when properly controlled the average person holds 4 items, not 7.  Try chunking to see if it improves your score.  

[Mike Morg]

I realize im a bit late but ive been playing on Cambridge brain science and this what i have to say: First of all, i do agree and believe that above average people would even come across the website, so it would be skewed in that sense to reiterate. But also, with the digit span the website goes pretty quick when it comes to showing the numbers - thats the purpose of my inquiry here. I did digit span on another site before this and my highest was 12 (9 not chunking). But on this site i barely get to 11 but i have and completed it which makes me above 99% of the population. Now is only 11 99% ABOVE the population? I dont know. Also my overall verbal fluency is above 99% of the population with a g score of around 3.5 - which would be equivalent to someone with an iq of 150+. I'd actually like to know if all this holds true but probably never will unless i take an iq test haha. But anyways to stop rattling on, results can definitely be skewed but i'd imagine a good way(relatively). 

[Toma Mircea]

I think the verbal test is hard for non native enlgish speakers, like me

[Mark Williams]

I found this site to be really good, its presented and organised well, and well, it's free!!! :) I've played most of the games plenty of times, and as has been mentioned above your scores increases a great deal with practice. At least they did with me. The only ones I've genuinely struggled with after a couple of weeks are the Memory ones. I've gotten around 5 of the games up to 100% now, and a couple more over 90%. This suggests that there is a great deal of technique, as opposed to purely mental adaptations going on. Take for example the one where you are given a word RED or BLUE, in either colour,  asked to note its colour and then asked to click on the word RED or BLUE below (which themselves can be in either colour). First time most people come across this they'll struggle. But then you figure out that if you kinda lower your field of vision, keeping the word as a sort of blur at the top of your peripheral vision so that only the colour is discernible - this helps enormously and you are then able to increase your score by 100-200% perhaps (I can't remember). Is this cheating?

Anyway, great site I think.

[yui takashima]

Your score on verbal intelligence  quite amazing. Which iq test have you been tested on ? I am assuming the WAIS IV? I got tested on that and I scored 120s on most of the indexes with the exception of my verbal comprehension 141, which made the overall Full scale iq to be 136. And like you, I am enjoying cambridge brain sciences, but there are some issues with their scoring. Have you noticed some of their bell curves have swings up and down which makes the normally regarded bell if the line was drawn above them. But I highly doubt the validity of their percentile giving. For example, you can't technically get 100th no matter what norms used. But they say that the digit span of 12 is 100% which makes it blurry as to its comparison. I strongly believe they should start using the percentile form instead of that dubious percentage ranking. But what do you think? 

[Mike Morg]

Are you asking me? 

[hera...@gmail.com]

Hi Yui, my 7 year old score 148 on the Stanford Binet, do you think this is actually an accurate projection to adulthood score?

On Thursday, December 19, 2013 7:51:14 AM UTC-5, yui takashima wrote:

[Brandon Woodson]

The score is 7 years old, or you were when you took it? :) Kidding.

I am not the addressee of your post, but I write this in the good faith that you won't especially offended by unsolicited advice. There might be quite a few things to consider. A few complications that easily come to mind are:

-modernity of the test when you took it since (because of the Flynn effect) testing with outdated tests can produce skewed results

-the quality of the test and its correlation with other tests

-the type of score reported, ratio or deviation

-the trajectory of your cognitive development; children develop asynchronously


And, of course, education, lifestyle, health, etc. are key to your current psychometric intelligence, which would affect comparisons between measurements of it and an older score.

Short answer: If much time has elapsed, I wouldn't personally bank on it; In cases like this, people regularly report higher, equivalent, and lower scores with wild disparity.

If you feel compelled to speculate, at the risk of inaccuracy, this resource may assist, if indeed the score in question is a ratio score:

http://hiqnews.megafoundation.org/John_Scoville_Paper.htm


--Brandon

--
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/d/optout.

[jttoto2]

Hate to resurrect this post, but since some people were wondering if this site represents population averages, I created a new account and bombed the digit span test after reaching 7 digits, which gives a percentile ranking of 48 (very close to 50)  If 7 really is the population average, then I'd say the Cambridge scores are close to representative of the general population.  Some researchers have noted that the number 7 is too high however and that the average young adult has something closer to 6:  https://books.google.com/books?id=i3vDCXkXRGkC&pg=PA266&lpg=PA266&dq=average+adult+digit+span&source=bl&ots=3b2P2ZZHGg&sig=0R5_XA5kACE1JTML2SLe5AmkbeM&hl=en&sa=X&ei=o97kVJmVCMGrNp7igaAD&ved=0CJoBEOgBMBM#v=onepage&q=average%20adult%20digit%20span&f=false