Mensa practice

in 20, 21, 22
use the elements on the left as a pattern and then select two elements with the same pattern

originally no but after explained it, i do

A sample size of 1 is bad for trying to evaluate a group. Not for evaluating the individual that submitted that sample.

polling the same individual 50 times is not the same as taking a sample size of 50

You're right, it's not. But that's not what's being discussed here. The question is whether or not the test can propeely evaluate individual ability. And the point is that if the test is not an adequate measure of individual ability, then it is a shakey measure of group ability. To suggest otherwise would be akin to measuring weights of animals with a shifting ill-calibrated scale, and then trying to make claims about different groups of animals when the scale was wonky for all of them at the individual level and we had weights all over the the place for all of the animals.

Either we are using a good calibrated scale or a fucked up one. If the scale is fucked up and not good for individual measure, it's likely not good for group measures.

no
bad at measuring individually doesn't mean bad at measuring groups.

Yes it does. Something that is ill fit to measure an individual cannot possibly be reliable for measuring groups. If a person can step on a scale three times back to back and get the readings 180lbs, 300lbs and -62 lbs, there is no way you can trust that scale for measuring a group of people.

It's either the scale is reliable for individuals and groups, or it is not reliable for either. I believe you are confusing using an individual measure as an aggregate for a group, which is definitely improper, for whether or not an individual measure is actually reliable for the individual, which is made reliable by the individual repeatedly testing (stepping on and off the scale, answering multiple questions on the voluntary IQ test, or taking multiple voluntary IQ tests) and getting measurement results in the same ballpark time and time again as the same individual. The individual measure has to be reliable for the individual or else we are using a bad scale. Bad scale for individual is a bad scale for the group.

Ergo if IQ is a legit group measure it has to be built on legit individual measures. If the individual measures are questionable, so is the group aggregate.

> If a person can step on a scale three times back to back and get the readings 180lbs, 300lbs and -62 lbs, there is no way you can trust that scale for measuring a group of people.
But if you got 180, 174 and 188, it could be a good scale for measuring a group of people.

We use inaccurate sensors and put them with other inaccurate sensors in my lab.
Sensor fusion is current research

>tfw actually arguing for iq tests

Yep. Basic statistics that the higher the variability of a test the larger the sample size required to get a meaningful result.

That's if you got numbers in those range and if the level of variability on the results returned was equal across the individuals measured.
It's not going to help you to get those numbers on one person and then the numbers originally posted on the next person. Innacuracy only can be worked around to find relationships among subsets when it's verifiably a consistent innacuracy.

mensa is, ironically, full of insecure simpletons

Indeed, however that logic applies when trying to coax out group aggregates. Not evaluate the individual sample for the individual. If we are working with a good enough scale for an aggregate, we are working with a good enough scale for an individual.
If we are dealing with the example provided by the other user where the scale reports 180, 174, and 188 on the individual, we can assume the scale is relatively accurate for estimating the weight of the individual. It's not precisely calibrated to an exactness, but we can definitely make statements about the individual as a result of the measurements obtained. Our scale may be off +/- 10 lbs on any given measurement possibly, but we can say with high certainty that the individual is under 200lbs.

If it works for the group and you can make statments about the group, it has to work for the individual and you can make statements about the individual.

*Consistent range innacuracy

Yes that is all true when it comes to the real world. However, from a purely theoretical perspective, you can have a test that is wildly variable to the point where it provides essentially no useful information about the individual, but as long as it varies relative to the "true" value for each individual it will provide an accurate mean for a large enough group.

That's IF it varies relative to the true value for each individual with relatively consistent variance range. You will get no accurate mean for a group of any size if the variability numbers are all over the place, and it would be entirely unethical to start making claims about the data with such problems on measurement. For example, finding variance ranges for of +10/-10, +25/-25,+5/-5 for different individuals won't help you discover any meaningful conclusions unless you either scale each reported individual to their true value based on their unique variance range, or you separate by variance ranges before separating by other dimensions and making claims. If we are dealing with a test which doesn't require such cleaning before making statements on groups, then we are dealing with a test that is suitable for making statements about the individual taking that test.

I will use this thread to ask a question myself.

Just half assed an IQ test test.mensa.no and got a pretty decent result. It would definitely have been better if was serious and prepped myself. I am confident I could join mensa if I wanted. They offer physical IQ tests almost monthly nearby.

So what are your thoughts on "high IQ clubs"? Are there any true benefits to being in one? And what do you think about an IQ certificate itself? Also, are physical mensa IQ tests actually good?

Obviously IQ is flawed and so on. I don't really care about all that stuff. But I happen to do decently well in these tests so of course I might as well reap any benefits from them.

Yes that's why I said as long as it varies relative to the true value because if it didn't then the thing you are trying to test/measure isn't affecting the result you are getting at all.

It's fine if it varies a different amount from the true value each time it is tested, that's part of what variance is. However, in the case you are talking about where you consistently get different variances for different individuals then of course you need to be more careful and there could be a confounding variable. However, that also applies to tests which have a variance low enough to provide useful information on an individual level.

M
15
8
6
5
4
1
2

I just did it properly and got 31/33. Should I apply for Mensa?

name a single noteworthy scientist or mathematician that was/is a member of a "high IQ club"

I can't

doesn't mean there isn't of course

I'm sure there are many, but I think there isn't really much benefit to being in one other than having to brag to people that you're smart.
Also, I think it costs to be a member of these clubs.

What about this is not a good measure of intelligence? This is literally what intelligence is - looking at data, finding patterns and then using those patterns to extrapolate predictions.

I think if you have someone working on these types of exercises all day they'll get a significantly better score than they would have had they just blindly taken the test, because the questions tend to be very similar in the types of patterns you have to spot.

Test how many can be wrong until it stops asking you to pay for real test, if it encourages you to go in test even with much worse results, then maybe...
I would say, that if one gets 130+ in every net-test, he has real chances - but even one site that gives something below that, and chances drop dramatically.

Also:
>I got 31 right and quite sure I cant get >130 on real test right now.

And for this you will have to provide evidence, user. The whole point of an iq test is that the patterns are of a relatively low order - only basic knowledge is required, like knowing what digits stand for, knowing the alphabet etc. Assuming everyone taking the test is familiar with these, the rest is actually difference in real cognitive ability.

This is only controversial for political reasons. The US military has been using this shit for a century and they seem content with the results.

Sorry for samefagging, but come on, claiming g doesn't exist is basically denying what every human population everywhere at every time in history has empirically observed to be true: smart people exist and stupid people exist.

What is 7 supposed to be?
I even looked it up on OEIS.
There's not enough information for any meaningful mathematical sequence even match that.

Are you supposed to be squaring and subtracting 1 like the one below?
Still though, giving 1 period of a periodic sequence is fucking retarded.

So you've never taken an IQ test twice and gotten different results?

isn't it just a sine/cosine function with π/2 intervals

so like sin(π/2), sin(π), sin(3π/2), sin(2π), etc

what's wrong with that?

>Still though, giving 1 period of a periodic sequence is fucking retarded.

>patterns are of low order
>tests literally include things such as the fibonacci sequence, factorial patterns, finding the length of the sides of shapes, etc.
These are things definitely of a lower academic order, but someone who has actually had quality schooling will blow someone else out of the water because these are all patterns they see regularly in academia. It's nonsense and unethical to try and test some malnourished fellow in the sticks of China or Nigeria who never had a quality day of schooling in his life and compare it against stats of a first world nation. Likewise the same issue is apparent when testing across stratified groups within developed nations.

youre overthinking it

wow the autism

1 0 -1 0 _ just look at the fucking symmetry you dumbasses

sure there is no perfect answer but you arent supposed to pick a perfect answer, you are supposted to pick the most likely answer. 1 3 6 10 the next number could be anything but 15 is obviously the simplest / most likely

Where did I say anything about Nigeria or China? Obviously there are a ton of factors tha can cause an individual to score below what his intelligence would allow. My point is, this intelligence variable still exists, and can be inferred from various forms of testing with varying degrees of accuracy.