SOME IDEAS ON INTELLIGENCE TESTING

There are many possible definitions of intelligence. There is
cognitive or intellectual intelligence, relating to the mind.
There is also emotional intelligence, how well an individual adapts
to his environment in an emotional sense, communication skills etc.

Basic ability to survive is an important form of intelligence,
as is longevity. Other forms of intelligence at the individual level,
are a person's health or degree of enjoyment of life.

At a broader level, there is intelligence of a group of people,
a community, a society, a whole species or even the intelligence of
an entire biosphere (i.e. the planet Earth).

A subgroup of the possible definitions are the thinking types of
intelligence. Thinking is used in all kinds of human activities, whether
its dancing, laying bricks, painting or whatever. Despite all these
possible definitions, it is still possible to categorize one type of
intelligence as the reading/writing thinking that is helpful in academic
success. IQ or intelligence tests,(or more accurately aptitude tests)
relate to this kind of thinking, and it is this that is the subject of the
current topic.

WRITTEN INTELLIGENCE TEST
The most well known type of IQ tests are the written tests given to older
children and teenagers in school and the aptitude tests sometimes given to
job applicants. There are similar tests in books and on the net.
Most tests are "speed tests", there are more items than you are expected
to answer, so time is a factor.

The word intelligence evokes a lot of feeling and has many different
meanings, the word aptitude is probably better.

WHAT DOES AN IQ TEST TEST?
In theory, aptitude related to potential scholastic success, or success in
certain types of thinking occupations.

The official IQ tests do apparently have a reasonable correlation with
academic achievement. When you look at this, its not as unusual as it
seems. Imagine having thousands of items tested on thousands of students.
With careful statistical analysis, the items that correlated best with
academic performance could be selected, while others filtered out.

The cynical answer is that intelligence tests only test skill at intelligence
tests. However, its simplistic to suggest that scores are entirely random.
Some credit should be given for doing well, (or especially for a big
improvement on a previous test), as (for example) getting most items correct
in a difficult test in a short space of time can take quite a supreme mental
effort.

OTHER REASONS FOR SUCCESS IN IQ TESTS
Let us suppose we have a test of pure skill, one in which participants
have a series of questions, in each of which there is a number of possible
answers of which only one is "correct". (Correctness is an issue in itself,
but in this case we will presume there a single answer to each question
that the vast majority of people will agree (in hindsight) on.

A good example of such a test might be a clerical checking test.
There are 2 columns addresses. The 2 addresses on the same row are
either identical or they are are not. Hence there are 2 possible answers
, true and false, or which only one is correct. Let us also say its a
timed test, eg. there are 500 questions and 10 minutes available to do
the test.

Hence we have a test of skill in which content of the questions is
irrelevant.

There is a very high probability that say with a group of students, there
will be a range of scores and there will be individual differences.
I speculate that these will correlate (at least weakly) will overall
academic achievement on written tests, because there are some common factors
for success in both.

Here are some possibilities:
Motivation
Motivation may vary between people. Some people might just not
be interested in achieving a high score. Others might feel that they
couldn't do well, so aren't motivated, perhaps as a result of years of
conditioning that they are "dumb". Hence the result may be a self-fulfilling
prophesy. Others might not like to stand out from the crowd and be content
with an average performance.

High academic achievers might like to feel good by gaining a high score on
*anything* so will be highly motivated to increase their focus. Perhaps the
tests correlate with intellectual competitiveness.

Focus
Someone who can bring all their attention to achieving on the test
will do better than if their mind is wandering. For a timed test, this is
critical. This ability to concentrate may be an important skill for
academic success.

Another form of focus, is to maintain some thinking to overview the
thought process (e.g. Is there a better way of approaching the current
question I am working on?) Its easy to see a question, recognise a way
of approaching it, but overlooking a more efficient approach.

Discipline 
Some people may have more self-discipline than others
(or compared with themselves at a different time!). To do your best in
any kind of speed test takes a high degree of self discipline.

Time Budgeting
If a test is timed and there are more items than be easily
done in the time allowed, time budgeting is critical. Strategies such as
skipping items that a person is stuck on (and perhaps returning to them later) and
spending just enough time on items to get them right can be very valuable.
Note that these skills are independent of the skills needed for actually
solving the questions.

Experience
People with more practice in written/speed tests in general,
may have more developed test skills such as time budgeting.
Other test skills include strategies such as intelligent guessing in
multiple-choice questions etc., (e.g. by recognising weak distracters).
There are others can be evolved.

Note that this is independent of motivation.

The point being made is that timed written tests may tend to have a weak
correlation with academic success, regardless of the content of the test.

COGNITIVE STYLE
People (or the same person at different times) think in different ways.
A person who often thinks in a style suited to doing iq test answers will
do better than if they don't. You could say if you're smart enough to get
the right answer but no smarter you will do better than if you're wildly
imaginative and able to come up with lots of ideas beyond the scope of the
test. Aptitude tests favour convergent thinking, so if you are usually a
divergent thinker, you need to converge fast.

WHAT APTITUDE TESTS SHOULD NOT BE
Aptitude tests should not be knowledge tests
Tests supposedly of abstract thinking, pattern recognition should not
require knowledge of specific items. This is not to say tests of general
knowledge might not be fun and useful for students or job applicants, but
knowledge of trivia is not what aptitude tests are supposed to be
testing.

Some test questions of the early 1900s were simply knowledge tests. One
way to see whether a test item is mainly a test of knowledge
is it change its context to a different time or place. One such question
was "What is Crisco?" and they have the gall to suggest the ability to
answer such questions is genetically determined.

I've seen test questions where knowledge of higher maths is almost
essential to solve test questions. This gives people who are currently
or have recently studied maths a huge advantage.

There are also questions where knowledge of specific proverbs is highly
advantageous. Unless you testing the number of proverbs a person is
familiar with, this sort of question is randomising the result.

A similar situation exists with well-known puzzles. If I am familiar
with the "This man's father is my fathers son" puzzle, I will tend to
solve any related questions much faster than if I'm not.

WEAKNESSES IN TEST DESIGN
Of the tests I've seen some seem to be better designed than others.
In some popular psychology books, some iq test questions required
obscure knowledge such as familiarity with words such as dram, andiron
and the knowledge that Wigan is a town in England. Its hard to imagine
items such as this correlating with even general vocabulary.

Questions can be difficult to invent, so in some tests some poorly refined
questions get through, which only serve to randomize the result.

Other poorly designed questions include those which are only either true
or false are possible (unless they are included in a battery of all t/f
questions) or other questions where this a high probability of guessing
the answer. Multiple choice questions of 3 are absurd, for example.
Multiple choice questions should have at least 6 choices, and the
distracters need to be plausible.

Why is it poor to mix a true/false question with multiple choice and/or
open ended questions? In a speed test you probably don't have all the time
you would like and may have to guess some answers. It becomes valid to guess
a true/false question and spend the saved time on questions with more
possibilities (increasing the chance factor).

To my knowledge tests generally assign 1 point for each correct answer.
A test weakness would be to deduct for incorrect answers. For example,
in a true-false test like the clerical checking example, this would
result in 1 correct + 1 wrong = 2 unanswered. This would tend to
randomize the result. A situation arises with incompleted questions.
One fair approach might be to score a percentage of unanswered questions
as the probability of guessing them.

A test weakness would be to assign different scores to different questions
without displaying the value of each question. Some tests do put
questions in rough order of difficulty, but even with this approach,
this weakness would tend to randomize the result.

With IQ tests raw scores are converted to a mental age. The relationship
is not linear as the mental age scores tend to follow a bell curve.
At the extremes the difference in a few questions might make a difference
in quite a few points, while in the middle they have much less impact.
Hence at the extremes, they become less accurate.

There may be a ceiling effect, if some candidates are able to complete all
questions (getting most right), they may get a score lower than they
deserve. This might be countered by having more questions than *anyone*
can possibly complete.

RIGID OR FLUID?
The first mass intelligence testing involved US recruits in the first
world war, designed as a quick method of categorizing people.

The idea that an iq score is some kind of magical constant used to
more prevalent. These days the more popular notion is that an
intelligence test tests skill at intelligence tests.

Iqs have sometimes been used in the past for ugly political purposes.
Test scores have been used to argue that races are different intellectually
or scores are mainly due to inherited factors. Ironically the dominance of
the genetic view largely happened at a time when tests were more
knowledge based than now.

The 11 plus tests helped determine what type of school a child went to.
A careful study of such tests might reveal serious design faults in some
questions.

Tests are sometimes used for recruitment purposes. In some cases it gives
recruits who display poor body language or be subject to various prejudices
an opportunity to impress in an alternative manner.

Its a total fallacy that iqs cant change. Anyone motivated enough
could increase their ability to do aptitude tests. Thinking involved
skills which can developed just like any other skill. As with most human
activities, performance is improved with practice.

If someone achieves something it proves that can do it, not achieving it
doesn't prove they can't do it.

LOW SCORES
There should be no stigma with a low score. In most cases a highly motivated
person could improve low scores.

A low score might suggest poor basic literacy and numeracy skills and lack
of experience with concepts used in the test (an example, might be knowledge
of the order of the alphabet). In some cases it may be due to emotional reasons
(eg. negative self talk). Low scores in schoolchildren might suggest extra
tuition in basics is needed.

For example, poor reading skills might be a large handicap in most tests.

People should forget totally the notion that a low score is genetically
created. We are all born with incredibly powerful minds and incredible
potential.

There are many thinking skills not catered for by written tests.

FAST OR SLOW?
Sometimes we learn the notion that some people are bright (quick) and
others slow (dumb). One wonders what self-fulfilling prophesies, years of
schooling in which one is told they are dumb creates.

In real life, quick thinking is not nearly as vital as in a timed test.
There is usually time to reflect, explore alternatives etc.

Are some people quicker mentally because their nervous systems are faster?
Its more likely they have more neural connections relevant to the current
topic (which is acquired after birth.)

All healthy babies are born very bright. Stimulation in infancy is very
important for development in every way.

WHAT THE TESTS SHOULD BE
One approach might be not to throw the baby out with the bathwater and
see what can be learned from these types of tests.

First thing is that aptitude tests can be fun for some people,
especially people who like puzzles and thinking challenges. The
sense of achievement or improving your score can be fun too.

A really good test would be one in which teaches something as well as
it scores. Rather than ranking students etc., a better aim might be to
increase students ability to do iq scores. This would not be so much to
do with increasing knowledge of iq test technique, but the underlying
useful thinking skills they could help develop.

OTHER IDEAS

Some possible exercises might be :


	

		
		

			
				(a) group work- students work in a team
			
		
	
	

		
		

			(b) higher scoring students tutor weaker students
		
	
	

		
		

			(c) students create their own tests
		
	
	

		
		

			(d) students discuss which test questions are more suitable than others
		
	
	

		
		

			(e) students what might be learned from the tests
		
	
	

		
		

			(f) invent your own ideas