IQ Is a Relative Position, Not an Absolute Value
The phrase intelligence quotient suggests an absolute quantity of ability, but the number is nothing more than a statement of where someone sits relative to others of the same age. Most tests are scaled so that the mean is 100 and the standard deviation is 15, which makes an IQ of 115 mean 'one standard deviation above the mean', or roughly the top 16 percent. A score of 130 is two standard deviations up, roughly the top 2.3 percent. Because the scaling depends entirely on the distribution of a reference group, the same person can receive a different number from a test normed on a different group. An IQ of 120 carries exactly one piece of information - that this was the person's position within that test's reference population - and nothing more absolute than that.
What a Standardized Test Actually Does
Standardized instruments such as the Wechsler and Binet scales collect age-banded data from reference samples numbering in the thousands and build conversion tables from them. The procedure for administering the test is specified just as tightly: the examiner's qualifications, the condition of the room, the wording read aloud to the examinee, even how time limits are measured are all held constant. This rigidity exists because when conditions drift, the score drifts with them. The same person taking the same test can produce a different result depending on fatigue, nervousness, or noise in the room. Standardization is precisely the work of suppressing that drift as far as possible so that scores become comparable at all.
The Decisive Difference from Quick Tests on the Web
Most of the 'IQ tests' available in a browser have never been through this process. They have no age-banded reference sample, they impose no consistency on the environment the test is taken in, and they have no way of tracking who has taken them or how many times. Even when the question format closely resembles that of a genuine instrument, the number that comes out is not an intelligence quotient. The effect of repeated attempts is the most serious problem of all. Getting familiar with a fixed question format raises the number of correct answers, but that is fluency with the task rather than a change in ability. Standardized instruments control the interval between retests; a test on the web imposes no such constraint.
Only Part of Intelligence Is Being Measured
An IQ battery is assembled from several subtests, each addressing a different facet: verbal comprehension, perceptual reasoning, working memory, processing speed. The composite score is a blend of these, so two people with an identical IQ can have completely different strengths and weaknesses once the profile is broken out. Beyond that, the ability to read a social situation, persistence, creativity, and physical skill are simply not among the things the battery examines. A high IQ means someone handles efficiently the kind of task the battery measures; it is not a ranking of human capability as a whole. Losing sight of this is what pushes interpretations of the number far away from what it can actually support.
What IQ Predicts and What It Does Not
Correlations between IQ and academic attainment, and between IQ and performance in some occupations, have been reported repeatedly, and the measure does carry real predictive power when a population is viewed statistically. A correlation, however, is not a formula that fixes an individual's future. Among people with the same IQ, academic results and income vary enormously, and much of that spread is accounted for by learning environment, health, household finances, and whether opportunities were available. The accuracy with which anyone can say of a particular person 'their IQ is this, therefore their life will go that way' is far lower than the statistical correlation might suggest. Refusing to convert a group-level tendency into a statement about an individual is the single most important discipline in handling this measure.
Using It as a Record of Your Own Trend
None of this makes an unstandardized test pointless. Drop the comparison with other people and the reading of the number as an absolute, compare it instead against your own earlier results taken under the same conditions, and the measurement becomes perfectly usable. Learning how far your results move with sleep, physical condition, and time of day is a way of reading your own state. Bench's Pattern Reasoning test does not produce an intelligence quotient either; it is built as an instrument for measuring your grasp of unfamiliar rules in a consistent format and watching how that moves over time. Working out the conditions under which you tend to perform well is more useful in practice than the size of the number itself.