Rules of conduct
These are rules of conduct and other points relevant when taking intelligence tests scored by Paul Cooijmans. By taking a test one agrees to abide by these rules. Their purpose is to prevent certain types of problems that have occurred in the past. The essential points are summarized first because it is a long read.
Possible measures against serious offenders include, but are not limited to, invalidation of one's score, exclusion from further test scoring, and expulsion from I.Q. societies. Treatment of offenders is described in more detail in the novel Field of eternal integrity.
Test scoring and score reporting come with the absolute and unconditional guarantee of sincerity and objectivity, which also implies that each test submission is scored independently of one's earlier scores; in other words, earlier scores have no influence on later scores.
Some candidates find their score is lower than expected. This is normal and the result of being blind to one's errors, like one can sometimes not find a lost object although it is staring one in the face. Naturally there will be no discussion with the candidate about one's answers or the construction of the test, and no "independent committee" to decide if one's answers are better than the intended ones. It will also not be revealed which test problems one had wrong.
The page with statistics and norms contains all of the available information on that aspect of the tests. It is not possible to predict when a new statistical report for a given test will appear because that depends on incoming test submissions. Initially, the statistics page of a test contains only preliminary norms. A histogram of scores is normally published after 16 submissions; a first full statistical report after 25 submissions; the norms stop being "preliminary" after 36 submissions. Beyond that, updated statistical reports may appear whenever the number of submissions reaches a square of an even whole number, not necessarily containing new norms every time. This policy reflects the general phenomenon in statistics that the accuracy of our estimations increases in proportion with the square root of sample size. This principle was first understood by Abraham De Moivre (1667-1754).
Some of the tests are accepted by various societies while others are not; it is up to the society in question to decide which tests are accepted.
One can take each test only once. Some good reasons for that are explained here.
Test scores are final, not open to debate. Accusations that the test was scored dishonestly, demands for a higher score, and attempts to intimidate, bully, or threaten the scorer into reporting a higher score are not effective.
Candidates who, before taking a test, ask for or try to trick the scorer into revealing correct answers to test problems, or into giving help in finding those, will either not receive a reply, or receive a reply which may be true or false to prevent them from inferring anything from the reply or its absence. This also means it will not be confirmed if particular answers are correct, either before or after the test submission. This also means it will not be revealed what the score of a particular set of answers would be when submitted; showing that set of answers means submitting it.
Report your age with each test submission as asked in the test instruction; note that age is not date of birth. An important difference for instance is that age tends to slowly evolve over time, most typically in an upward direction and sometimes even by as much as a whole year per twelve months, while date of birth does not. Age as meant is a whole number representing how many years one has lived, rounded off downward. An example of an age is "34". Even though age is explicitly requested in every test, about half of the candidates neglect to report it, which is why a penalty of 20 (twenty) I.Q. points has been effective regarding this omission for all test submissions scored after March 31, 2007.
Additional answers or corrections received after or apart from the actual submission are not used in scoring. This is so because it would be impossible to treat such consistently; no received submission could ever be scored, as there might always be an additional answer yet to come. For information, such answers are nearly always wrong, being the result of one's imagination running wild. Submitting one's answers is saying "These are my final answers", and that one and final submission is considered scored the moment it is received.
In case the above is not clear: Once the submission has been scored and reported on, rescoring it based on additional answers would mean to give the candidate information about the correctness of those answers (which is not acceptable) and to allow a retest (idem). This paragraph also contains the answer to questions like "What happened to the 'No time limit' clause? I thought I could take as much time I wanted to send answers?"
Then there are those who say, "Then only use additional answers received before you have scored the submission, and do not use those you receive after having scored it". But that would not be fair as it would mean to treat candidates unequally; for instance, candidate A. submits answers while the scorer happens to be working so that the scorer can score them at once, and candidate B. submits answers while the scorer is away for the day so that the scorer can score them only the next day. Then candidate B. has a whole day more to submit additional answers that would occur to the good candidate, and the conditions are unequal. Therefore every submission must be considered scored the moment it is received, regardless the actual date and time of scoring.
Then one could say, "Then allow a fixed amount of time for additional answers". But that is what the scorer is already doing; to wit, zero days, zero hours, zero minutes, zero seconds and zero milliseconds. This is logically the only consistent thing to do.
Only one answer per test item is scored, unless explicitly indicated otherwise in the test. This prevents one from gaining unfair advantage through "shooting with grapeshot". When multiple answers are given to a problem, only the first is scored. The others are studied nevertheless with regard to item validity.
It is not possible to send answers conditionally: "Do not score these answers if…" (they contain errors, the score is lower than..., a retest is not allowed, et cetera). When answers are received they are scored, period.
It is not allowed to discuss test items, answers, alternative answers et cetera with anyone, either before, during, or after taking a test, and regardless if one intends to take the test. Just give the answer that seems best.
It is possible to include explanations or comments with one's submission, which may help to detect bad items but not influence the score. There will be no response back on those comments. Make sure to distinguish the comments from the answers; only the answers are scored and yield credit, never the comments. A possible score based on explanations or comments would not be comparable to the scores of any other candidates, and it would be impossible to have norms for such incomparable scores. Possible bad items are removed or revised in a future revision of the test, after which the test is renormed.
One should avoid sending explanations or comments and asking feedback on them while still working on the test, because one is then taking the risk that a possible response from the scorer helps in solving the problems in question.
Some think they see alternative or multiple possible correct answers. In most cases this is because one is overlooking something or making a reasoning error. In some cases it is true, and those problems are then treated as bad items; that is, removed, revised (to lead to a single solution) or replaced in a future revision of the test, after which the test is renormed. In the meantime, the alternative answer is either given full credit, or the test problem is discarded for all candidates, depending on various circumstances. When a bad item is discovered in a new test that has only been taken by a small handful of candidates, the item is revised or replaced "in place" (so without resulting in a new version of the test) and the few existing submissions are rescored, if needed after inviting the candidates to send a new answer to the changed problem. In any case, great care is taken not to let candidates miss points by giving valid alternative answers; not to disadvantage them compared to those who give the intended answer.
Candidates who are disappointed about their score sometimes worry about having lost points by giving brilliant alternative valid answers that have not been recognized, but this is in almost all or all cases unjustified. These tests and problems have been designed expressly for the kind of people who could find such answers, and to measure that ability through objective statistics rather than arbitrary subjective judgement. Problems that require arbitrary subjective judgement at every test submission to decide whether a candidate's answer falls within a yet undefined scope of acceptable answers are less than ideal for this purpose.
It is not allowed to publish tests or test items and/or answers in any way, for instance on Internet fora, regardless if the answers are correct.
When reference aids or reference books are allowed, this includes the Internet, unless otherwise indicated. For clarity, and obviously, the use of reference aids means for instance the looking up of a word one does not know; it does not include cheating of any kind.
Self-obvious as it seems that the test to which one is submitting answers must be identified, there do exist candidates who simply send a collection of unidentified answers, leaving it to the scorer to connect them to a test of the latter's choice. And indeed, the scorer is in most cases able to infer from the answers, their number, and the tests known to be in the candidate's possession, which test is likely meant. However, it is unwise to leave the delicate task of identifying the test up to the scorer, because (1) the scorer is then helping the candidate by finding the test on which the particular set of answers receive their maximum score (after all, on all other tests the score would be close to zero), such help not being allowed, and (2) the candidate could claim, after seeing the score report, not to have submitted answers to the test in question at all (which technically would be correct) and thus trick the scorer into allowing a retest with the advantage of already knowing the score on the first set of answers. Finally, (3) the scorer, being in a bad mood, might decide to interpret the answers as meant for another test altogether, resulting in a near-zero score, and boy would the candidate be dismayed then!