I'm a school psychologist so I administer and interpret intelligence tests every day at work. Yes, intelligence is supposed to remain stable over time, in the absence of any medical conditions, injuries, emotional, or behavioral influences that can alter one's performance during test administration. For example, if I am evaluating a student with hyperactive and impulsive behaviors that are present during the session, I interpret the results with caution, and note in my report that the results most likely represent a low estimate of that student's present level of cognitive ability due to said behaviors. Anyways, IQ isn't supposed to be referred to as a single number, but rather, as a range of scores (i.e, average, low average, borderline, high average, superior, very superior). When it is scored up, you do obtain the single score, but also a range of scores (confidence interval stating that if you were to repeat the assessemnt over time, you would have a 90 or 95% chance of obtaining scores within that range without the effect of error). The number is reported as a means of comparison to other assessment scores (they are all presented in standard score format, with the mean being 100 and a standard deviation of 15 or 10 depending on the test). That being said, to answer the OPs question, 100 is considered to be the average score, with 90-110 making the average range on tests like the WISC-IV, and 120 and above making the superior and very superior ranges.
Sorry for the clinical reponse. I've been sitting here writing up some psych reports that are due since I was away last week (at Disney, of course!) and I guess I'm on a roll.
There are so many different websites that say so many different things.
