There are a lot of myths on the internet perpetuated about the Personality Type, and Myers-Briggs Type Indicator in particular. As I come across them, I'll attempt to bust them here.
1. Personality (tests/profiles) are like horoscopes. You can read anything you want into them.
It's more like horoscopes are attempting to be like personality profiles. In the case of the MBTI, there are 16 basic divisions, all of which are ultimately different from the other. There are similarities between many of the Types, but that's because they share common attributes (they may share ExFx in their code for example). Horoscope profiles are not generated by a question and answer instrument structured by an underlying theoretical structure. The MBTI is.
2. The MBTI doesn't predict success at a recommended career/job.
It's not supposed to. The MBTI reports job satisfaction vis-a-vis Type. When you see a "preferred careers" table, what you're seeing is the results of people reporting high job satisfaction compared to their Type. The MBTI shows what you prefer, not what you're good at doing. There are plenty of any given Type that aren't very good at being their Type. You may prefer to play baseball right-handedly, but that doesn't mean you're any good at hitting a ball.
3. The MBTI is frequently used to screen job applicants.
This is an unethical application of the MBTI according to APT International. Refer to myth #2 above. The exception to this rule is an organization that has conducted validated research to support such application, and I can nearly garuntee you that most HR departments don't. Personally, I'd be skeptical of working for any organization that uses any psychological instrument in their hiring process, but that's just me (being an ENFP).
4. I'm an Extravert, but I don't like to socialize with people.
Extraversion and socialablity are not the same thing. Extraversion and/or Introversion refers to where one focuses one's attention, or consciousness, and how one prefers to reenergize. Extraversion refers to the outside world, and Introversion refers to the inside world. The outside world isn't all people - it's things, ideas, animals, the sky, you name it. Socialability (or being a "people person") is more accurately indicated by the Thinking/Feeling dichotomy than the E/I dichotomy. When combined (as ExFx, IxFx, ExTx, IxFx) you get an even more accurate indicator. Additonally, remember that just because you prefer it, doesn't mean you're any good at it. You can prefer Extraversion and Feeling and be a pretty rotten person to be around at a party.
5. I'm an Introvert, but when I am with my friends and/or family, I'm very Extraverted.
Yep, and that's pretty indicative of Introverts. Those who prefer Introversion frequently feel "safe" enough around intimates to "let their guard down" and show a non-preferred side of themselves that isn't normally presented to the Extraverted world. Extraverts do the same thing in reverse - being uncommonly quiet and introspective with those they are intimately familiar.
6. I took the MBTI (in college, at work, in therapy) and then took the one online and it came out differently!
Well, yeah. And there's a few different possible reasons why.
a) It's a different indicator. If you took this one, then be advised that it's not the official MBTI. In fact, unless you're taking it through a qualified MBTI trainer, you're probably not taking the official MBTI. While I think the referred online indicator is pretty accurate, it's not the real thing.
b) You answered differently. That's hardly the indicator's fault. Now indicators are supposed to compensate for some degree of variation on the part of the taker when it comes to answers, but it's not psychic. Maybe you need to be a little more consistent ;-)
c) Officially, test-retest reliability scores for the MBTI are pretty high. Test-retest reliability is usually done with a 9-month time scale, so if it's been years since you took it, there's more of a chance of it changing. Now if you only changed on one letter, that's hardly the end of the world; see the next point.
d) The first time you took it (or maybe this time) you scored pretty close to the middle of the pack on at least one of the dichotomies, so the chances that you "flipped" letters is pretty good.
e) Possibly what happened (especially if you took it at work) you answered it in such a way as to satisfy an image you wanted to project, or expectations you expected to meet.
e) That usually indicates something is pulling you in one direction away from your preferred function. When we're younger, we're more under the influence of authority figures and family dynamics.
If you flipped on more than one letter, a little life-examination may be in order. ;-)
7. I read the profile and that isn't me!
First, the profiles aren't the point of the indicator. They just provide a nice snapshot of the larger dynamic of what's really going on (which most people don't get to learn, especially if they just took an online version without a trainer).
Secondly, well, then why did you answer the questionnaire the way you did? The resultant code only comes from one source - you. Garbage in, garbage out. A little self-awareness may be in order ;-)
[And guess what? That's partly what the MBTI is designed to help you start doing]