"Gambler's fallacy" refers to a mistake that people often make in assessing the probability of a particular outcome from a repeated random action. For example, a coin toss actually has a fifty-fifty chance of either a heads or tails outcome every time the coin is flipped. However, most of us tend to think that a greater probability of a heads would follow . . . oh, say a run of twenty tails in sequence.
Unless there is something wrong with the coin or the flipper (and the process is not truly random), the chance remains fifty-fifty even after a series of highly improbable outcomes, such as 1,000 heads in a row. Outcomes do not become any less random or any more predictable just because the process is repeated.
This piece from NPR's Shankar Vedantam suggests that current popular myths about interviewing for a job might have basis in the gambler's fallacy. Is it better to have your interview scheduled late in the day? Perhaps so--if you happen to follow a series of poorly-perceived candidates. The interviewer is likely to commit this fallacy and assume that a series of tails (bad interviews) should not persist. By following a string of unqualified wannabees you will look good if only because the interviewer is mistakenly inclined to expect a heads (favorable interview) to be more likely.
I suspect that some interviewers think this way. But I have also (personally) observed another kind of bias among interviewers. I think of it as a failure to "zero-out" all of the registers or re-initialize all of the variables from one interview to the next. What happens, it seems to me, is that an interviewer carries impressions of what occurred in the first interview forward into the second, third, and so on.
For example, if the first candidate was moderately bright, outgoing and friendly, then (at least for a while) subsequent candidates seem brighter, more outgoing and friendly. Once the interviewer loads these values into his mental variables they tend to remain there, influencing perceptions until something forces a reset.
This pattern persists when candidates are generally similar; extreme cases tend to force a reset. A seriously sullen candidate will likely erase the predilection to regard candidates as friendly, and may foster a new tendency to see them as moody.
When you have an interviewer like this, that cheerful, confident candidate who breezed through the door just in front of you may have done you a favor instead of creating an expectation that you will land as a tails.
Deciphering Hidden Biases During Interviews