You need to understand that polygraph "testing" has
no scientific basis. It is a
fraudulent procedure that depends not on science, but on
trickery, as it entails the polygraph operator lying to and otherwise deceiving the person being "tested."
While the polygraph operator you hired will have told you that all questions must be answered truthfully, and that any hint of doubt or deception will be detected by his highly sensitive polygraph instrument, he will have secretly assumed that your answers to certain questions, called "control" questions, were less than truthful.
Your reactions to the question, "Since July have you had any sexual encounter with any other woman other than your partner?" will have been compared to your reactions to a question like, "Have you ever deliberately lied to someone who loved and trusted you?" You will have been steered into a denial regarding this latter question, perhaps with a warning that the kind of person who would lie to someone who loved and trusted them is the same kind of person who would cheat on a partner and then lie about it. If you made admissions to having lied, after a sermon about honesty, the question would have been rephrased to something like, "Other than what you told me, did you ever deliberately lie to someone who loved and trusted you?" Your denial at this point is
still assumed to be a lie, or at best an answer about which you cannot be completely certain and which should cause a reaction.
If the relevant question about infidelity causes you to breath more erratically, to sweat more profusely, or causes your heart rate and blood pressure to change more so than does the "control" question about lying to someone who trusted you, then you fail. Conversely, if you show stronger reactions to the question about lying to someone who loved and trusted you, then you pass.
The procedure also includes irrelevant questions such as "Are the lights on in this room?" The polygrapher falsely explains that these questions provide a baseline for truth, as the true answer is obvious. (The polygrapher may have also falsely described these questions as "control" questions.) But in fact, such irrelevant questions are not scored, and they merely serve as "buffers" between pairs of relevant and "control" questions.
As I mentioned before, this simplistic procedure has
no scientific basis, and to make matters worse, it's inherently
biased against the truthful. Think about it: the more honestly one answers the control questions, and as a result feels less anxiety when answering them, the more likely one is to fail!
For a thorough debunking of polygraphy, and a more detailed explanation of polygraph procedure, see our e-book,
The Lie Behind the Lie Detector (1 mb PDF).
I think that rather than throwing good money after bad for another séance with a polygraph chartgazer, you'd be better off educating yourself (and your girlfriend) about polygraphy. See the statement of Gary Smith for an example of the harm that misplaced faith in the pseudoscience of polygraphy has caused to couples' relationships:
https://antipolygraph.org/statements/statement-028.pdf