Often, in any case

Tom Stafford has a BBC article about men and sex:

We've all been told that men think about you-know-what far too often; every seven seconds, by some accounts. Most of us have entertained this idea for long enough to be sceptical. However, rather than merely wonder about whether this is true, stop for a moment to consider how you could or could not prove it.
If we believe the stats, thinking about sex every seven seconds adds up to five hundred-plus times an hour. Or approximately seven thousand times during each waking day. Is that a lot? It sounds like a big number to me, I’d imagine it’s bigger than the number of thoughts I have about anything in a day. So, here’s an interesting question: how is it possible to count the number of mine, or anyone else’s thoughts (sexual or otherwise) over the course of a day?
The scientific attempt to measure thoughts is known to psychologists as "experience sampling". It involves interrupting people as they go about their daily lives and asking them to record the thoughts they are having right at that moment, in that place.
Terri Fisher and her research team at Ohio State University did this using 'clickers'. They gave these to 283 college students, divided into three groups, and asked them to press and record each time they thought about sex, food, or sleep.
Using this method they found that the average man in their study had nineteen thoughts about sex a day. This was more than the women in their study, who had about ten thoughts a day. However, the men also had more thoughts about food and sleep, suggesting perhaps that men are more prone to indulgent impulses in general. Or they are more likely to decide to count any vague feeling as a thought. Or some combination of both.
The interesting thing about the study was the large variation in number of thoughts. Some people said they thought about sex only once per day, whereas the top respondent recorded 388 clicks, which is a sexual thought about every two minutes.
However, the big confounding factor with this study is "ironic processes", more commonly known as the "white bear problem". If you want to have cruel fun with a child tell them to put their hand in their air and only put it down when they've stopped thinking about a white bear. Once you start thinking about something, trying to forget it just brings it back to mind.
This is exactly the circumstances the participants in Fisher's study found themselves in. They were given a clicker by the researchers and asked to record when they thought about sex (or food or sleep). Imagine them walking away from the psychology department, holding the clicker in their hand, trying hard not to think about sex all the time, yet also trying hard to remember to press the clicker every time they did think about it. My bet is that the poor man who clicked 388 times was as much a victim of the experimental design as he was of his impulses.
Another approach, used by Wilhelm Hoffman and colleagues, involved issuing German adult volunteers with smartphones, which were set to notify them seven times a day at random intervals for a week. They were asked to record what featured in their most recent thoughts when they received the random alert, the idea being that putting the responsibility for remembering onto a device left participants' minds more free to wander.
The results aren't directly comparable to the Fisher study, as the most anyone could record thinking about sex was seven times a day. But what is clear is that people thought about it far less often than the seven-second myth suggests. They recorded a sexual thought in the last half hour on approximately four percent of occasions, which works out as about once per day, compared with nineteen reported in the Fisher study.
The real shock from Hoffman's study is the relative unimportance of sex in the participants' thoughts. People said they thought more about food, sleep, personal hygiene, social contact, time off, and (until about 5pm) coffee. Watching television, checking email, and other forms of media use also won out over sex for the entire day. In fact, sex only became a predominant thought towards the end of the day (around midnight), and even then it was firmly in second place, behind sleep.
Hoffman's method is also contaminated by a white bear effect, though, because participants knew at some point during the day they'd be asked to record what they had been thinking about. This could lead to overestimating some thoughts. Alternately, people may have felt embarrassed about admitting to having sexual thoughts throughout the day, and therefore underreported it.
So, although we can confidently dismiss the story that the average male thinks about sex every seven seconds, we can't know with much certainty what the true frequency actually is. Probably it varies wildly between people, and within the same person depending on their circumstances, and this is further confounded by the fact that any efforts to measure the number of someone's thoughts risks changing those thoughts.
There’s also the tricky issue that thoughts have no natural unit of measurement. Thoughts aren't like distances we can measure in centimeters, meters and kilometers. So what constitutes a thought, anyway? How big does it need to be to count? Have you had none, one or many while reading this? Plenty of things to think about!

Rico says he'd have voted for more often, but maybe not...