The tidal effects are there (in principle) for the human body has mass. However, the tidal effects due to my computer next to me are much larger, not to forget my keyboard. They all exert more force on me than the moon does. Much too small to be of any significance. Some of the theoretical explanations cited above just hope that their readers don't know enough to laugh out loud.

Lunar phase effects are something very different. The amount of light at night differs a lot and that could in principle have an influence upon the incidence human activities needing more or less light. If there was such an influence it could easily be smaller now due to our growing independence of moon light level since many decades.

Contrary to popular belief, subjective experiences and some postings above, there is close to nil support for lunar phase influences on human behaviour from a very large number of studies. How are these studies done? For instance and mostly, like Donuel recommends: If you survey police blotters and hospital logs you will discover full and new moons have twice the business. Only that what Donuel thinks would be found isn't.

The persistence of (not only) Donuel's belief in humans contrary to data won by counting and comparing is in my eyes the most interesting aspect in this discussion. Many surveys have been done in the very same wards in which e.g. the nurses said they knew for sure that there was e.g. an excess of births. A look at the very same data from which they formed their impression showed that what they knew just wasn't there.

It has been done worldwide, with data across decades for a whole country, with many dependent variables (murder, births, lunacy,...). Nothing, or, perhaps, a very tiny effect.

A metaanalysis of all available studies at that time:
Rotton and Kelly, Much ado about the full moon, Psychological Bulletin 1985, 97, pp. 286-306.
I haven't seen new research sionce than that add something new to the conclusions of the authors.

Two of the factors that contribute to the belief of a correlation in absence of a real correlation:
(1) Humans often do not know exactly when there is full moon or not. They often also consider the night before and the night after as 'full moon'. Now that's three nights instead of one and if you compare the number of incidents on three nights with 1/28 you get a spurious excess of hits.
(2) Some studies with positive results have made a fatal mistake: They looked at data only from a short time and overlooked a counfounding. Since the moon phase is close to being a multiple of the number of days in a week, when the full moon in one month is at a weekend, the next full moon probably is as well. It is well known that suicides, murders, onset of psychiatric admission etc. vary largely with the day of the week, more of those incidents happening at the weekend. If you now look for moon influences over a period of let's say three months, all full moons during that period might have been midweek or weekend. You'd publish then as a moon phase effect what has its origin with the weekly ups and downs of human activity. Any moon phayse study below 18 months observation time has to be discarded for that reason.

Wolfgang