I am going to add one more point: whenever you talk about voltage, current, power factor, etc. for an AC circuit, the numbers that you are using are various types of _averages_ taken over time.
For example, when we say '120V', this is really the 'root mean square' average taken over at least one AC cycle. The actual _instantaneous_ voltage is continuously changing, going from roughly -170V to +170V. Similarly, we report a single number for current, which is really a form of average.
Okay, for _DC_ circuits, and for _instantaneous_ measurements, power is _always_ simply volts times amps. No power factor, no RMS, no nuttin'. But when the voltage and the current are constantly changing, the only way to figure out power is to multiply _instantaneous_ voltage and _instantaneous_ current to get _instantaneous_ power...then you have to average that power result over time, and you are back to the averages used to describe AC circuits.
If you look at an AC circuit with a pure resistive load, you will find that the current is precisely in phase with the voltage. The instantaneous power is simply the product of the two at each moment in time...but since the voltage and current are proportional, you will find that the instantaneous power is proportional to the _square_ of each. So for a pure resistive load you can simply multiply the RMS voltage and the RMS current.
But when the circuit has reactive components, you will find that the voltage and current are no longer in phase. The peak voltage point in the cycle does not correspond to the peak current point, and for portions of the cycle, the voltage is positive when the current is negative, or vice-versa. What this means is that for portions of the cycle, the power is _negative_.
Now don't get scared by 'negative' power; this is really just accounting. Negative power is still real juice flowing; it simply means that the net power is flowing in the opposite direction, from the device that is nominally the load back to the device that is nominally the source. This happens all the time, for example, with an induction motor attached to an overhauling load, where the motor becomes a generator and supplies power back to the mains.
When you have a load with reactive components, what happens is that for portions of the AC cycle, power flows to the load. Some of it gets used up in the load, and some of it gets stored in the reactive portion of the load (building up a magnetic field or charging a capacitor plate). During other portions of the AC cycle power actually flows back from the load to the source, being supplied by this stored energy. The net result is that more total current is flowing in the wires than needed to deliver power to the load itself.
We use the term power factor to correct the results that we get from multiplying the _averaged_ voltage and current (using RMS average) in order to find the true power being delivered to the load.
If you remember that the values for voltage/current/power factor/crest factor/etc. are all tools to let you use _average_ values to describe continuously varying quantities, then the whole thing of power factor becomes much less mysterious.
For example, above I described reactive power factor. But the term power factor is actually used whenever the product of RMS voltage and RMS current does not yield power delivered to the load.
With many power supplies, the input rectifier only draws current during the peak of the AC voltage cycle. During this 'on' period, the current is _much_ higher than average. The heating in the wires is accurately determined by the RMS current flow, but the power delivered to the rectifier is not correctly given by RMS current times RMS voltage. Again power factor could be used to describe this. Note, however, that this is neither a capacitive nor an inductive power factor, and normal power factor correction techniques wouldn't help. Thus a _different_ correction factor is sometimes used: 'crest factor' which describes the _shape_ of the current waveform.
Regards,
Jonathan
