I never actualy tryed measure speaker DC vs AC resistance aka impedance, but i do pretty much know that they can be different in sense on several decades.
It all depend on material and construction of speaker driver and crossover alsso add little thing to it.
In a coil or capacitor, the reactance changes due to the way the component works with electricity. A capacitor, will increase its resistance as the frequency gets lower. A coil on the other hand, will increase it's resistance as the frequency gets higher. They will both continue like this until their resistance is so very high, that no useable current will pass, at which point they are said to be "saturated". In a normal resistor, at any given frequency, we can say that it's resistance is still the same. When graphed, the value of a resistor will be a straight line across the graph. In a reactive component, however, the line will gently slope upward, or downward, with frequency.
This is reactance. Reactance is necessary for crossovers to do their job. Whenever you look inside a crossover (passive crossover, at least), you will see usually nothing more than a few coils and capacitors, and occasionally the odd sand block resistor. Remember, coils resist high frequencies, and capacitors resist low frequencies. When the two are combined, they form crossover networks. A crossover network typically uses a capacitor to keep low frequencies from going to a tweeter, and a coil to keep high frequencies from going to a woofer.
Another component that has reactance, is the speaker itself. A speaker's voice coil behaves electrically just like a coil in a crossover network. Because of this, speaker designers face special problems when designing midranges and tweeters based on voice coil drivers. Also, due to it's reactance, a speaker is almost never at it's rated "Impedance" (a word often used incorrectly by speaker manufacturers). When a speaker is measured at 4 ohms, it is measured using a device that puts out DC current to do the measuring. The only other time the subwoofer will have anything near 4 ohms is when it is at resonance. The rest of the time, the voice coils impedance is very high.
Crossover should be designed well enough that it actually prevent speaker goes into resonance.
Even if you measure DC resistance around 3 ohms, it still doesn't meant it will not have more then 8 ohm when driven by your receiver.
Did OP actually measured resistance of whole speaker including crossover or took speaker apart and measured driver coils separately ?
Only really revealing way how to measure speaker impedance is to use frequency generator and measure current and voltage using AC signal with several different frequencies.
This is actually what speaker manufacturers or test labs do, and 4/6/8 ohm impedance specified by manufacturer is simple "average or approximated" value of measured impedance in full frequency spectrum they should be able provide.