In the NLP practitioner course I am taking, it talks about how people are more receptive to either the visual, acoustic, or kinesthetic representational system. That is some people are more likely to respond to sights(words that conjure images), vs sounds (words that have certain tone to it), vs feelings (words that emphasize feelings). And it states that we use predominately one of the 3 and that effective communication is based on knowing what type your partner is.

Does any one have any ideas about this and how it works and if it is valid or not?

It's a half truth that has been around for YEARS. It can be a useful model to help build skills but think of it as training wheels.
In reality, you'll be hard pushed finding somebody who isn't predominantly visual, whether they know it or not. It's more useful, as you progress, to match their representational systems the same way they do. Or to start mixing them together, overlapping each other. Later on it can be useful to help you better understand Dynamic Mental Imagery, or Revivification and virtually any kind of induction/trance work.
It's all down to what you're trying to achieve, but for now, you may as well get sensitive towards them and do it in the way your training provides.

Just don't get caught up in eye cues

Has there been any empirical research on this showing that people do use different senses more often than other senses? and that by communicating using the same sense as the people they better understand our perspective on things?

After all, NLP tries to teach us to better communicator.

Nope. There is no research that *I'm aware of* that supports it. There is research that has been done which shows no evidence of any relationship there.
Truthfully, it's often not useful to think of it as a specific method/technique, but to think of it as just a way of communicating, with no inherent value. I view them as just that, a different way to bring attention to certain senses. None are better or worse, it's all down to appropriate or innapropriate.
Over time you'll notice what effect each rep. system has, and which is appropriate, but to be honest it's very difficult for me to distill why they are appropriate or not at any time. Each sense has a different effect at different times for different people, it's all down to what you're trying to achieve.

DISCLAIMER: I don't use much NLP, at least the way most people think it works.

Do you still have the name of the research that refuted the support for this? I want to have a more thorough read on this.

Aside, when I was referring to validity, I meant like do people actually communicate more effectively (gets point across easier) when talking with the same representational system? That is the point that this NLP topic in the course is trying to get across. What are your thoughts on this one? and in regards to research evidence on this, was it also not supported?

I can't remember the research but it should be easy to find. Just type "represenational systems NLP" into google or something. Think it references it on Wiki too.

As far as it making you communicate better, I'm not too sure. I have matched rep. system succesfully in the past to induce trance or when having a conversation, but I'm not sure if there is a way of quantifying how much the rep system affected it and how much was down to the other factors, or even just the use of different senses alone (and the pacing and leading that naturally flows from it).
I've never used it to match somebody's supposed "primary rep. system", and I don't believe that works.

If I'm honest, I know there was a bit of controversy about the testing methods involved in the research, but I don't know to what extent, and *usually* in cases like this I trust the scientist.