Well, I see 2 options :
- You evaluate the results
- You evaluate the path to the result
If you choose to evaluate the result, you will focus on the retention rate. This means that you will have to use "hooks" in the conversation that you can come back to and see whether the person retained the information, and in which level of detail.
The other option is to focus on the words used. Someone visual will most likely use words like "see", "picture", maybe even draw something (on a board, on the floor, or just using hands). The details of the story will be visual.
Kinestetic people will focus more on sensations. I guess audio will focus on sounds and maybe speak faster ?
So when describing the place where they spent their vacation :
- Visual : We were in a house with ancient stone walls, wooden doors and red roof
- Kinestetic : this was a peaceful place, the woods around the house were ideal to relax and the fresh air filled me with energy in the morning and we had a fire to keep us warm in the evening.
- audio : we could hear the birds every morning. It was so silent at night that you could notice the noise of the wooden floor.
I believe nobody is a pure visual/audio/kinestetic person. Everyone is a bit of each, so usually the best technique is to combine everything (particularly when speaking to several persons like when doing a public speech).
Also if you use the wrong style, the person will most likely drop you hints (such as drawing something on paper, saying to slow down...). In this case, the best strategy is always to listen and adapt. If you see something fails, try something else.
So I would base myself more on the reaction that the type of learning, which is IMO secondary if you get the message across anyway.
I believe there are some statistics somewhere, if memory serves I think most males are visual learners for example. You might also find some disparities depending on the job the person does. I expect a mathematician to be visual, while a musician would be audio for example (though this is not a bulletproof rule at all).
I don't have anything more to back my thoughts than my own experience with it, but maybe someone has some study or document to share on the topic ?
Does it answer the question, or did you have some specific problematic in mind when writing your message ? Maybe with some more context it would be possible to narrow down the answer.