What kind of data collection method will you be using? Also, you might consider what are the standard practices in your field; in particular, if reactivity is difficult to escape with the most commonly used methods, then it is likely to be widely accepted.
I am not certain you can control that. Seems like the best thing you can do is just record what you observe, unless you intend to do a study where some participants know they are involved in your research and others don't which would make it unethical. You have to trust that your participants are being honest.