Predictive validity is the extent to which a theory or study allows us to predict a person's behaviour. In the case of Gottman's research, he argues that he can predict if a relationship will survive by carrying out an interview with the couple. There is a problem with the predictive validity of his research because the couples that come to him are already in crisis. He then debriefs them on the research and tells them about the state of their relationship. They often then get a divorce. It is questionable to what extent his research actually does predict the likelihood of a relationship ending, even though his findings make sense to us.

When is a study is retrospective, it must create a "history" of a behaviour from the memories of those involved. In this case, couples must talk about the history of their relationship. This is then open to memory distortion as well as demand characteristics. There is also often no way to verify the claims made by the participants.

According to research by Gottman, contempt not only makes an individual feel inferior, but long-term exposure to contempt can also have a negative effect on a person's immune system.

A relationship-enhancing pattern of communication is one in which we do not blame our partners when bad things happen, and we give them credit when good things happen.

Disclosure is a difficult thing to measure. There also appear to be gender and cultural differences in what is expected with regard to the level of disclosure in a relationship. Finally, disclosure is often determined by self-reported data.

Much of the research on relationships is done on young couples. More often than not, it is young women from an individualistic culture. So, there is a gender, age, and culture bias in the samples.