Operationalising rewards and costs is hugely subjective, making comparisons between people and relationships in controlled settings very difficult. Most studies which are used to support Social Exchange Theory account for this by using artificial procedures in laboratory settings, reducing the external validity of the findings