In order to keep the development of a new product secret from competitors, convenience sampling within a company may be unavoidable. Convenience sampling, such as recruiting friends or coworkers, can introduce biases into the data collection process. These participants could be distinctly different from the rest of the population and the relationships between the interviewee and the interviewer could skew what kinds of feedback can be collected. In our study, we recruited friends of six interviewers to participate in a usability test on a high-fidelity prototype that was designed to be frustrating in different ways. The goal was to investigate what features of friendships and of the interviewee determine how likely one is to report frustration in an “objectively” frustrating prototype. The interviewees reported their ease and frustration with requested tasks, comfort in being honest and how much they adjusted their ratings from the truth. We collected friendship details from both the interviewer and the interviewee and categorized those friendships to be mutual or lopsided/one-sided (friendship type) based on how much their responses agreed with one another. We identified certain qualities of friendships rated by the interviewee (companionship, transcending problems) and by the interviewer (aid), or combining both (reflected appraisal) that predicted the interviewee’s frustration ratings, but did not find any differences in friendship type, gender dynamics or big five personality traits. We discuss implications for selecting participants.