Which is NOT a way to ensure that the results from usability testing are consistently reliable?

Coaching users during usability testing compromises the reliability of the results because it introduces bias. The goal of usability testing is to observe natural user behavior and identify genuine pain points or areas of confusion, not to guide users to a predetermined outcome. If users are coached, their actions and feedback will reflect the coaching rather than their unassisted experience, making the test data inaccurate and unreliable for improving the product. The other options contribute to reliable testing: avoiding hints ensures unbiased responses, using a script standardizes the testing process, and recording tests provides objective data for analysis.