The validity of Criteria Based Content Analysis (CBCA) as a tool for differentiating between fabricated and experience-based statements has proven robust in several meta-analyses. However, while field studies lack ground truth, lab studies regularly fail to consider relevant forensic aspects, such as the voluntary nature of lying. Previous lab studies neglected this notion of self-selection and requested fabricated statements from all participants instead. Motivational research, however, would suggest that voluntary liars are more motivated and put more effort into their fabricated statements. Involuntary liars, in contrast, might even show reactance. Consequently, voluntary lies would be expected to be of a higher quality than involuntary lies. This, however, could threaten the validity of content-related lie detection methods such as CBCA. In its worst case, CBCA (and similar methods such as RM and SCAN) might not distinguish between experience-based statements and voluntary lies – those lies that matter most in the real world. I will present two studies addressing this issue. In Study 1, I provided participants (n = 462) with a scenario that was previously used in lie detection research and openly asked them how they would react in this situation. Only about 20-30% of all participants indicated that they would have fabricated a statement. In the preregistered Study 2 (n = 445), I analyzed the quality of voluntary and involuntary lies and compared them to the quality of experience-based statements. Several implications for lie detection research seem to follow; practical implications will be discussed.