People often learn about science from various sources including scientists, journalists, and friends. Many studies assume people pay different levels of attention to expert and non-expert sources. This foundational assumption has largely been tested with selective exposure and reading time measures. In Study 1, we used eye-tracking to measure attention and found that individuals paid more attention to experts than non-experts. The results are promising as it suggests that people can discriminate between expert and non-expert information. But in Study 2, we showed that expertise cues do not survive person-to-person transmission via serial reproduction. Our studies highlight the need to use new methods to validate key theoretical assumptions about attention and the presence of expertise cues during social transmission.