In the reviews received from Environmental Communication, Reviewer 1 asked why we did not conduct moderated mediation analyses or MANOVA given our hypotheses (particularly that we were questioning the gateway belief model) and that our DVs are likely correlated. We respond with the following points: 1. The gateway belief model depends, in part, on the relationship between the presence of a consensus message (compared to a control condition) and an increase in the perceived number of scientists who agree about the issue. This is because perceptions of scientific agreement is thought to mediate the relationship between the manipulation and beliefs about the topic (climate change in their model, and GMOs in ours). However, there is not a relationship between the presence of the consensus message (we used the NASEM message to test this) and the percent of scientists perceived to agree. Thus, the model would not be supported. 2. We did not conduct a more complex moderated mediation model based on our hypothesis that pre-existing anti-GMO attitudes (measured by preference for non-GMO labeled foods) would interact with the message to influence whether participants believe the message is representative of what the scientific community as a whole believes, which would then influence perceptions of the % of scientists who agree with the statement, and then post-message GMO concern. However, we did not measure the extent to which people believe the message is representative of the scientific community in the control condition for which the message discussed the history of baseball. Therefore, our only point of comparison would be the pro consensus statements versus the anti-consensus statements. 3. We hadn't thought about conducting a MANOVA. However, the MANOVA can only tell us whether there is an overall effect among conditions, we would still need to conduct the post-hoc tests at the level of each DV to determine where differences exist. Thus, it seems that the power conserved for the use of the MANOVA is then lost when having to conduct the follow-up tests. Thus, it appears to us that the most conservative estimate is to simply conduct the ANCOVAs. Although these were our reasons not to include those analyses in the body of the paper, to show that we've made a good faith effort to address reviewer concerns, we include the requested analyses here as part of our online supplementary materials.