**Principal investigator(s):** **Irina Iles** National Cancer Institute Email: irina.iles@gmail.com **Arielle S. Gilman** National Cancer Institute Email: arielle.gillman@nih.gov **Lauren O'Connor** National Cancer Institute Email: lauren.oconnor@nih.gov **Rebecca Ferrer** National Cancer Institute Email: rebecca.ferrer@nih.gov **William Klein** National Cancer Institute Email: kleinwm@mail.nih.gov **Sample size:** 1027 **Field period:** 12/10/2020-05/06/2021 **Abstract** Individuals are regularly exposed to conflicting nutrition information that can negatively affect perceptions of health science as well as individual behaviors. Research on how individuals respond to different types of conflicting information regarding nutrition is limited, however. The main goal of this study is to expand research on conflicting information by testing the effects of three different types of conflicting information that are often encountered in the communication of scientific findings – conflict in evidence (source A and source B agree the evidence is mixed), conflict between expert sources about the evidence (source A and source B present conflicting evidence about a topic), and conflict within the same expert source (source A changes its own recommendation about the evidence) – on several key responses explored in past research (confusion, trust in scientists, affective reactions, and behavioral intentions to follow topic-specific and non-specific recommendations). We specifically focus on conflict arising between new scientific findings and an established body of evidence in a domain that constitutes the basis for public health guidelines. This is because public health guidelines are critical to preventing and controlling disease and loss of adherence to such guidelines (e.g., cancer screening guidelines; nutrition guidelines) by the public could substantially increase disease burden (Simms et al., 2020). In this study, we focus on nutrition guidelines as an important cancer prevention tool (Islami et al., 2018). **Hypotheses** H1. Consistent with previous research, we hypothesized that exposure to any conflicting information (vs. control), as well as exposure to each conflicting information type (vs. control), will lead to higher nutrition confusion (i.e., higher confusion about nutrition recommendations, lower perceptions of scientific consensus about the amount of red meat one should consume; higher cancer fatalism), higher negative affective reactions, lower trust in scientists, and lower intentions to follow (topic-related and topic-unrelated) public health recommendations. H2. Because people prefer conflict in evidence vs. conflict between sources about the evidence (Smithson, 1999), we hypothesized that exposure to conflict between sources (vs. conflict in evidence) will lead to higher confusion, higher negative affective reactions, lower trust in scientists, and lower intentions to follow (topic-related and topic-unrelated) public health recommendations. RQ1. We further explored whether a public health organization changing its own recommendation (vs. conflict between sources and conflict in evidence) had different effects on confusion, affective reactions, trust, and behavioral intentions. RQ2. As a sensitivity analysis, we also examined if the above comparisons were significantly modified by the direction of the initial recommendation to parse whether it was exposure to conflicting information that caused reactance or only conflicting information that suggested a behavior change. **Experimental Manipulations** A representative sample of 1027 US adults was randomly assigned to read and respond to 1 of 8 instances of conflicting information about nutrition and cancer risk. The messages were introduced as news articles encountered while browsing the internet and depicted 1 of 4 conflicting information types (conflict in evidence - sources A and B agree the evidence is mixed; conflict between two expert sources - sources A and B present conflicting evidence about nutrition and cancer risk; conflict within the same expert source - source A changes its own recommendation about the evidence; no conflict control) crossed by 1 of 2 baseline recommendations with which new information conflicted (limit vs. do not limit red meat intake to reduce cancer risk). **Outcomes** Participants first rated on a 7-point Likert scale several items that captured confusion regarding the state of science on nutrition and cancer: 1) confusion about nutrition recommendations (six items; e.g., It is not always clear to me what food are best to eat; α=0.76) (Nagler, 2014); 2) perception of the scientific consensus about the amount of red meat they should consume (one item: “The scientific community is in agreement about the amount of red meat you should consume.”) (Cook et al., 2017); and 3) cancer fatalism (one item: “There’s not much you can do to lower your chances of getting cancer.”) (National Cancer Institute, 2008). Next, participants reported affective reactions by rating on a 5-point scale how much of three negative emotions (frustrated/annoyed/distressed) they felt as a result of the article they read (α=0.85) (Nagler et al., 2019). Next, participants reported their trust in scientists or science experts by rating four adjective pairs on a 7-point scale (non-collaborative/collaborative; incompetent/competent; untrustworthy/trustworthy; not protective of public interest/protective of public interest) (α=0.86) (Jensen & Hurley, 2012). Finally, participants reported their intentions to follow (topic-related and unrelated) public health recommendations: plans about red meat in their diet over the next month (1 – eat a lot less to 7 – eat a lot more; and 8 – “I don’t eat red meat now and will not eat red meat in the future”) and how likely they were to have five or more servings of fruits and vegetables most days in the next month (5-point scale) (Nagler, 2014). **Summary of Results** Compared to the control, each conflicting information type led to lower perceived scientific consensus about how much red meat one should eat (p<.001); conflict in evidence (p=.004) and between sources (p=.006) led to lower trust in scientists. Intentions to consume red meat more frequently were higher in the conflicting information conditions than control in the group initially told to “limit red meat” and lower in the “do not limit red meat” group (p=.022). Conflict within the same source led to higher perceived scientific consensus compared to conflict in evidence (p=.007) and between sources (p=.013); it also lowered intentions to consume red meat more frequently compared to conflict in evidence, but only in the “do not limit red meat” condition (p=.033). Conflict in evidence (p=.007) and within the same source (p=.013) increased cancer fatalism compared to conflict between sources. **References** Iles, I., Gillman, A.S., O’Connor, L.E., Ferrer, R.A., & Klein, W.M.P. (2022, April). Effects of communicating different types of conflicting information about nutrition and cancer risk. Abstract accepted for presentation at the Society for Behavioral Medicine 43rd Annual Meeting & Scientific Sessions, Baltimore MD