Power Analysis A power analysis indicated that a sample size of 967 would be necessary to detect a statistically significant interaction effect in a 2 (direction of comparison; upward vs. downward) x 3 (source of comparison; internet, friends, pornography) ANOVA, assuming a small effect size (f = .10) with a power level of 0.80 (power estimated using GPower 3.1; Erdfelder, Faul, & Buchner, 1996; Faul, Erdfelder, Lang, & Buchner, 2007). The target sample for a medium effect size (f = .25) with 80% power is only 158, so we estimate a small effect size to be conservative. Planned Sample Participants will be recruited from Amazon's Mechanical Turk (MTurk), an online crowdsourcing platform that is commonly used for psychological research. An advertisement will be placed on the MTurk website for all MTurk users with active accounts to see, with information about the inclusion criteria as well as a link to the survey. Participants must be at least 18 years of age, speak English fluently, be involved in a sexually active romantic relationship, and have an MTurk approval rating of 97% or more to be eligible for the study. Eligible and interested participants will follow the link, and will be redirected to a survey hosted on Qualtrics, where the letter of information and consent will be presented. Informed consent will be received from each participant digitally (each participant must indicate they read the consent form and agree to take part before proceeding). Participants included in the final sample must correctly answer 2 attention checks and 2 instruction checks. To account for loss of participants due to factors such as failing attention checks, our target sample prior to these cuts was approximately 1150 participants. Importantly, no analyses will be conducted until an approximation of the final target sample (*n* = 967) has been acquired.