Support for “doing your own research” is associated with COVID-19 misperceptions and scientific mistrust

---

Implications 

Concerns about misinformation are prevalent among the U.S. public and elsewhere in the world (Berthel et al., 2016). People are aware of the need to be skeptical consumers of information and that some information they encounter may be misleading or false (Guess et al., 2020). Yet there is also a rising sense of distrust in experts, including scientists and journalists, as trustworthy sources of information (Rainie et al., 2019) and vocal skepticism of information about public health, novel technologies, and scientific research (Merkley, 2020; Rutjens et al., 2022). Though trust in science remains high (Funk et al., 2020), a growing number of people assert it is better to “do your own research” (DYOR) than rely on information provided by scientists, doctors, or other credentialed experts (Carrion, 2018). 

The phrase “do your own research” (DYOR) was made a slogan by American conspiracy theorist Milton William “Bill” Cooper in the 1990s (Ballantyne et al., 2022) and has been frequently used by anti-vaccine advocates online since the 2010s (Kata, 2012). These calls encourage individuals to seek additional or alternative sources of information, verify facts, and examine evidence to make informed decisions that best suit one’s individual circumstances. However, DYOR calls became more visible to the broader public during the COVID-19 pandemic, where they were common in online content about QAnon, a conspiratorial group describing themselves as a “community of researchers” dedicated to uncovering hidden truths behind current events (Marwick & Partin, 2022; Pasquetto et al., 2022). 

We leveraged panel survey data to investigate how perceptions of “doing your own research” were associated with trust in science, beliefs, and concerns about COVID-19. Our results revealed that positive DYOR perceptions were associated with more COVID-19 misperceptions and less trust in scientific institutions, even when controlling for previous beliefs and levels of trust. This may be because DYOR reflects epistemic beliefs that one’s gut feelings and “alternative facts” from independent research are as or more legitimate means of forming beliefs as scientific evidence and that truth is relative to each person’s unique experience (Carrion, 2018; Kata, 2012; Marwick & Partin, 2022; Pasquetto et al., 2022). As people more strongly adopt DYOR epistemic beliefs, they may have greater justification for engaging in motivated reasoning and cherry-picking expert voices to support appealing beliefs, even when those beliefs oppose a consensus of institutional experts. 

The panel analyses found DYOR perceptions only explain about 1% of the variance in both trust in science and COVID-19 misperceptions when controlling for prior levels of the dependent variables (and about 5% of the variance in cross-sectional analyses). This suggests that potential effects of DYOR perceptions are small but may accumulate over time. Though survey waves were only three months apart, DYOR perceptions were nevertheless associated with decreased trust in science and increased misperceptions in that short time. Over more months or years, positive DYOR perceptions may have substantial influence on trust in institutions and false beliefs across a range of issues. 

These findings have several implications for understanding how people make sense of contemporary information environments. First, though we do not identify what kinds of research or information seeking people may do, it suggests that favorable DYOR perceptions may be associated with biased information processing or seeking untrustworthy information. People often overestimate their abilities to seek and interpret information (e.g., Atir et al., 2015), and tend to search for information that aligns with preexisting values, beliefs, and identities (Lord et al., 1979). Perceptual biases can lead to inaccurate conclusions, particularly when individuals lack topic knowledge, training in scientific methods (Ballantyne & Dunning, 2022), or rely on gut feelings (Garrett & Weeks, 2017).

Second, calls to DYOR may promote skepticism of science institutions and mainstream information sources by highlighting how they might mislead the public. These calls may reflect perceptions that science institutions or mainstream news media are corrupt or have a hidden agenda hostile to one’s worldviews and goals (Kata, 2012; Perloff, 2015). While DYOR messages are not exclusively used to promote misinformation (e.g., some emphasize the safety of COVID-19 vaccines; Kelsey, 2021), our results show that DYOR perceptions are associated with distrust of science institutions and COVID-19 misperceptions. It is possible that even well-intentioned calls to DYOR could foster skepticism of official information sources. 

Finally, doing your own research may not be about seeking information but instead be expression of anti-establishment political identity. In cases of anti-vaccine sentiment, DYOR is often invoked in conjunction with resentment towards doctors and scientists who dismiss personal experience and intuition (e.g., Carrion, 2018). DYOR may be a manifestation of contempt for elites rather than an affirmation of the importance of independent research.

 “Doing your own research” is both a prudent and flawed endeavor. Informed and engaged individuals are widely considered normatively good in health, science, and politics (e.g., Korda & Itani, 2013). Marginalized groups who have legitimate reasons to be wary of institutions often benefit from outside research and expertise (e.g., Bond et al., 2021). Further, blind trust in experts can lead to politicization of science and policies that do not align with public interest, making healthy skepticism and independent research democratically desirable (Krause et al., 2021). Rather than discouraging DYOR, which could elicit reactance for infringing on freedom of thought (Brehm, 1966), it may be fruitful to advance media literacy interventions that address information seeking biases and provide logic-based corrective interventions about rhetorical strategies, including DYOR messages, that might promote misinformation (Vraga et al., 2020).     

Scientific research and institutions are not without problems, from news biases to structural discrimination in healthcare (Bond et al., 2021). However, the idea that expert institutions and mainstream news have nefarious motivations can foster distrust and influence information-seeking in ways that undermine democratic norms and values (Hutchens et al., 2021). Our research provides initial support for concerns some have expressed about DYOR messages (Ballantyne & Dunning, 2022). However, while controlling for prior levels of the outcomes provides some evidence that DYOR perceptions lead to changes trust and beliefs, these data do not account for exposure to DYOR messages or whether such messages cultivate positive DYOR perceptions. Future research should explore whether support for DYOR is a response to fears about disinformation or reflective of anti-establishment superordinate worldviews, and the potential effects of DYOR messages on information seeking and attitudes about science, politics, and media.

Findings 

Finding 1: Positive perceptions of “doing your own research” were associated with less trust in science institutions.

DYOR perceptions were significantly associated with decreased trust in science institutions, b = -0.09(.02), p < .001, accounting for prior levels of trust in science.

Finding 2: Positive perceptions of “doing your own research” were associated with holding more COVID-19 misperceptions.

DYOR perceptions were significantly associated with decreased belief accuracy about COVID-19, b = -0.05(.01), p < .001, accounting for prior levels of COVID-19 belief accuracy.

Finding 3: Positive perceptions of “doing your own research” were not associated with concern about COVID-19.

DYOR perceptions had no association with either health concerns about COVID-19, b = -0.05(.03), p = .13, or economic concerns about COVID-19, b = 0.02(.03), p = .50, when accounting for prior levels of concern.

Methods

The data for this study are from a two-wave panel survey of U.S. adults administered by YouGov in December 2020 (N_W1 = 1,500) and March 2021 (N_W2 = 1015). Respondents were matched to a sampling frame on gender, age, race, and education constructed by stratified sampling from the full 2018 American Community Survey. Of Wave 1 respondents, 67.66% returned to complete Wave 2. After removing participants who failed attention checks in either Wave 1 or Wave 2, the final sample was N_W1 = 1421 and N_W2 = 975. From this sample, 54% of Wave 1 were women, and the average age was 49 years old (SD = 17.96). Sixty-five and a half percent of respondents identified as White, 11.9% as Black, 14.8% as Hispanic, 2.7% as Asian, and the remaining 5% identified as mixed race or other. In the sample, 30.7% had at least a 4-year college degree, and the median income was between $50,000-$59,000. Political affiliation was measured on a 7-point scale from “Strong Democrat” to “Strong Republican” (M = 3.57, SD = 2.14). For more information and full question wording, see Appendix A.

Perceptions of “doing your own research” were measured with three items in both waves. These measures intended to capture agreement that personal research is as valuable as expert knowledge. The items were, “Anyone can be an expert on something if they do enough research,” “I prefer to do my own research rather than rely on experts and intellectuals,” and “The opinions of people who have done their own research are just as valid as the opinions of experts and intellectuals.” Agreement was measured on a scale from 1 (“Strongly disagree”) to 7 (“Strongly agree”) with “neither disagree nor agree” as the scale midpoint. Responses were averaged into a scale of DYOR perceptions (M_W1 = 4.16, SD_W1 = 1.42, α_W1 = .80; M_W2 = 3.99, SD_W2 = 1.45, α_W2 = .82). 

Trust in science was measured in both waves with four items asking how trustworthy scientists, colleges and universities, doctors and medical scientists, and the Center for Disease Control were. Response options ranged from 1 (“Extremely untrustworthy”) to 7 (“Extremely trustworthy”). These items averaged into a scale of trust in science institutions (M_W1 = 4.79, SD_W1 = 1.39, α_W1 = .86; M_W2 = 4.74, SD_W2 = 1.42, α_W2 = .87). 

We asked participants their beliefs about nine COVID-19 statements in both waves (e.g., “It is possible to transmit COVID-19 to others even if you do not have symptoms”). Respondents indicated their beliefs on a scale from 1 (“Definitely false”) to 5 (“Definitely true”) with “unsure” as the scale midpoint. Items were coded so high values reflected accurate beliefs, then averaged into a scale (M_W1 = 3.63, SD_W1 = 0.76; α _W1 = .81; M_W2 = 3.63, SD_W2 = 0.84; α _W1 = .78). 

We had two measures of concern about COVID-19, measured in both waves. Health concerns were measured using two items asking how COVID-19 may affect the health of oneself and others. Economic concerns were asked with two items measuring concern economic impacts to oneself and community. Response options ranged from 1 (“Strongly disagree”) to 7 (“Strongly agree”). These items were averaged into a scale of health concern about COVID(M_W1 = 5.05, SD_W1 = 1.77, r_W1 = .75; M_W2 = 4.79, SD_W2 = 1.82, r _W2 = .75) and economic concern about COVID(M_W1 = 5.33, SD_W1 = 1.38, r _W1 = .43; M_W2 = 5.12, SD_W2 = 1.41, r _W2 = .40). 

To examine how DYOR perceptions were associated with trust, COVID-19 beliefs, and concern about COVID-19, we ran four OLS regression models with Wave 2 DYOR perceptions as the main independent variable and Wave 2 measures for each dependent variable. All models controlled for age, gender, race, education, political party affiliation, political interest, and news media use, including traditional news, online news, and social media news. We additionally controlled for Wave 1 trust in science, and Wave 1 measures of the respective dependent variable in each model. All analyses are unweighted.