Ecological Fallacy Research Articles

Adoption of partially automated vehicle technology features and impacts on vehicle miles of travel (VMT)

In this paper, we examine vehicle owners’ adoption of five different types of partially automated features (PAFs); lane keeping system, backup camera (BUC), adaptive cruise control (ACC), automatic braking system (ABS), and blind spot monitoring; as well as PAF effects on vehicle miles of travel (VMT). The joint modeling of PAF adoption and VMT is achieved using both individual demographic characteristics as well as psycho-social characteristics. A Generalized Heterogeneous Data Model (GHDM) is estimated, which controls for possible self-selection effects in PAF adoption based on VMT, and thus is able to provide “true” PAF effects on VMT. Our analysis specifically indicates that ignoring this self-selection can lead to a significant underestimation of the VMT increase due to PAF adoption. The results also indicate that women and older individuals (65 years or older) appear to be more inclined to invest in assistive PAFs, because of a perception that these assistive features still leave the human driver in control. However, women are less likely than men to invest in the more active ABS PAF because of heightened safety concerns with technology. In terms of PAF effects on VMT, PAFs focusing on lateral movement assistance appear to have a smaller VMT effect than those that serve longitudinal movement assistance. The highest estimated VMT change of 2,462 miles (13.8% change) is for the case when the package of BUC, ACC, and ABS is installed for middle-aged men. The highest percentage VMT change (40%), though, is for the same package of BUC, ACC, and ABS for older women. Overall, there are considerable variations in VMT impact across demographic groupings, suggesting that a single aggregate percentage improvement in safety benefits may suffer from the well-known ecological fallacy.

Association of latitude and altitude with adverse outcomes in patients with COVID-19: The VIRUS registry.

BACKGROUNDThe coronavirus disease 2019 (COVID-19) course may be affected by environmental factors. Ecological studies previously suggested a link between climatological factors and COVID-19 fatality rates. However, individual-level impact of these factors has not been thoroughly evaluated yet.AIMTo study the association of climatological factors related to patient location with unfavorable outcomes in patients.METHODSIn this observational analysis of the Society of Critical Care Medicine Discovery Viral Infection and Respiratory Illness Universal Study: COVID-19 Registry cohort, the latitudes and altitudes of hospitals were examined as a covariate for mortality within 28 d of admission and the length of hospital stay. Adjusting for baseline parameters and admission date, multivariable regression modeling was utilized. Generalized estimating equations were used to fit the models.RESULTSTwenty-two thousand one hundred eight patients from over 20 countries were evaluated. The median age was 62 (interquartile range: 49-74) years, and 54% of the included patients were males. The median age increased with increasing latitude as well as the frequency of comorbidities. Contrarily, the percentage of comorbidities was lower in elevated altitudes. Mortality within 28 d of hospital admission was found to be 25%. The median hospital-free days among all included patients was 20 d. Despite the significant linear relationship between mortality and hospital-free days (adjusted odds ratio (aOR) = 1.39 (1.04, 1.86), P = 0.025 for mortality within 28 d of admission; aOR = -1.47 (-2.60, -0.33), P = 0.011 for hospital-free days), suggesting that adverse patient outcomes were more common in locations further away from the Equator; the results were no longer significant when adjusted for baseline differences (aOR = 1.32 (1.00, 1.74), P = 0.051 for 28-day mortality; aOR = -1.07 (-2.13, -0.01), P = 0.050 for hospital-free days). When we looked at the altitude’s effect, we discovered that it demonstrated a non-linear association with mortality within 28 d of hospital admission (aOR = 0.96 (0.62, 1.47), 1.04 (0.92, 1.19), 0.49 (0.22, 0.90), and 0.51 (0.27, 0.98), for the altitude points of 75 MASL, 125 MASL, 400 MASL, and 600 MASL, in comparison to the reference altitude of 148 m.a.s.l, respectively. P = 0.001). We detected an association between latitude and 28-day mortality as well as hospital-free days in this worldwide study. When the baseline features were taken into account, however, this did not stay significant.CONCLUSIONOur findings suggest that differences observed in previous epidemiological studies may be due to ecological fallacy rather than implying a causal relationship at the patient level.

Bringing Geospatial Awareness to Community Pediatrics and Primary Care.

Bringing Geospatial Awareness to Community Pediatrics and Primary Care.

Culture and corruption: An experimental comparison of cultural patterns on the corruption propensity in Poland and Russia

The development of effective anti-corruption measures relies on a sound understanding of underlying country-specific cultural patterns of corruption. However, finding these patterns faces the problem of ecological fallacies when tracing back the results of comparative macro-studies to the national level or of using ex-post explanations for cultural variances in experimental research designs. Thus, we ask how cultural patterns can explain country differences in the propensity to act corrupt without neglecting the aforementioned problems. Based on institutional theory, we model path-dependent cultural patterns at the macro, meso and micro levels promoting propensity to act corrupt in Poland and Russia. The results of experimental data gathered from students in Poland and Russia show that the extent to which legal nihilism and ethical dualism are institutionalized at the macro level, as well as the micro factors of gender-specific socialization and studying law, has a significant effect on the propensity to act corrupt.

A Methodological Approach to Use Contextual Factors for Epidemiological Studies on Chronic Exposure to Air Pollution and COVID-19 in Italy.

The large availability of both air pollution and COVID-19 data, and the simplicity to make geographical correlations between them, led to a proliferation of ecological studies relating the levels of pollution in administrative areas to COVID-19 incidence, mortality or lethality rates. However, the major drawback of these studies is the ecological fallacy that can lead to spurious associations. In this frame, an increasing concern has been addressed to clarify the possible role of contextual variables such as municipalities’ characteristics (including urban, rural, semi-rural settings), those of the resident communities, the network of social relations, the mobility of people, and the responsiveness of the National Health Service (NHS), to better clarify the dynamics of the phenomenon. The objective of this paper is to identify and collect the municipalities’ and community contextual factors and to synthesize their information content to produce suitable indicators in national environmental epidemiological studies, with specific emphasis on assessing the possible role of air pollution on the incidence and severity of the COVID-19 disease. A first step was to synthesize the content of spatial information, available at the municipal level, in a smaller set of “summary indexes” that can be more easily viewed and analyzed. For the 7903 Italian municipalities (1 January 2020—ISTAT), 44 variables were identified, collected, and grouped into five information dimensions a priori defined: (i) geographic characteristics of the municipality, (ii) demographic and anthropogenic characteristics, (iii) mobility, (iv) socio-economic-health area, and (v) healthcare offer (source: ISTAT, EUROSTAT or Ministry of Health, and further ad hoc elaborations (e.g., OpenStreetMaps)). Principal component analysis (PCA) was carried out for the five identified dimensions, with the aim of reducing the large number of initial variables into a smaller number of components, limiting as much as possible the loss of information content (variability). We also included in the analysis PM2.5, PM10 and NO2 population weighted exposure (PWE) values obtained using a four-stage approach based on the machine learning method, “random forest”, which uses space–time predictors, satellite data, and air quality monitoring data estimated at the national level. Overall, the PCA made it possible to extract twelve components: three for the territorial characteristics dimension of the municipality (variance explained 72%), two for the demographic and anthropogenic characteristics dimension (variance explained 62%), three for the mobility dimension (variance explained 83%), two for the socio-economic-health sector (variance explained 58%) and two for the health offer dimension (variance explained 72%). All the components of the different dimensions are only marginally correlated with each other, demonstrating their potential ability to grasp different aspects of the spatial distribution of the COVID-19 pathology. This work provides a national repository of contextual variables at the municipality level collapsed into twelve informative factors suitable to be used in studies on the association between chronic exposure to air pollution and COVID-19 pathology, as well as for investigations on the role of air pollution on the health of the Italian population.

Supervised Spatial Regionalization using the Karhunen-Loève Expansion and Minimum Spanning Trees

The article presents a methodology for supervised regionalization of data on a spatial domain. Defining a spatial process at multiple scales leads to the famous ecological fallacy problem. Here, we use the ecological fallacy as the basis for a minimization criterion to obtain the intended regions. The Karhunen-Loève Expansion of the spatial process maintains the relationship between the realizations from multiple resolutions. Specifically, we use the Karhunen-Loève Expansion to define the regionalization error so that the ecological fallacy is minimized. The contiguous regionalization is done using the minimum spanning tree formed from the spatial locations and the data. Then, regionalization becomes similar to pruning edges from the minimum spanning tree. The methodology is demonstrated using simulated and real data examples.

Equity Implications of Robo-Taxis on Job Accessibility: Avoiding the Ecological Fallacy with Agent-Based Models

Equity Implications of Robo-Taxis on Job Accessibility: Avoiding the Ecological Fallacy with Agent-Based Models

Limits and potential implementation of the present pN classification for oral squamous cell carcinoma.

The purpose of this review is to highlight the most important changes in the eighth TNM classification system for oral squamous cell carcinomas compared with the seventh edition with focus on lymph node staging (pN). Nodal involvement is crucial when addressing prediction of survival, and staging must mirror the disease extension. pN classification will be evaluated with respect to lymph node yield (LNY), lymph node density (LND), and a recently proposed classification: pN-N+ reflecting positive regional lymph nodes (metastatic burden) and extra nodal extension. TNM8 was introduced in 2018, and the most noteworthy changes were depth of invasion (DOI) and extranodal extension (ENE). Recent studies indicate, that TNM8-related pN is not superior to TNM7 with respect to predicting survival. LNY and LND are biased with ecological interference fallacy, and currently not recommended in future iterations of TNM. In contrast, the pN-N+ classification has demonstrated improved survival prediction compared with TNM8. The recent findings support the inclusion of pN-N+, that is, metastatic burden and extranodal extension in future iterations of TNM.

Predictive Role of Population Density and Use of Public Transport for Major Outcomes of SARS-CoV-2 Infection in the Italian Population: An Ecological Study

Background: This study aimed at assessing how population density (PD), aging index (AI), use of public transport (URPT), and PM 10 concentration (PI) modulated the trajectory of the main COVID-19 pandemic outcomes in Italy, also in the recrudescence phase of the epidemic. Study design: Ecological study. Methods: For each region, we recovered data about cases, deaths, and case fatality rate (CFR) recorded since both the beginning of the epidemic and September 1, 2020. Data about total hospitalizations were included as well. Results: PD correlated with, and was the best predictor of, total and partial cases, total and partial deaths, and total hospitalizations. Moreover, URPT correlated with, and was the best predictor of, total CFR. Besides, PI correlated significantly with total and partial cases, total and partial deaths, and total hospitalizations. Conclusions: PD explains COVID-19 morbidity, mortality, and severity while URPT is the best predictor of disease lethality. These findings should be interpreted with caution due to the ecological fallacy.

Social distancing and influenza mortality in 1918 did not increase suicide rates in the United States

Recent research has suggested that the social distancing mandates introduced in the United States during the main waves of the 1918–20 influenza pandemic caused an increase in suicide rates. However, that finding relies on poor-quality, temporally mismatched data and has signs of omitted variable bias. Similarly, a long-standing finding that American suicide rates in 1918–20 were also boosted by the influenza mortality of the time has gone unquestioned in the literature, despite the original research admitting its risk of ecological fallacy. Using higher-powered mortality data, I cast doubt on both findings by analyzing the experiences of the pandemic in 43 of the largest American cities of the time. In line with some populations’ experiences of COVID-19, I report tentative evidence that social distancing mandates during the 1918–20 pandemic may have been associated with decreased suicide rates. Larger, cross-national investigations of the effects of historical pandemics and social distancing mandates on mental health and suicide are needed.

The relationship between urban refuse with fecundity and nestlings' success of a generalist seabird in the Río de la Plata Estuary - Uruguay

Resource acquisition and allocation impacts individual fitness. Using pellet analysis of breeding adults and stable isotopes of carbon and nitrogen of down feathers of Kelp Gull (Larus dominicanus) nestlings, we evaluated the relationship between urban refuse (beef and chicken) and natural food (fish) consumption of individual females during the pre-incubation period, with fecundity and young nesting's success in the Río de la Plata Estuary in Uruguay. Assimilated urban refuse positively correlated with egg weight and negatively with young nestling's success. This suggests a possible impact of urban refuse foraged by females during the pre-incubation period on their immediate fecundity (positively) and young nestling's survival (negatively). Differences between studies at the individual and colony levels are also discussed in light of an “ecological fallacy” of interpretation and we thus argue for the need of additional research to evaluate this relationship further, considering potential confounding factors.

Trends in depression incidence in China, 1990–2019

BackgroundDepression is one of the most widespread health burdens for the general population in China. This study aims to assess the long-term trends of depression incidence in China between 1990 and 2019. MethodsThe incidence data were drawn from the Global Burden of Disease Study 2019, and an age–period–cohort model was used in the analysis. ResultsThe net drift was -0.66% (95% CI: -0.79% to -0.53%) per year for both genders. For males, the local drift was lower than 0 (P<0.05) in those aged 10-54 years, and higher than 0 (P<0.05) in those aged 60–69 years. For females, the local drift was lower than 0 (P<0.05) in those aged 10–49 years and higher than 0 (P<0.05) in those aged 55-84 years. Females had a higher risk of depression incidence than males. Compared with the 1990-1994 period, the relative risk (RR) of depression incidence in 2015–2019 decreased by 12.2% in males and 12.3% in females, and compared to the 1903–1907 birth cohort, the cohort RRs in the 2008–2012 birth cohort decreased by 42.1% in males and 34.5% in females. Period and cohort RRs all showed an increased tendency in recent periods and birth cohorts. LimitationsThese data are macrolevel estimates at the national level, may have ecological fallacies. ConclusionsAlthough the age-standard incidence of depression has declined in China as a whole in the last three decades, the incidence of depression among older individuals has increased. More efforts are needed to promote the mental health of elderly individuals in China.

The Impact of Socioeconomic Factors on State Suicide Rates: Revisited

Abstract This paper revisits the purported impact of socioeconomic and social environment factors on annual, U.S. state-level suicide rates. Special attention is paid to the right-hand-side linking covariates directly to Durkheim's (1897/1951) significant contributions to established ecological suicide research. Results from a Haus-man-Taylor panel specification lend little support to Durkheim's social integra-tion/regulation hypothesis that aggregate social forces matter in explaining varia-tions in regional suicide rates. Data from 1990-2019 and the advanced empirical method support the mounting sentiment of an abiding ecological fallacy plaguing suicidology. JEL classification numbers: C51, R11, I31. Keywords: Hausman-Taylor, Suicide rates, Socioeconomic factors.

Wildfire smoke as a driver of covid-19 fatality rates? A synthetic control analysis of San Francisco Bay Area Counties

BACKGROUND AND AIM: The 2020 wildfire season broke records for the most active fire year on the West Coast, resulting in the worst air quality observed in decades. Concurrently, the public health threat of COVID-19 has caused approximately 3.6 million cases and 59,000 deaths in California to date. Due to the association between air pollution and respiratory disease outcomes, wildfire-specific particulate matter exposure is a hypothesized driver of COVID-19 severity. While several studies have considered this association, methodological challenges exist to study this question due to the ecological fallacy and threat of unmeasured confounders in this proposed etiological question. Additionally, no study has considered wildfire smoke as a natural experiment. This study proposes synthetic control methods (SCM) to study the impact of wildfire smoke on COVID-19 mortality rates, considering smoke exposure as a natural experiment. The San Francisco Bay Area (SFBA)—a densely populated region with record-high smoke exposure—is ideal to study this question. METHODS: The CDC COVID data tracker was used to compute weekly case-fatality ratios for each county in the United States from May 11th to November 10th, 2020. The NOAA hazard mapping Smoke Product was used to identify county-level wildfire smoke exposure; counties with area covered by heavy smoke surpassing 70% on average in a given week were considered exposed. Unexposed counties were considered as potential controls. Counties were considered eligible if more than 100 COVID-19 deaths were reported during the study period. RESULTS:Exposure to heavy smoke started between the weeks starting August 26th to September 7th in six eligible counties in which 1,307 COVID-19 deaths occurred within the study period. Results indicate some counties observed an increase in case fatality ratios, while other counties didn’t exhibit change. CONCLUSIONS:Results can be used to understand environmental drivers of COVID-19 mortality, and protect vulnerable populations from concurrent and potentially concomitant public health threats. KEYWORDS: Covid-19, wildfire smoke, synthetic control method

Maternal stress and sex ratio at birth in Sweden over two and a half centuries: a retest of the Trivers-Willard hypothesis.

STUDY QUESTIONIs there a negative relationship, as predicted in the Trivers–Willard hypothesis (TWH), between the intensity of maternal stress and sex ratio at birth (SRB)?SUMMARY ANSWERUsing a comprehensive data set with multiple indicators of maternal stress, most measures of stress show no statistically significant association with SRB over a period spanning 243 years, indicating no support for the TWH.WHAT IS KNOWN ALREADYEvolutionary biologists have proposed a widely discussed hypothesis that women in poor and stressful conditions during pregnancy are more likely to give birth to girls, and exposure to stressful events may therefore lead to a reduction in sex (male-to-female) ratio at birth. The empirical evidence so far is mixed.STUDY DESIGN, SIZE, DURATIONAnnual time series data, spanning 243 years between 1749 and 1991 for Sweden at the national level, were drawn from multiple sources. The outcome is defined as the percentage of male births relative to all births in Sweden in a given year. The covariates include a set of economic and climatic variables as proxies for maternal stress.PARTICIPANTS/MATERIALS, SETTING, METHODSWe conduct a series of ARIMA (autoregressive integrated moving average) models to examine the relationship between maternal stress and SRB during three periods: 1749–1991, 1749–1861 and 1862–1991.MAIN RESULTS AND THE ROLE OF CHANCEIn 1749–1991, economic proxies for maternal stress showed no statistically significant association with SRB. In 1749–1861, two indicators were significantly associated with SRB, but the coefficients were opposite in direction to the TWH. In 1862–1991, five out of six covariates showed no significant association with SRB. An additional analysis found no significant correlation between sex ratio of stillbirths and all covariates in 1862–1991. Our results are incompatible with the TWH and suggest that previous findings in support of the TWH are not robust.LIMITATIONS, REASONS FOR CAUTIONThis study provides population-level evidence that may not necessarily reflect the nature of all individuals due to the ecological fallacy. The time series analysed in this study are annual data, and we cannot examine the potential seasonality due to the lack of disaggregated monthly data. Our findings may not be generalised to the contexts of extreme maternal stress conditions such as famine and war.WIDER IMPLICATIONS OF THE FINDINGSThe results from existing studies in this topic may be speculative, and additional research with more comprehensive design, data and covariates is needed to reconsider the robustness of previous findings.STUDY FUNDING/COMPETING INTEREST(S)The author receives no external funding and has no conflict of interest to declare.TRIAL REGISTRATION NUMBERN/A.

Residential radon exposure and leukemia: A meta-analysis and dose-response meta-analyses for ecological, case-control, and cohort studies

IntroductionIn this study, the authors conducted a comprehensive systematic review and meta-analysis (including a dose-response meta-analysis) for a possible causal association between residential radon exposure and leukemia. All 3 types of study design, including ecological, case-control, and cohort studies, were included in this study. In particular, different measurement units of radon exposure among studies were dealt with and analyzed thoroughly. MethodsA medical librarian searched MEDLINE (PubMed), EMBASE, and the Cochrane Library (from January 01, 1970 to November 05, 2020). For ecological studies, a conventional meta-analysis and subgroup analyses with meta-ANOVA analyses were conducted. For case-control and cohort studies, a two-stage dose-response meta-analysis was conducted. ResultsA total of 8 ecological, 9 case-control, and 15 ecological-cohort studies were analyzed. For ecological studies, the pooled correlation coefficient was 0.48 (95% CI 0.41–0.54). In the meta-analysis of variance (ANOVA) analyses, the age group (childhood vs. adult) showed a statistically significant result (Q = 7.93 and p = 0.019) with the pooled correlation coefficient for childhood, adult, and all age group of 0.67 (95% CI 0.53–0.77), 0.46 (95% CI 0.05–0.74), and 0.44 (95% CI 0.36–0.51), respectively. For case-control studies, the dose-response meta-analysis showed the pooled OR increase of 1.0308 (95% CI 1.0050–1.0573) for each 100 Bq/m3 increase of radon dose. The pooled OR increase was 1.0361 (95% CI 1.0014–1.0720) for each 100 Bq/m3 increase of radon dose for lymphoid leukemia subgroup and 1.0309 (95% CI 1.0050–1.0575) for each 100 Bq/m3 increase of radon dose for childhood leukemia subgroup. Because of the inclusion of ecological studies with larger exposure assessment units, the pooled RR from ecological-cohort studies should be interpreted conservatively (a tendency towards a higher risk estimate). The overall pooled RR increase for each 100 Bq/m3 increase of radon dose was 1.1221 (95% CI 1.0184–1.2363). The pooled RR increase was 1.2257 (95% CI 1.0034–1.4972) for each 100 Bq/m3 increase of radon dose for the myeloid leukemia subgroup and 1.2503 (95% CI 1.0233–1.5276) for each 100 Bq/m3 increase of radon dose for adult leukemia subgroup. DiscussionA number of epidemiologic concepts, including the issue of sample size justification, the possibility of differential participation selection bias for case-control studies, the possibility of random and systematic errors in radon measurement, ecological fallacy for ecological studies, were discussed. The effect of age group, socioeconomic status, and gamma radiation exposure was also discussed. Future more accurate and conclusive large-scale case-control and cohort studies are needed.

Most published meta-regression analyses based on aggregate data suffer from methodological pitfalls: a meta-epidemiological study

BackgroundDue to clinical and methodological diversity, clinical studies included in meta-analyses often differ in ways that lead to differences in treatment effects across studies. Meta-regression analysis is generally recommended to explore associations between study-level characteristics and treatment effect, however, three key pitfalls of meta-regression may lead to invalid conclusions. Our aims were to determine the frequency of these three pitfalls of meta-regression analyses, examine characteristics associated with the occurrence of these pitfalls, and explore changes between 2002 and 2012.MethodsA meta-epidemiological study of studies including aggregate data meta-regression analysis in the years 2002 and 2012. We assessed the prevalence of meta-regression analyses with at least 1 of 3 pitfalls: ecological fallacy, overfitting, and inappropriate methods to regress treatment effects against the risk of the analysed outcome. We used logistic regression to investigate study characteristics associated with pitfalls and examined differences between 2002 and 2012.ResultsOur search yielded 580 studies with meta-analyses, of which 81 included meta-regression analyses with aggregated data. 57 meta-regression analyses were found to contain at least one pitfall (70%): 53 were susceptible to ecological fallacy (65%), 14 had a risk of overfitting (17%), and 5 inappropriately regressed treatment effects against the risk of the analysed outcome (6%). We found no difference in the prevalence of meta-regression analyses with methodological pitfalls between 2002 and 2012, nor any study-level characteristic that was clearly associated with the occurrence of any of the pitfalls.ConclusionThe majority of meta-regression analyses based on aggregate data contain methodological pitfalls that may result in misleading findings.

Hofstede's individual-level indulgence dimension: Scale development and validation

In spite of the important role of culture in consumer behavior, there has been no study of the effects of indulgence vs. restraint, the sixth dimension of Hofstede's cultural framework, at the individual level. This is mainly because there has been no scale to measure indulgence vs. restraint at the individual level. Given that ascribing national-level cultural scores to individuals may cause an ecological fallacy, the four studies of this research conceptualize and develop a valid and reliable measure for individual-level indulgence. In study 1 (n= 290), using PCA, the items were purified to the final six items. The result of an EFA in study 2 (n=230) suggested a unidimensional six-item structure for the scale. This structure was confirmed by a CFA in study 3 (n=257). Study 3 established the reliability, convergent, discriminant, and criterion validities of the scale. In study 4 (n=219), through developing and testing the nomological network, the nomological validity of the scale was supported by showing that the impacts of individual-level indulgence on hedonic attitudes and repurchase intentions are mediated by positive post-purchase emotions and hedonic attitudes, respectively.

On Simpson’s paradox: To remain or not to remain a population-based science

In this article, I critically reflect on J. F. Arocha’s (2021) contention that psychologists need to use methods and tools that are suitable for data analysis at the individual level. First, I discuss the beleaguered nature of the philosophical underpinnings of the standard practices in psychological research. Of the five assumptions he presented, the aggregate assumption results in Simpson’s paradox, a form of ecological fallacy. While the other assumptions need urgent attention, the proposals Arocha makes for addressing the aggregate assumption are still unsettled in many ways. I show that while perceptual control theory informed by the Aristotelian concept of final cause or telos allows for embracing variability as a psychological fact of human behaviour, one cannot say the same for his recommendation for the use of observation-oriented modelling (OOM) to address the aggregate assumption or to circumvent Simpson’s paradox.

Evaluations of science are robustly biased by identity concerns

People are known to evaluate science based on whether it (dis)affirms their collective identities. We examined whether personal identity concerns also bias evaluation processes by manipulating the degree to which summaries of ostensible scientific research about an unfamiliar topic were or inconsistent with how participants thought about themselves. In three preregistered experiments ( N = 644) conducted across two continents, participants were more likely to believe the science when its conclusions aligned with prior understanding of their self, effects that were mediated through positive emotional reactions. Two of the experiments also tested a de-biasing intervention: prior to evaluating science, participants received a brief tutorial on the ecological fallacy (of which, self-related biases represent a special case). The tutorial did not mitigate identity-biased evaluations. This experimental evidence raises questions about whether it is possible to engage global citizens more fully in science consumption while not further triggering identity-based biasing processes.