KnowraEcological fallacyLinked fromLinked fromThe 28 pages that link to Ecological fallacy, each with the reason it gives.All 28Related 18Compared with 10Causal inferenceRelated: Aggregate associations can suggest causal patterns that do not hold for individuals.Public opinionRelated: Group-level opinion patterns do not establish what any particular person believes.Selection biasCompared with: It arises from mistaking aggregate patterns for individual ones, not from selective inclusion.CorrelationCompared with: A correlation across groups need not describe associations among individuals within those groups.Simpson's paradoxCompared with: It also warns against cross-level inference, but does not require an association to reverse.Collective efficacyCompared with: Neighborhood-level associations do not establish that every resident shares the measured beliefs or behavior.Choropleth mapRelated: Area-level shading cannot establish that individuals share the mapped area's pattern.Individual differencesCompared with: Population statistics do not automatically describe any particular person.Spatial analysisRelated: Area-level patterns do not necessarily describe the people or objects within each area.Ecological validityCompared with: It shows that real-world data can still mislead when conclusions cross levels of analysis.Mortality rateRelated: A population mortality rate does not establish an individual’s probability of death.Thematic mapRelated: Area-level patterns on thematic maps do not necessarily describe individuals within those areas.Excess mortalityRelated: Population-level excess deaths do not identify which individuals or causes account for them.Modifiable areal unit problemRelated: Aggregation can make area-level patterns appear to describe individuals when they do not.Copycat crimeRelated: Population-level timing after coverage cannot by itself prove that particular offenders copied it.Gravity modelRelated: Aggregate place-to-place flows do not directly reveal individual choices.Social disorganization theoryCompared with: Neighborhood-level crime patterns do not establish that every resident shares the same traits or risks.Scale (geography)Related: Aggregating observations can conceal differences among individuals.Opinion pollRelated: Poll summaries about groups do not establish what each individual believes.Statistical populationCompared with: Population-level summaries do not automatically establish properties of individual members.Suicide contagionRelated: Population rate changes cannot establish which exposed individuals were affected.Isarithmic mapRelated: Interpolated map surfaces can invite unsupported conclusions about individuals from geographic patterns.Bradley effectRelated: Aggregate poll errors cannot by themselves establish that individual respondents concealed their preferences.Offender profilingRelated: Group-level crime patterns do not establish that a particular offender has the group’s average traits.Cinderella effectRelated: A group-level difference cannot establish that a particular stepparent poses a greater risk.Modifiable temporal unit problemCompared with: Both problems warn that conclusions depend on the level at which data are analyzed.Population Characteristics (demography)Compared with: Population summaries do not establish that every individual shares a group-level pattern.Statistical geographyRelated: Area summaries can conceal variation among people within each statistical unit.
KnowraEcological fallacyLinked fromLinked fromThe 28 pages that link to Ecological fallacy, each with the reason it gives.All 28Related 18Compared with 10Causal inferenceRelated: Aggregate associations can suggest causal patterns that do not hold for individuals.Public opinionRelated: Group-level opinion patterns do not establish what any particular person believes.Selection biasCompared with: It arises from mistaking aggregate patterns for individual ones, not from selective inclusion.CorrelationCompared with: A correlation across groups need not describe associations among individuals within those groups.Simpson's paradoxCompared with: It also warns against cross-level inference, but does not require an association to reverse.Collective efficacyCompared with: Neighborhood-level associations do not establish that every resident shares the measured beliefs or behavior.Choropleth mapRelated: Area-level shading cannot establish that individuals share the mapped area's pattern.Individual differencesCompared with: Population statistics do not automatically describe any particular person.Spatial analysisRelated: Area-level patterns do not necessarily describe the people or objects within each area.Ecological validityCompared with: It shows that real-world data can still mislead when conclusions cross levels of analysis.Mortality rateRelated: A population mortality rate does not establish an individual’s probability of death.Thematic mapRelated: Area-level patterns on thematic maps do not necessarily describe individuals within those areas.Excess mortalityRelated: Population-level excess deaths do not identify which individuals or causes account for them.Modifiable areal unit problemRelated: Aggregation can make area-level patterns appear to describe individuals when they do not.Copycat crimeRelated: Population-level timing after coverage cannot by itself prove that particular offenders copied it.Gravity modelRelated: Aggregate place-to-place flows do not directly reveal individual choices.Social disorganization theoryCompared with: Neighborhood-level crime patterns do not establish that every resident shares the same traits or risks.Scale (geography)Related: Aggregating observations can conceal differences among individuals.Opinion pollRelated: Poll summaries about groups do not establish what each individual believes.Statistical populationCompared with: Population-level summaries do not automatically establish properties of individual members.Suicide contagionRelated: Population rate changes cannot establish which exposed individuals were affected.Isarithmic mapRelated: Interpolated map surfaces can invite unsupported conclusions about individuals from geographic patterns.Bradley effectRelated: Aggregate poll errors cannot by themselves establish that individual respondents concealed their preferences.Offender profilingRelated: Group-level crime patterns do not establish that a particular offender has the group’s average traits.Cinderella effectRelated: A group-level difference cannot establish that a particular stepparent poses a greater risk.Modifiable temporal unit problemCompared with: Both problems warn that conclusions depend on the level at which data are analyzed.Population Characteristics (demography)Compared with: Population summaries do not establish that every individual shares a group-level pattern.Statistical geographyRelated: Area summaries can conceal variation among people within each statistical unit.