Start typing to search across the full knowledge system.
Reasoning error · causal reasoning
A correlation/causation error treats an observed association as sufficient evidence that one variable caused the other, without resolving alternative causal directions, confounding or selection effects.
Explore connections →Correlation can be part of causal evidence. The error is not “correlation is useless”; it is treating association as if all competing causal structures had already been ruled out.
Good causal inference asks what else could generate the same pattern and what design or additional evidence can distinguish those possibilities.
Research basis
Review of causal illusions and contingency judgments, including conditions that make unrelated events feel causally connected.
Examines how readers infer causal conclusions from correlational study descriptions and how explicit alternative causal models can improve reasoning.
Investigation ↔ Learn
High-frequency voltage transients on electrical wiring are real and measurable, but the claim that they are an established major cause of chronic disease is not supported by the current human evidence. Reviews have found inconsistent exposure definitions, weak or nonblinded designs and limited replication. The electrical phenomenon should be distinguished from the much stronger health-causation claim.
Eugene Mallove was murdered, but a January 2025 appellate decision ordered vacatur of one defendant’s convictions and a new trial after a due-process violation. Neither this case nor a defined comparison rate establishes widespread motive-linked murder of disruptive-energy inventors.
Pyramidal forms appear in multiple cultures, but their dates, materials, functions, construction histories and local architectural sequences differ substantially. Egypt and Teotihuacan, for example, developed within distinct regional traditions separated by thousands of years. Shared broad geometry is not enough to establish one global designer.
A latitude line at 29.9792458° N does pass through the Great Pyramid, echoing the digits of 299,792,458 metres per second. But the pyramid's surveyed center is at a different latitude, and the numerical match depends on modern decimal degrees and the SI value of c—whose exact numerical value is fixed by the modern SI definition of the metre. No independent evidence shows Fourth Dynasty builders intended this encoding.
Alfred Watkins' original ley idea concerned alignments and possible old trackways, not a measured planetary energy network. Archaeological sites can form striking alignments by chance when many points and many possible lines are available. No reproducible physical field corresponding to a global ley grid has been established.
Churches often occupy old or symbolically important sites, and their orientations can reflect astronomy, terrain and earlier structures. None of that demonstrates a measurable Earth-energy field or an architecture that harvests human energy. The stronger physical claim remains unsupported; versions that never define measurable energy become unfalsifiable.
Human activity can induce earthquakes when it changes stress or fluid pressure in the crust, for example through injection, withdrawal, mining or reservoirs. The High-frequency Active Auroral Research Program (HAARP) transmits high-frequency radio waves into the ionosphere, and USGS identifies no plausible mechanism by which that process would generate major tectonic earthquakes.
Dietary patterns marketed for particular blood types can be nutritious, but available research has not demonstrated benefits caused by an ABO-specific food match. The 1,455-person 2014 study was observational—not a diet intervention—and a 2021 analysis of 68 participants in a vegan-diet intervention found no meaningful ABO-associated response difference. Neither is a randomized direct test of all matched versus mismatched ABO diets.
Aluminum salts can interact with biology and aluminum has been measured in breast tissue, so the proposed mechanism is not meaningless. Epidemiologic evidence, however, is limited and conflicting: a large 2002 case-control study found no increased risk, while a smaller 2017 study reported an association in a very high-use subgroup and higher tissue aluminum in cases. A 2023 systematic review found inconsistent evidence, while a 2024 meta-analysis of seven case-control studies of antiperspirants or deodorants found no overall association (pooled odds ratio 0.96, 95% confidence interval 0.78–1.17; substantial heterogeneity). The pooled exposure was not restricted to aluminum salts and the studies were retrospective. A causal increase in breast-cancer risk is therefore not established, but neither the pooled null result nor the high-use subgroup signal establishes certainty.
Excess mortality persisted after vaccine rollout and rare serious vaccine adverse events are real, but large linked individual-level and self-controlled studies do not show the population mortality increase predicted by a claim that vaccination caused a substantial share of excess deaths.
Natural-catastrophe explanations of Exodus have a long modern history, and the historical core of Exodus remains debated. None of that establishes a close planetary encounter. The astronomical component lacks the independent orbital and geological evidence it would require.
Parasitic infections are common in many endemic regions and can be asymptomatic or produce nonspecific gastrointestinal, skin or systemic symptoms. But risk is strongly shaped by geography, sanitation, travel, food, water and exposure, and major public-health guidance does not support assuming that most people with ordinary nonspecific symptoms have parasites. Diagnosis is based on exposure context and appropriate stool, blood or other testing.
Naturalistic chains have been proposed for several plague motifs, but historical and medical reviews do not establish the biblical sequence as one documented outbreak, and no independent evidence ties it to a planetary or cosmic catastrophe.
Some comparative analyses find no clear incremental benefit from the most restrictive measures, while stronger local quasi-experiments find substantial transmission reductions. Because voluntary behavior, policy bundles and endogeneity are difficult to separate, the broad incremental effect remains uncertain.
Diet changes renal acid load and can substantially change urine pH, but blood pH in healthy people is tightly regulated by respiratory and renal physiology. Human evidence does not support using alkaline diets or alkaline water to prevent or treat cancer by alkalizing the blood.
Spinal manipulation can be evaluated for musculoskeletal conditions without accepting the broader vitalistic subluxation theory. Reviews and randomized trials have not established chiropractic subluxations as a major cause of systemic disease or shown manipulation to outperform sham for a range of non-musculoskeletal disorders.
Cold exposure causes real stress, inflammatory and circulatory responses, but evidence that it broadly prevents infections is limited. A large cold-shower trial reduced sickness absence without reducing illness days, and a 2025 systematic review found no clear pooled immediate immune-function benefit. Immune biomarkers, feeling resilient and actually having fewer infections are different outcomes.
People reporting electromagnetic hypersensitivity can experience real and sometimes disabling symptoms. The causal question is separate: blinded provocation studies have generally not found that participants can detect actual RF exposure better than chance or that below-limit exposure reliably produces more symptoms than sham exposure. The strongest recent systematic review addresses mainly acute exposure and does not settle every possible long-term hypothesis.
A small 2001 study of 11 aged cadavers found substantial fluoride in pineal tissue and a positive correlation between pineal fluoride and calcium, so claims that fluoride can occur in calcified pineal tissue have a real empirical basis. The study did not show that fluoride caused pineal calcification, impaired pineal function or reduced cognition. Separately, current systematic reviews report an association between higher fluoride exposure and lower childhood IQ, while evidence at drinking-water concentrations relevant to 0.7 mg/L community fluoridation remains limited and disputed. Neither line of evidence establishes a pineal-mediated cognitive effect or an intentional population-control program.
Soft green lumps passed after oil-and-citrus liver flushes are often presented as gallstones that were already inside the gallbladder. Direct chemical examination and ingredient-reconstruction evidence show that characteristic lumps can form from flush ingredients rather than represent pre-existing biliary stones. Ultrasound-measured stone clearance and claims about unnamed toxins are different endpoints.
Stress and psychological state can affect health, symptoms and behavior, but German New Medicine's specific conflict-to-organ laws are not supported by modern cancer biology or validated clinical evidence. Treating its universal mapping as a replacement for diagnosis and evidence-based care can therefore misattribute disease and delay treatment.
In a normally oxygenated person, hemoglobin is already close to saturated. Hyperventilation mainly lowers carbon dioxide, which can constrict cerebral blood vessels and shift oxygen binding rather than simply increasing tissue oxygen delivery. Breathwork may have other psychological or autonomic effects, but the 'more breathing equals more tissue oxygen' mechanism is misleading.
Smart-meter emissions must be assessed for two different endpoints: cancer incidence and recurring headaches in nearby residents. Original meter studies measure radiofrequency (RF) exposure and variable transmission duty cycles, not diagnosed cancer or headache causation. Neither clinical claim has been established by the reviewed record; evidence quality for typical emissions is more direct than for either health outcome.
Research into natural conception and assisted reproduction does not establish that vaccination commonly causes persistent infertility. Some inactivated-COVID-19-vaccine frozen-transfer cohorts report conflicting live-birth associations, including a concerning 2025 result requiring replication. Temporary menstrual changes and selected embryo-transfer outcomes are not measures of lasting population infertility.
Two different claims need different tests: whether Bluetooth headsets produce more external radiofrequency (RF) power density than an active phone at the ear, and whether earbud RF causes brain tumours. A historical 2019 comparison measured lower headset external power density in tested configurations, not brain specific absorption rate (SAR); no causal brain-tumour result was established. A 2024 thyroid-nodule survey does not measure brain cancer.
Sugar-rich diets can matter for cancer risk, especially through excess energy intake and body fatness, and newer cohort syntheses report associations between sugar-sweetened beverages and some cancers. But that does not establish a universal direct carcinogenic effect of dietary sugar independent of obesity and metabolic pathways. The broad claim is misleading.
The vaccine-autism claim has been tested in large cohort studies, case-control studies, meta-analyses and evidence reviews. A 2019 Danish cohort of 657,461 children found no increased autism risk after measles–mumps–rubella (MMR) vaccination, including among children with elevated baseline risk, and a 2014 meta-analysis summing more than 1.2 million cohort study enrollments found no increased autism risk associated with vaccination, MMR, thimerosal or mercury exposure; overlap among contributing national cohorts means this is not a verified count of unique children. The 1998 paper that catalyzed the MMR claim was a 12-child case series, did not establish causation and was later retracted.
Vaccines can cause adverse events, including some rare serious events, and vaccine safety evidence has real limitations that require continued surveillance. Those facts do not establish an intentional program to create chronic disease. A 2013 National Academies review found no evidence that the then-current U.S. childhood schedule was unsafe while also identifying research gaps; post-licensure systems such as VAERS and VSD are designed to detect and investigate safety signals. Neither that historical schedule review nor the surveillance systems prove every vaccine policy optimal, but no verified documentary, epidemiological or mechanistic evidence establishes the claimed design-for-chronic-illness scheme.