🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
ascertainment_bias
Ascertainment bias occurs when the method of identifying study participants systematically distorts the sample composition. The way cases are detected, referred, or selected determines who ends up in the study, and if this process favors certain characteristics, the results will not reflect the true population. This is especially common in clinical and genetic studies.
A genetic study recruits families through hospitals, finding a strong association between a gene variant and a rare disease. However, families with multiple affected members are more likely to seek medical attention, inflating the apparent genetic risk.
A public health department tracks opioid overdose rates using emergency room records and concludes that overdoses are concentrated in urban areas. Rural overdoses that result in death before an ambulance arrives or are misclassified as accidents never enter the hospital system, making rural areas appear far safer than they are.
An online survey about social media addiction is promoted through Facebook and Instagram, recruiting participants who are already active users. People who have quit social media entirely due to addiction problems are unreachable through this channel, so the most severe cases are systematically missed.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Were participants identified or recruited through a method that could favor certain types of cases?
Type: binaryCould the detection or referral process systematically overrepresent certain conditions or demographics?
Type: binaryDoes the sample source (hospital, clinic, database) introduce a systematic skew?
Type: binaryAre findings generalized beyond the specific population from which cases were ascertained?
Type: binaryAscertainment bias occurs when the method of identifying study participants systematically distorts the sample composition. The way cases are detected, referred, or selected determines who ends up in the study, and if this process favors certain characteristics, the results will not reflect the true population. This is especially common in clinical and genetic studies.
The recruitment pathway acts as an invisible filter. Researchers often assume their sample is representative without examining how the identification process itself shaped the group. Cases that are more visible, more severe, or more connected to the medical system are overrepresented.
Use population-based sampling rather than clinic-based recruitment. Document the ascertainment process transparently. Conduct sensitivity analyses to estimate how different recruitment methods might alter results.
Cancer registries that rely on hospital records overrepresent aggressive cancers (which require treatment) and underrepresent slow-growing ones (which may never be diagnosed). This distorts estimates of cancer incidence, survival, and treatment effectiveness.
Systematic exclusion of certain participants from a study distorts results.
Diagnostic test accuracy varies when evaluated across different disease severity levels.
A spurious correlation appears between two independent variables when the sample is conditioned on a common effect (collider). For example, among hospitalized patients, two unrelated diseases may appear negatively correlated because admission is the collider.
Systematic differences in how outcomes are identified between comparison groups.
Participants who choose to join a study differ systematically from those who do not.
Use these tools to detect, analyze, or train this aspect.