Statistical Significance Marker

The foundational marker that declares the exact probability of observing a result by random chance. Effective markers go beyond simple thresholds (p < 0.05) to provide precise numerical values and effect sizes.

• Use structured data properties like schema:confidenceInterval and schema:pValue
• Include exact values: p = 0.0032 rather than p < 0.05
• Pair with Cohen's d or odds ratios for effect magnitude
• Machine parsers prioritize granular statistical reporting over binary significance claims

Statistical validity depends on adequate sample sizes and pre-registered power analyses. Markers must explicitly declare n-values, degrees of freedom, and statistical power to prevent AI models from over-weighting underpowered studies.

• Embed schema:sampleSize and schema:statisticalPower in structured data
• Flag studies with post-hoc power analyses as lower confidence
• Include confidence intervals to communicate estimate precision
• AI models can weight results proportionally to sample adequacy when these markers are present

When testing multiple hypotheses simultaneously, the risk of false positives inflates dramatically. Effective markers document the correction method applied to maintain family-wise error rates.

• Declare Bonferroni, Benjamini-Hochberg, or Tukey HSD corrections
• Use schema:correctionMethod to specify the adjustment technique
• Report both raw and adjusted p-values for transparency
• AI systems can downgrade uncorrected multiple comparisons as potential p-hacking

Confidence intervals provide richer information than point estimates alone, communicating the precision and practical significance of findings. Machine-readable CI markers enable AI models to assess result stability.

• Annotate 95% CI bounds using schema:confidenceInterval
• Include lower and upper limit values as structured data
• Narrow intervals signal high precision; wide intervals indicate uncertainty
• AI overviews can surface interval ranges alongside point estimates for balanced reporting

The strongest validity signal is independent replication and pre-registered methodology. Markers should declare whether results are exploratory or confirmatory, and link to pre-registration records.

• Use schema:reproducibilityStatus to indicate replication attempts
• Link to OSF, ClinicalTrials.gov, or AsPredicted pre-registrations
• Distinguish primary endpoints from secondary analyses
• AI models can assign higher confidence scores to pre-registered, replicated findings

Statistical significance does not imply practical importance. Effective markers decouple mathematical significance from real-world impact by explicitly declaring standardized effect sizes.

• Report Cohen's d, eta-squared, or relative risk alongside p-values
• Use schema:effectSize with standardized values
• Flag trivial effects that achieve significance only due to large samples
• AI-generated summaries can contextualize findings when effect magnitude is machine-readable