Statistical Significance Level
(Redirected from Size of Test)
Jump to navigation
Jump to search
A Statistical Significance Level is an ordinal classification that categorizes the strength of statistical evidence against a null hypothesis based on comparing p-values to significance level thresholds.
- AKA: Significance Classification, Statistical Evidence Level, P-Value Category, Significance Category.
- Context:
- It can typically classify hypothesis test results into discrete evidence categorys using standardized thresholds.
- It can often use conventional thresholds of 0.05, 0.01, and 0.001 to define significance boundarys.
- It can be expressed through descriptive labels like "not significant," "significant," or "highly significant."
- It can range from being a Non-Significant Statistical Significance Level to being an Extremely Significant Statistical Significance Level, depending on its evidence strength.
- It can be produced by a statistical significance measure when evaluating test statistics.
- It can determine whether to produce a statistically significant result or non-significant result.
- It can influence null hypothesis rejection decisions in statistical hypothesis testing tasks.
- It can be reported with confidence scores and effect size measures for complete interpretation.
- It can be affected by sample size, where large samples may yield significant levels for trivial effects.
- It can require multiple testing adjustment to maintain family-wise error rate control.
- It can vary across scientific disciplines in interpretation conventions.
- ...
- Example(s):
- Standard Statistical Significance Levels, such as:
- Non-Significant Statistical Significance Level (p > 0.05) indicating insufficient evidence.
- Marginally Significant Statistical Significance Level (0.05 < p ≤ 0.10) suggesting borderline evidence.
- Significant Statistical Significance Level (p ≤ 0.05) showing sufficient evidence.
- Highly Significant Statistical Significance Level (p ≤ 0.01) demonstrating strong evidence.
- Very Highly Significant Statistical Significance Level (p ≤ 0.001) providing very strong evidence.
- Research Context Statistical Significance Levels, such as:
- Specialized Statistical Significance Levels, such as:
- Genome-Wide Statistical Significance Level (p < 5×10⁻⁸) for genetic associations.
- Discovery Statistical Significance Level (5σ, p < 3×10⁻⁷) for physics discoverys.
- Suggestive Statistical Significance Level (p < 1×10⁻⁵) for preliminary findings.
- Multiple Testing Statistical Significance Levels, such as:
- ...
- Standard Statistical Significance Levels, such as:
- Counter-Example(s):
- Significance Level, which is the probability threshold rather than the classification result.
- P-Value, which is the continuous probability rather than the ordinal category.
- Test Statistic, which is the calculated value rather than the significance classification.
- Effect Size, which measures practical significance rather than statistical significance.
- Confidence Interval, which provides range estimates rather than categorical classifications.
- Bayes Factor, which quantifies evidence ratio rather than threshold-based categories.
- See: Statistical Significance Measure, Significance Level, Statistically Significant Result, Statistical Hypothesis Testing Task, Null Hypothesis, False Positive Error Rate, Confidence Score, Sample Size Determination Task, Statistical Power Measure.