scientific-critical-thinking

Scientific Critical Thinking

Overview

Critical thinking is a systematic process for evaluating scientific rigor. Assess methodology, experimental design, statistical validity, biases, confounding, and evidence quality using GRADE and Cochrane ROB frameworks. Apply this skill for critical analysis of scientific claims.

When to Use This Skill

This skill should be used when:

• Evaluating research methodology and experimental design
• Assessing statistical validity and evidence quality
• Identifying biases and confounding in studies
• Reviewing scientific claims and conclusions
• Conducting systematic reviews or meta-analyses
• Applying GRADE or Cochrane risk of bias assessments
• Providing critical analysis of research papers

Visual Aids (Optional)

Only add figures when the user explicitly requests a diagram (for example, a GRADE flowchart, bias decision tree, or evidence-quality framework).

When figures help:

• Critical thinking framework diagrams
• Bias identification decision trees
• Evidence quality assessment flowcharts
• GRADE or risk-of-bias evaluation frameworks

How to create figures:

• Preferred: Use the scientific-schematics skill for AI-generated diagrams from a natural-language description
• Alternative: Build figures in your usual tools (draw.io, PowerPoint, matplotlib, etc.)

From the scientific-schematics skill directory, with OPENROUTER_API_KEY set:

python scripts/generate_schematic.py "GRADE evidence assessment flowchart with downgrade and upgrade factors" -o figures/grade_flowchart.png --doc-type report

Disclosure: AI schematic generation sends your prompt to OpenRouter (a third-party API). Do not include unpublished sensitive details unless that transmission is appropriate for your project.

Core Capabilities

1. Methodology Critique

Evaluate research methodology for rigor, validity, and potential flaws.

Apply when:

• Reviewing research papers
• Assessing experimental designs
• Evaluating study protocols
• Planning new research

Evaluation framework:

1. Study Design Assessment
2. Is the design appropriate for the research question?
3. Can the design support causal claims being made?
4. Are comparison groups appropriate and adequate?
5. Consider whether experimental, quasi-experimental, or observational design is justified

1. Validity Analysis
2. Internal validity: Can we trust the causal inference?
3. Check randomization quality
4. Evaluate confounding control
5. Assess selection bias
6. Review attrition/dropout patterns
7. External validity: Do results generalize?
8. Evaluate sample representativeness
9. Consider ecological validity of setting
10. Assess whether conditions match target application
11. Construct validity: Do measures capture intended constructs?
12. Review measurement validation
13. Check operational definitions
14. Assess whether measures are direct or proxy
15. Statistical conclusion validity: Are statistical inferences sound?
16. Verify adequate power/sample size
17. Check assumption compliance
18. Evaluate test appropriateness

1. Control and Blinding
2. Was randomization properly implemented (sequence generation, allocation concealment)?
3. Was blinding feasible and implemented (participants, providers, assessors)?
4. Are control conditions appropriate (placebo, active control, no treatment)?
5. Could performance or detection bias affect results?

1. Measurement Quality
2. Are instruments validated and reliable?
3. Are measures objective when possible, or subjective with acknowledged limitations?
4. Is outcome assessment standardized?
5. Are multiple measures used to triangulate findings?

Reference: See references/scientific_method.md for detailed principles and references/experimental_design.md for comprehensive design checklist.

2. Bias Detection

Identify and evaluate potential sources of bias that could distort findings.

Apply when:

• Reviewing published research
• Designing new studies
• Interpreting conflicting evidence
• Assessing research quality

Systematic bias review:

1. Cognitive Biases (Researcher)
2. Confirmation bias: Are only supporting findings highlighted?
3. HARKing: Were hypotheses stated a priori or formed after seeing results?
4. Publication bias: Are negative results missing from literature?
5. Cherry-picking: Is evidence selectively reported?
6. Check for preregistration and analysis plan transparency

1. Selection Biases
2. Sampling bias: Is sample representative of target population?
3. Volunteer bias: Do participants self-select in systematic ways?
4. Attrition bias: Is dropout differential between groups?
5. Survivorship bias: Are only "survivors" visible in sample?
6. Examine participant flow diagrams and compare baseline characteristics

1. Measurement Biases
2. Observer bias: Could expectations influence observations?
3. Recall bias: Are retrospective reports systematically inaccurate?
4. Social desirability: Are responses biased toward acceptability?
5. Instrument bias: Do measurement tools systematically err?
6. Evaluate blinding, validation, and measurement objectivity

1. Analysis Biases
2. P-hacking: Were multiple analyses conducted until significance emerged?
3. Outcome switching: Were non-significant outcomes replaced with significant ones?
4. Selective reporting: Are all planned analyses reported?
5. Subgroup fishing: Were subgroup analyses conducted without correction?
6. Check for study registration and compare to published outcomes

1. Confounding
2. What variables could affect both exposure and outcome?
3. Were confounders measured and controlled (statistically or by design)?
4. Could unmeasured confounding explain findings?
5. Are there plausible alternative explanations?

Reference: See references/common_biases.md for comprehensive bias taxonomy with detection and mitigation strategies.

3. Statistical Analysis Evaluation

Critically assess statistical methods, interpretation, and reporting.

Apply when:

• Reviewing quantitative research
• Evaluating data-driven claims
• Assessing clinical trial results
• Reviewing meta-analyses

Statistical review checklist:

1. Sample Size and Power
2. Was a priori power analysis conducted?
3. Is sample adequate for detecting meaningful effects?
4. Is the study underpowered (common problem)?
5. Do significant results from small samples raise flags for inflated effect sizes?

1. Statistical Tests
2. Are tests appropriate for data type and distribution?
3. Were test assumptions checked and met?
4. Are parametric tests justified, or should non-parametric alternatives be used?
5. Is the analysis matched to study design (e.g., paired vs. independent)?

1. Multiple Comparisons
2. Were multiple hypotheses tested?
3. Was correction applied (Bonferroni, FDR, other)?
4. Are primary outcomes distinguished from secondary/exploratory?
5. Could findings be false positives from multiple testing?

1. P-Value Interpretation
2. Are p-values interpreted correctly (probability of data if null is true)?
3. Is non-significance incorrectly interpreted as "no effect"?
4. Is statistical significance conflated with practical importance?
5. Are exact p-values reported, or only "p < .05"?
6. Is there suspicious clustering just below .05?

1. Effect Sizes and Confidence Intervals
2. Are effect sizes reported alongside significance?
3. Are confidence intervals provided to show precision?
4. Is the effect size meaningful in practical terms?
5. Are standardized effect sizes interpreted with field-specific context?

1. Missing Data
2. How much data is missing?
3. Is missing data mechanism considered (MCAR,