[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M792DAPBR18S2W5BPFPF8RXZ":3,"skill-same:01M792DAPBR18S2W5BPFPF8RXZ":62},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":31,"ownerHtmlUrl":32,"ownerBlogUrl":33,"ownerDescription":34,"updatedAt":35,"install":36,"resources":49},"01M792DAPBR18S2W5BPFPF8RXZ","research-methods",null,"Design the study itself: question framing, method choice, sampling, controls and which statistics actually fit. Use when designing qualitative, quantitative, or experimental research methods.","poorvith-mp\u002Fskills-education","poorvith-mp","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F277478855?v=4","01M78Z5K0YH1Z3PX815PGQ96ND","Poorvith M P","01M792ATSH2NKESX1VP51ASDH0",3,"MIT","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fresearch-methods\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: research-methods\nlast_reviewed: 2026-09-06\ngroup: Research\ndescription: >-\n  Design the study itself: question framing, method choice, sampling, controls and which\n  statistics actually fit. Use when designing qualitative, quantitative, or experimental research\n  methods.\n---\n\n# research-methods\n\n## Core Philosophy\nResearch methodology is the epistemic foundation of science: it is the disciplined architecture that allows researchers to draw valid causal inferences while systematically neutralizing cognitive bias, confounding variables, and measurement noise. A study is only as strong as its methodological rigor. Choosing the wrong sampling strategy, failing to control for confounders, or applying parametric statistics to non-normal ordinal data invalidates every chart, table, and conclusion in the final manuscript.\n\n---\n\n## 4-Step Research Design Architecture\n\n### Step 1: Question Framing & Methodological Paradigms\n1. **Quantitative vs Qualitative vs Mixed Methods**:\n   - **Quantitative**: Tests specific causal hypotheses, measures relationships between variables, uses numerical data and inferential statistics (e.g. *RCTs, quasi-experiments, cohort studies*).\n   - **Qualitative**: Explores phenomena, unpacks mental models, understands social processes (e.g. *phenomenology, grounded theory, thematic interviews*).\n   - **Mixed Methods (Convergent Parallel \u002F Sequential)**: Combines statistical breadth with qualitative depth.\n2. **Variable Operationalization**:\n   - Define exact, observable metrics for abstract constructs:\n     - *Construct*: \"Developer Cognitive Load\".\n     - *Operationalization*: NASA-TLX score (0-100) + Pupil dilation metrics during debugging tasks.\n\n### Step 2: Sampling Strategy & Statistical Power ($1 - \\beta$)\n1. **Sampling Methodologies**:\n   - **Probability Sampling**: Simple random, stratified (preserves subgroup ratios), cluster sampling. Enables statistical generalizability.\n   - **Non-Probability Sampling**: Convenience, purposive, snowball. High risk of selection bias; strictly bounded generalizability.\n2. **Sample Size & Statistical Power Calculations**:\n   - Calculate sample size $N$ in advance based on desired effect size (Cohen's $d$), significance level ($\\alpha = 0.05$), and statistical power ($1 - \\beta = 0.80$):\n     - Underpowered studies ($N$ too small) produce high false-negative rates and inflated effect sizes when statistical significance is achieved by chance (Winner's Curse).\n\n### Step 3: Experimental Controls & Threats to Validity\n1. **Neutralizing Threats to Internal Validity**:\n   - *Selection Bias*: Mitigated by strict computer-generated **random assignment**.\n   - *History \u002F Maturation*: Mitigated by parallel **Control Groups** experiencing identical time intervals without the intervention.\n   - *Observer \u002F Hawthorne Effect*: Mitigated by **Double-Blind** protocols where neither participant nor researcher knows group allocation.\n2. **Threats to External Validity**:\n   - Ensure the sample demographic and testing environment reflect the real-world operational domain (ecological validity).\n\n### Step 4: Statistical Test Alignment (Parametric vs Non-Parametric)\n1. **The Decision Matrix for Statistical Tests**:\n   - Continuous DV + Normal Distribution + Independent Groups (2 groups) $\\to$ **Two-Sample Independent $t$-Test**.\n   - Continuous DV + Normal Distribution + Repeated Measures (Same group) $\\to$ **Paired $t$-Test**.\n   - Continuous DV + Normal Distribution + $\\ge 3$ Groups $\\to$ **One-Way ANOVA** (with Tukey HSD post-hoc).\n   - Ordinal DV (Likert scales) or Skewed Distribution (2 groups) $\\to$ **Mann-Whitney $U$ Test** (Wilcoxon Rank-Sum).\n   - Categorical IV + Categorical DV $\\to$ **Chi-Square ($\\chi^2$) Test of Independence**.\n   - Continuous DV + Multiple Continuous\u002FCategorical IVs $\\to$ **Multiple Linear Regression**.\n\n---\n\n## Deliverable Format: Methodological Study Blueprint (RCT Protocol)\n\n```markdown\n# Study Protocol: Evaluating Spaced Retrieval in Software Engineering Onboarding\n\n### 1. Research Hypothesis\nEngineers onboarded using a spaced-retrieval CLI interactive tool will demonstrate significantly higher code architecture comprehension ($p \u003C 0.01$, Cohen's $d \\ge 0.5$) at Day 30 compared to engineers onboarded using traditional monolithic wiki documentation.\n\n### 2. Experimental Design\n- **Type**: Randomized Controlled Trial (Between-Subjects).\n- **Independent Variable (IV)**: Onboarding Modality:\n  - *Treatment Group*: Spaced retrieval CLI tool (10-min active recall challenges on Days 1, 3, 7, 14, 28).\n  - *Control Group*: Standard company Confluence onboarding wiki (unconstrained self-study).\n- **Dependent Variable (DV)**: Architecture Comprehension Score (0-100 on a validated 20-item diagnostic scenario exam administered on Day 30).\n\n### 3. Sampling & Power Analysis\n- Power calculation ($\\alpha = 0.05, 1 - \\beta = 0.80$, expected Cohen's $d = 0.60$): Requires minimum $N = 45$ participants per group ($N_{\\text{total}} = 90$).\n- Recruitment: 96 newly hired engineers across 4 enterprise business units, stratified by years of prior software experience, then randomly assigned 1:1.\n\n### 4. Statistical Analysis Plan\n1. Test normality of Day 30 score distribution via Shapiro-Wilk test.\n2. If normal: Independent two-sample $t$-test; report mean difference and 95% Confidence Intervals.\n3. If non-normal: Mann-Whitney $U$ non-parametric test.\n4. Multiple regression controlling for years of experience and prior programming language familiarity.\n```\n\n---\n\n## Worked Example: Rescuing a Flawed Product A\u002FB Test Methodology\n\n- **Problem**: A growth team ran an A\u002FB test claiming a new checkout button increased conversions by 14% ($p = 0.04$). Two months later, total company revenue had declined.\n- **Methodological Audit**:\n  1. *Peeking Bias*: The growth team had checked the dashboard every morning and stopped the test the moment $p \u003C 0.05$ appeared (Day 4), inflating false-positive rates to over 30%.\n  2. *Novelty Effect*: Returning power users clicked the bright new button out of curiosity, not increased purchase intent.\n- **Methodological Remedy**: Implemented fixed-horizon testing: mandatory 14-day run time to capture full weekly seasonal cycles, with sample sizes locked in advance via power analysis.\n- **Result**: The 14% conversion lift vanished in the controlled re-test, preventing a disastrous permanent UI overhaul.\n\n---\n\n## Verification Checklist\n\n- [ ] Research paradigm matches the underlying epistemic question.\n- [ ] Variables clearly operationalized with validated measurement instruments.\n- [ ] Sample size determined prior to data collection via formal statistical power calculations.\n- [ ] Threats to internal validity (selection, maturation, confounders) controlled.\n- [ ] Statistical tests aligned with variable data types (parametric vs non-parametric).\n\n---\n\n## Anti-Patterns\n\n- **P-Hacking \u002F Data Dredging**: Running 40 different statistical combinations until one randomly shows $p \u003C 0.05$ and inventing a hypothesis after the fact (HARKing).\n- **Premature Test Peeking**: Halting an A\u002FB test or experiment early the moment the significance indicator flashes green.\n- **Convenience Bias Generalization**: Running an experiment exclusively on 30 undergraduate psychology students and claiming the findings apply to all human beings.\n","# research-methods\n\n## Core Philosophy\nResearch methodology is the epistemic foundation of science: it is the disciplined architecture that allows researchers to draw valid causal inferences while systematically neutralizing cognitive bias, confounding variables, and measurement noise. A study is only as strong as its methodological rigor. Choosing the wrong sampling strategy, failing to control for confounders, or applying parametric statistics to non-normal ordinal data invalidates every chart, table, and conclusion in the final manuscript.\n\n---\n\n## 4-Step Research Design Architecture\n\n### Step 1: Question Framing & Methodological Paradigms\n1. **Quantitative vs Qualitative vs Mixed Methods**:\n   - **Quantitative**: Tests specific causal hypotheses, measures relationships between variables, uses numerical data and inferential statistics (e.g. *RCTs, quasi-experiments, cohort studies*).\n   - **Qualitative**: Explores phenomena, unpacks mental models, understands social processes (e.g. *phenomenology, grounded theory, thematic interviews*).\n   - **Mixed Methods (Convergent Parallel \u002F Sequential)**: Combines statistical breadth with qualitative depth.\n2. **Variable Operationalization**:\n   - Define exact, observable metrics for abstract constructs:\n     - *Construct*: \"Developer Cognitive Load\".\n     - *Operationalization*: NASA-TLX score (0-100) + Pupil dilation metrics during debugging tasks.\n\n### Step 2: Sampling Strategy & Statistical Power ($1 - \\beta$)\n1. **Sampling Methodologies**:\n   - **Probability Sampling**: Simple random, stratified (preserves subgroup ratios), cluster sampling. Enables statistical generalizability.\n   - **Non-Probability Sampling**: Convenience, purposive, snowball. High risk of selection bias; strictly bounded generalizability.\n2. **Sample Size & Statistical Power Calculations**:\n   - Calculate sample size $N$ in advance based on desired effect size (Cohen's $d$), significance level ($\\alpha = 0.05$), and statistical power ($1 - \\beta = 0.80$):\n     - Underpowered studies ($N$ too small) produce high false-negative rates and inflated effect sizes when statistical significance is achieved by chance (Winner's Curse).\n\n### Step 3: Experimental Controls & Threats to Validity\n1. **Neutralizing Threats to Internal Validity**:\n   - *Selection Bias*: Mitigated by strict computer-generated **random assignment**.\n   - *History \u002F Maturation*: Mitigated by parallel **Control Groups** experiencing identical time intervals without the intervention.\n   - *Observer \u002F Hawthorne Effect*: Mitigated by **Double-Blind** protocols where neither participant nor researcher knows group allocation.\n2. **Threats to External Validity**:\n   - Ensure the sample demographic and testing environment reflect the real-world operational domain (ecological validity).\n\n### Step 4: Statistical Test Alignment (Parametric vs Non-Parametric)\n1. **The Decision Matrix for Statistical Tests**:\n   - Continuous DV + Normal Distribution + Independent Groups (2 groups) $\\to$ **Two-Sample Independent $t$-Test**.\n   - Continuous DV + Normal Distribution + Repeated Measures (Same group) $\\to$ **Paired $t$-Test**.\n   - Continuous DV + Normal Distribution + $\\ge 3$ Groups $\\to$ **One-Way ANOVA** (with Tukey HSD post-hoc).\n   - Ordinal DV (Likert scales) or Skewed Distribution (2 groups) $\\to$ **Mann-Whitney $U$ Test** (Wilcoxon Rank-Sum).\n   - Categorical IV + Categorical DV $\\to$ **Chi-Square ($\\chi^2$) Test of Independence**.\n   - Continuous DV + Multiple Continuous\u002FCategorical IVs $\\to$ **Multiple Linear Regression**.\n\n---\n\n## Deliverable Format: Methodological Study Blueprint (RCT Protocol)\n\n```markdown\n# Study Protocol: Evaluating Spaced Retrieval in Software Engineering Onboarding\n\n### 1. Research Hypothesis\nEngineers onboarded using a spaced-retrieval CLI interactive tool will demonstrate significantly higher code architecture comprehension ($p \u003C 0.01$, Cohen's $d \\ge 0.5$) at Day 30 compared to engineers onboarded using traditional monolithic wiki documentation.\n\n### 2. Experimental Design\n- **Type**: Randomized Controlled Trial (Between-Subjects).\n- **Independent Variable (IV)**: Onboarding Modality:\n  - *Treatment Group*: Spaced retrieval CLI tool (10-min active recall challenges on Days 1, 3, 7, 14, 28).\n  - *Control Group*: Standard company Confluence onboarding wiki (unconstrained self-study).\n- **Dependent Variable (DV)**: Architecture Comprehension Score (0-100 on a validated 20-item diagnostic scenario exam administered on Day 30).\n\n### 3. Sampling & Power Analysis\n- Power calculation ($\\alpha = 0.05, 1 - \\beta = 0.80$, expected Cohen's $d = 0.60$): Requires minimum $N = 45$ participants per group ($N_{\\text{total}} = 90$).\n- Recruitment: 96 newly hired engineers across 4 enterprise business units, stratified by years of prior software experience, then randomly assigned 1:1.\n\n### 4. Statistical Analysis Plan\n1. Test normality of Day 30 score distribution via Shapiro-Wilk test.\n2. If normal: Independent two-sample $t$-test; report mean difference and 95% Confidence Intervals.\n3. If non-normal: Mann-Whitney $U$ non-parametric test.\n4. Multiple regression controlling for years of experience and prior programming language familiarity.\n```\n\n---\n\n## Worked Example: Rescuing a Flawed Product A\u002FB Test Methodology\n\n- **Problem**: A growth team ran an A\u002FB test claiming a new checkout button increased conversions by 14% ($p = 0.04$). Two months later, total company revenue had declined.\n- **Methodological Audit**:\n  1. *Peeking Bias*: The growth team had checked the dashboard every morning and stopped the test the moment $p \u003C 0.05$ appeared (Day 4), inflating false-positive rates to over 30%.\n  2. *Novelty Effect*: Returning power users clicked the bright new button out of curiosity, not increased purchase intent.\n- **Methodological Remedy**: Implemented fixed-horizon testing: mandatory 14-day run time to capture full weekly seasonal cycles, with sample sizes locked in advance via power analysis.\n- **Result**: The 14% conversion lift vanished in the controlled re-test, preventing a disastrous permanent UI overhaul.\n\n---\n\n## Verification Checklist\n\n- [ ] Research paradigm matches the underlying epistemic question.\n- [ ] Variables clearly operationalized with validated measurement instruments.\n- [ ] Sample size determined prior to data collection via formal statistical power calculations.\n- [ ] Threats to internal validity (selection, maturation, confounders) controlled.\n- [ ] Statistical tests aligned with variable data types (parametric vs non-parametric).\n\n---\n\n## Anti-Patterns\n\n- **P-Hacking \u002F Data Dredging**: Running 40 different statistical combinations until one randomly shows $p \u003C 0.05$ and inventing a hypothesis after the fact (HARKing).\n- **Premature Test Peeking**: Halting an A\u002FB test or experiment early the moment the significance indicator flashes green.\n- **Convenience Bias Generalization**: Running an experiment exclusively on 30 undergraduate psychology students and claiming the findings apply to all human beings.\n",{"name":5,"group":24,"description":7,"last_reviewed":25},"Research",1788652800000,"main","7ae884a84e46720ee4774a3d0e15b310064a9c4e","skills\u002Fresearch-methods",0,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskillary","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp","poorvithmp.com","Shipping tools for problems people notice too late.","2026-10-02 21:32:26",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add poorvith-mp\u002Fskills-education --skill research-methods",{"tool":42,"command":43},"bunx","bunx skills add poorvith-mp\u002Fskills-education --skill research-methods",{"tool":45,"command":46},"pnpm","pnpm dlx skills add poorvith-mp\u002Fskills-education --skill research-methods","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fresearch-methods\u002FSKILL.md",true,[50,56],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"evals\u002Frouting.yaml","REFERENCE","参考文档",1067,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fresearch-methods\u002Fevals\u002Frouting.yaml",{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"research-methods.skill","OTHER","其他",4201,"https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fresearch-methods\u002Fresearch-methods.skill",[63,70,77,83,90,96],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M792EBE3X3WHX57AMJNNMGTN","training-design","Design enterprise training: needs analysis, curriculum structure, blended delivery and evaluation. Use when developing corporate workshops, course modules, or adult learning.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Ftraining-design\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M792E6RWHNQV19CHD9WMMVSW","topic-selection","Turn a broad interest into a tractable research question or essay topic with defensible scope and available sources. Use when narrowing research questions, thesis scopes, or project ideas.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Ftopic-selection\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M792E447NN2HJQAPWEKX2KNQ","study-plan","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fstudy-plan\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M792E17ZWKR12JQ89MN7PK0D","study-plan-builder","Moved to modernized catalog in skillary. Removed in v4.1.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fstudy-plan-builder\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":19},"01M792DZSR1NRRDRABT0XACRWA","study-abroad-advisor","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fstudy-abroad-advisor\u002FSKILL.md",[],[],{"skillNo":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":19},"01M792DVAE1YB5T8DTS3B2V52J","source-evaluation","Judge whether sources carry claims: primary vs secondary, funding, method quality, and replication. Use when auditing source credibility or methodology.","https:\u002F\u002Fgithub.com\u002Fpoorvith-mp\u002Fskills-education\u002Fblob\u002Fmain\u002Fskills\u002Fsource-evaluation\u002FSKILL.md",[],[]]