Create systematic data analysis plans with our free data analysis plan generator. No registration, no fees - just comprehensive tools for planning quantitative, qualitative, and mixed methods analyses.
What is a Data Analysis Plan?
A data analysis plan is a detailed document specifying exactly how you'll analyze research data to answer your research questions. It outlines variables, measurement levels, analytical techniques, software tools, and step-by-step procedures. Creating analysis plans before data collection ensures appropriate methods and prevents ad-hoc analyses lacking clear rationale.
Why Plan Analysis in Advance?
- Methodological rigor - Pre-specified analyses demonstrate scientific integrity
- Resource planning - Identify needed software, training, or statistical consultation
- IRB requirements - Many IRBs require analysis plans in proposals
- Dissertation defense - Committees evaluate analysis appropriateness before data collection
- Grant applications - Funders want evidence you can successfully analyze proposed data
- Prevents fishing - Pre-registration reduces questionable research practices
Quantitative Analysis Plans
Variable Specification
List all variables with characteristics:
- Variable names - Clear, descriptive labels
- Measurement level - Nominal, ordinal, interval, ratio
- Variable type - Independent, dependent, moderator, mediator, control
- Measurement method - How variable is assessed
- Expected values - Range or categories
Example table: | Variable | Level | Type | Measurement | |----------|-------|------|-------------| | Test Score | Ratio | Dependent | 0-100 scale | | Study Time | Ratio | Independent | Hours per week | | Gender | Nominal | Control | Male/Female/Other |
Descriptive Statistics
Specify descriptive analyses for sample characterization:
- Continuous variables - Mean, median, standard deviation, range
- Categorical variables - Frequencies, percentages
- Distribution assessment - Histograms, skewness, kurtosis
- Missing data summary - Patterns and percentages of missingness
Assumption Testing
Identify assumptions requiring testing:
- Normality - Shapiro-Wilk test, Q-Q plots (for t-tests, ANOVA, regression)
- Homogeneity of variance - Levene's test (for t-tests, ANOVA)
- Independence - Durbin-Watson statistic (for regression)
- Linearity - Scatterplots (for correlation, regression)
- Multicollinearity - VIF values (for multiple regression)
Primary Analyses
Link each research question to specific analyses:
RQ1: Is there a difference in test scores between groups?
- Analysis: Independent samples t-test
- Variables: DV = test scores (continuous), IV = group (two categories)
- Assumptions: Normality, homogeneity of variance, independence
- Effect size: Cohen's d
- Alpha level: .05
RQ2: What factors predict test performance?
- Analysis: Multiple regression
- Variables: DV = test scores, IVs = study time, prior GPA, motivation
- Assumptions: Normality of residuals, homoscedasticity, linearity, no multicollinearity
- Effect sizes: R², adjusted R², β coefficients
- Alpha level: .05
Handling Violations
Specify contingency plans if assumptions are violated:
- Non-normality: Transform variables or use non-parametric alternatives (Mann-Whitney U, Kruskal-Wallis)
- Heterogeneity: Use Welch's t-test or Brown-Forsythe F-test
- Multicollinearity: Remove redundant predictors or combine into composites
- Outliers: Winsorize, remove with justification, or use robust methods
Missing Data Strategy
Document missing data handling:
- MCAR, MAR, MNAR assessment - Test patterns of missingness
- Imputation methods - Multiple imputation, mean substitution, or deletion approaches
- Sensitivity analyses - Compare results with different missing data approaches
Qualitative Analysis Plans
Data Preparation
Specify preparation procedures:
- Transcription - Verbatim or intelligent verbatim, notation conventions
- Data organization - File naming, folder structure, backup procedures
- Anonymization - How identifiers will be removed or pseudonymized
- Quality checks - Accuracy verification, member checking plans
Analytical Approach
State your qualitative methodology:
- Thematic analysis - Inductive or deductive coding
- Grounded theory - Open, axial, selective coding procedures
- Phenomenology - Horizonalization, imaginative variation
- Narrative analysis - Structural or thematic narrative approaches
- Content analysis - Manifest or latent content coding
Coding Procedures
Detail systematic coding process:
- Coding rounds - How many passes through data
- Codebook development - Initial codes vs. emergent codes
- Multiple coders - Inter-rater reliability procedures if applicable
- Memo writing - Frequency and focus of analytical memos
- Theme development - How codes become categories become themes
Quality Strategies
Document trustworthiness approaches:
- Credibility - Member checking, prolonged engagement, triangulation
- Transferability - Thick description, purposeful sampling rationale
- Dependability - Audit trail, reflexive journal
- Confirmability - Data-grounded claims, researcher positioning statements
Software and Tools
Specify analysis software:
- QDAS programs - NVivo, MAXQDA, Atlas.ti, Dedoose
- Coding methods - Software-assisted or manual coding
- Version control - How project files will be backed up and versioned
Mixed Methods Analysis Plans
Integration Strategy
Specify how qualitative and quantitative data will be integrated:
Convergent Design - Analyze separately, then merge results
- Quantitative phase: Descriptive and inferential statistics
- Qualitative phase: Thematic analysis
- Integration: Joint display comparing and contrasting findings
Sequential Design - One phase informs the next
- Initial phase: Survey providing quantitative overview
- Follow-up phase: Interviews exploring significant quantitative findings
- Integration: Qualitative data explains quantitative patterns
Embedded Design - One dataset supports the other
- Primary: Experiment with quantitative outcome measures
- Supplementary: Interviews during intervention explaining participant experiences
- Integration: Qualitative data contextualizes quantitative results
Meta-Inferences
Plan how final integrated conclusions will be drawn:
- Confirmation - Do qualitative and quantitative findings agree?
- Expansion - Do findings from different methods address different questions?
- Discordance - If findings conflict, how will conflicts be resolved?
Timeline and Resources
Analysis Timeline
Estimate time for each phase:
- Data cleaning: 2 weeks
- Descriptive statistics: 1 week
- Assumption testing: 3 days
- Primary analyses: 2 weeks
- Visualization: 1 week
- Results interpretation: 1 week
- Total: 7-8 weeks
Resource Requirements
Identify needed resources:
- Software: SPSS Statistics Standard (quantitative), NVivo Pro (qualitative)
- Training: Workshop on multilevel modeling
- Consultation: Statistical consultant for power analysis validation
- Hardware: Computer with minimum 16GB RAM for large dataset handling
Common Pitfalls
Analysis-Question Mismatch
Ensure analyses actually answer research questions. Statistical sophistication doesn't compensate for analyses that don't address your questions directly.
Fishing Expeditions
Pre-specify primary analyses. Running dozens of tests hoping something is significant is poor practice. Exploratory analyses are fine when labeled as exploratory, not confirmatory.
Ignoring Assumptions
Don't skip assumption testing. Violating assumptions produces unreliable results. Test assumptions and use appropriate corrections or alternative methods when violated.
Over-Complexity
Simpler analyses clearly addressing research questions beat unnecessarily complex analyses impressing reviewers. Complexity should serve understanding, not showcase statistical knowledge.
Transform Your Research Analysis
Stop approaching data analysis reactively. Create comprehensive analysis plans that ensure appropriate methods, satisfy review requirements, and produce trustworthy findings.
Visit https://www.subthesis.com/tools/data-analysis-plan-generator - Start planning your analysis today, no registration required!