A Graduate Student's Guide to Thesis Data Analysis

Your thesis data analysis can make or break your research project. It's the part where many graduate students feel the most anxiety—and for good reason. Choosing the wrong statistical test or misinterpreting results can derail months of work.

This guide will walk you through the process step by step.

Before You Collect Data: Plan Your Analysis

The biggest mistake students make is collecting data first and figuring out the analysis later. Your analysis approach should be determined before data collection.

During your proposal stage, answer:

• What are your research questions/hypotheses?
• What type of data will you collect? (Quantitative, qualitative, mixed?)
• What variables are you measuring?
• What statistical tests align with your research design?

Write this into your methodology chapter. Your supervisor will thank you.

Choosing the Right Statistical Test

Here's a simplified decision guide:

Comparing two groups?

• Independent groups → Independent samples t-test
• Same group measured twice → Paired samples t-test

Comparing three or more groups?

• One factor → One-way ANOVA
• Two factors → Two-way ANOVA

Looking at relationships between variables?

• Two continuous variables → Correlation (Pearson or Spearman)
• Predicting an outcome → Regression analysis

Categorical data?

• Frequencies in categories → Chi-square test
• Association between categories → Cross-tabulation

Important: These are starting points. Your specific research design may require more nuanced approaches. Consult your supervisor or a biostatistician.

Data Collection Tips

Good analysis starts with good data:

1. Use a well-designed questionnaire — pilot test with 5-10 people first
2. Code your data consistently — decide coding schemes before you start
3. Enter data carefully — double-check a random sample of entries
4. Back up everything — multiple copies in multiple locations
5. Document your process — keep a research diary

Cleaning Your Data

Before any analysis:

• Check for missing data and decide how to handle it
• Look for outliers — are they real or data entry errors?
• Verify data types (numeric, categorical, date)
• Check for duplicate entries
• Ensure coding is consistent

Document every cleaning decision. Your examiner may ask why you removed certain data points.

Performing the Analysis

Whether you use SPSS, R, or Python:

1. Start with descriptive statistics — means, frequencies, standard deviations
2. Check assumptions — normality, homogeneity of variance
3. Run your tests — one hypothesis at a time
4. Record effect sizes — not just p-values
5. Create clear tables and figures

Presenting Results

Your results chapter should:

• Present findings systematically (often in order of research questions)
• Use tables for detailed numbers
• Use figures for trends and comparisons
• Report statistical values properly (e.g., t(48) = 2.31, p = .025)
• Separate results from interpretation (save that for the discussion)

Common formatting standards:

• APA style for social sciences and health
• Report exact p-values (not just "p < .05")
• Include confidence intervals where appropriate
• Use consistent decimal places

Common Thesis Data Analysis Mistakes

1. p-hacking: Running multiple tests until something is "significant"
2. Ignoring assumptions: Every test has conditions that must be met
3. Over-interpreting results: A significant result isn't always a meaningful one
4. Incomplete reporting: Missing sample sizes, test statistics, or effect sizes
5. Not addressing limitations: Every study has them—own yours

When to Ask for Help

There's no shame in seeking support:

• When you're unsure which test to use
• When your results don't make sense
• When a reviewer asks for additional analyses
• When you need help with statistical software

---

Need hands-on support with your thesis data analysis? Our Research Support service helps health science students with everything from methodology design to results interpretation.