Key Takeaways
- Not understanding your data is more common than students admit — you're not alone.
- The problem is almost always one of three things: the wrong analysis approach, unfamiliarity with the software, or unclear research questions.
- Understanding your data doesn't mean being a statistician — it means being able to answer your research questions with evidence.
- Expert data analysis support can transform confusing outputs into a clear, examiner-ready chapter.
First: You're Not as Lost as You Think
"I don't understand my dissertation data" is one of the most common things students say when they reach out for help. It doesn't mean you're bad at research. It usually means one of three things:
- You collected data without a clear enough analytical plan
- You're using analysis software you weren't adequately trained in
- Your research questions are too vague to guide your interpretation
All of these are fixable. Let's diagnose which applies to you.
Diagnose the Real Problem
Problem 1: Your Analysis Approach Doesn't Match Your Data
This is the most serious issue — but also the most common. It happens when students choose a methodology in their proposal without fully understanding the analytical implications.
Signs this is your problem:
- You collected qualitative data but don't know how to identify "themes"
- You ran statistical tests in SPSS but don't know what the output means
- You have data but can't connect it to your research questions
- Your results section is empty because you don't know what to put in it
Problem 2: Software Unfamiliarity
SPSS, NVivo, R, and Atlas.ti all have steep learning curves. Many students choose these tools because their university recommends them, without adequate training. If you can navigate the software but don't understand what the outputs mean, this is your issue.
See our comparison guide on best software for dissertation data analysis for a breakdown of what each tool does and when to use it.
Problem 3: Vague Research Questions
If your research questions are unclear, your analysis will be aimless. You'll end up with lots of data and no idea which parts answer what. Tightening your research questions — even at this stage — can provide the analytical framework your interpretation needs.
What "Understanding Your Data" Actually Means
Many students think understanding data means being able to perform complex statistical operations or produce sophisticated visualisations. That's not what examiners are evaluating.
Understanding your data means:
- Being able to state clearly what your data shows in relation to your research questions
- Knowing which findings are significant and which are not
- Being able to explain patterns, relationships, or themes in plain language
- Linking your findings back to the literature you reviewed
Quantitative Data: What to Do When SPSS Output Makes No Sense
Start With Descriptive Statistics
Before running inferential tests, run descriptive statistics: frequencies, means, standard deviations. These tell you the basic characteristics of your data and help identify anomalies before you proceed to analysis.
Understand What Each Test Is Telling You
| Test | What It Answers | What to Report |
|---|---|---|
| T-test | Is there a difference between two groups? | t-value, degrees of freedom, p-value, mean difference |
| ANOVA | Is there a difference between three or more groups? | F-value, p-value, post-hoc results |
| Pearson's r | Is there a correlation between two variables? | r-value, direction, p-value |
| Regression | Does one variable predict another? | β values, R², p-values per predictor |
| Chi-square | Is there an association between categorical variables? | χ² value, df, p-value |
Focus on p-values and Effect Sizes
A result is statistically significant when p < 0.05. But statistical significance alone is not enough — report effect sizes (Cohen's d, η², r) to demonstrate the practical importance of your findings.
Qualitative Data: What to Do When You Can't Find Themes
Start With Open Coding
Read your transcripts or documents line by line and attach descriptive labels (codes) to meaningful segments. Don't try to find themes immediately — codes come first, themes emerge from clusters of codes.
Group Codes Into Categories, Then Themes
Once you have 50–100+ codes, look for patterns. Which codes appear repeatedly? Which are conceptually similar? Group related codes into categories, then identify what overarching theme each category represents.
Check Themes Against Your Research Questions
Every theme should answer part of a research question. If you have a theme that doesn't connect to any of your research questions, it's either tangential (exclude it) or signals that your questions need refining.
When to Get Expert Help With Data Analysis
If you've attempted analysis and still don't understand what your data is telling you, or if your deadline doesn't allow time for a learning curve, professional data analysis support is the right choice.
Our statisticians and qualitative analysts can:
- Run the appropriate statistical tests and explain what the outputs mean
- Code and thematically analyse qualitative data
- Write a clear, examiner-ready results chapter based on your data
- Produce a discussion chapter that interprets findings in relation to your literature review
Summary
Not understanding your dissertation data is common and fixable. Diagnose whether the problem is methodological mismatch, software unfamiliarity, or vague research questions — then address it systematically. If you need expert support to analyse or interpret your data, our team of PhD-qualified analysts is ready to help.