Data analysis can feel like one of the hardest parts of a CIPD dissertation. You may have spent weeks designing a questionnaire, arranging interviews or collecting information from an organisation, only to reach the analysis stage and wonder what you are actually supposed to do with all of it.
The good news is that data analysis does not have to mean complicated statistics or endless tables.
In my experience, the best place to start is much simpler: look at your research question first. Once you know exactly what you are trying to find out, it becomes much easier to decide what data you need, how you should analyse it and what your findings actually mean.
For CIPD research, this matters because evidence-based practice is central to the people profession. HR decisions should be supported by credible evidence rather than assumptions, personal opinions or workplace anecdotes.
What is data analysis in a CIPD dissertation?
Data analysis is the process of making sense of the information you have collected during your research.
Depending on your dissertation topic, that information could come from employee questionnaires, interviews, focus groups, HR records, engagement surveys, turnover figures or existing research.
For example, imagine your dissertation investigates whether flexible working affects employee engagement.
You might collect questionnaire responses from employees and ask them about their working arrangements, job satisfaction, workload and engagement. You could then analyse the responses to identify patterns.
Alternatively, you might interview employees and ask them how flexible working has affected their motivation, relationships with colleagues and work-life balance.
Neither approach is automatically better. The right choice depends on what you are trying to discover.
The CIPD’s guidance on people analytics makes a similar distinction between numerical information, which can help identify patterns and relationships, and qualitative information, which can provide deeper insight into people’s experiences.
Start with your research question
Before you open Excel or statistical software, go back to your research objectives.
This sounds obvious, but it is surprisingly easy to lose sight of them once you have a large amount of data in front of you.
Suppose one of your objectives is:
To examine whether flexible working is associated with employee engagement.
You now have a clear direction.
You may want to compare engagement scores between employees with different working arrangements. You might also investigate whether employees who have greater flexibility report higher levels of engagement.
If your research instead asks:
How do employees experience flexible working?
then interviews may be more useful because you are interested in people’s experiences and perceptions rather than simply numerical differences.
I find it helpful to create a basic analysis table before starting the actual analysis.
| Research objective | Information needed | Possible approach |
|---|---|---|
| Measure employee engagement | Survey responses | Frequencies, percentages and averages |
| Compare employee groups | Grouped survey responses | Comparative analysis |
| Explore employee experiences | Interview transcripts | Thematic analysis |
| Examine relationships | Numerical variables | Correlation or regression |
| Understand unexpected results | Survey and interview findings | Combined interpretation |
This simple exercise can save a lot of time later.
Decide what type of data you have
Your analysis will usually fall into one of two broad categories: quantitative or qualitative.
Quantitative data
Quantitative data is numerical.
You might have information such as:
- Employee engagement scores
- Number of employees leaving
- Absence rates
- Training hours
- Salary levels
- Survey ratings
- Length of service
You can use this information to identify patterns and differences.
For instance, if 80 out of 120 employees say they are satisfied with their current training opportunities, you can report that 66.7% of respondents expressed satisfaction.
But don’t stop at the percentage.
The more interesting question is what that finding tells you about your research problem.
Qualitative data
Qualitative data is generally concerned with people’s experiences, opinions and perceptions.
Interviews are a common example.
Imagine that you interview ten employees about working from home. Several might talk about having greater control over their working day. Others might mention feeling disconnected from colleagues.
Instead of simply listing these comments, you can organise them into broader themes.
You might eventually identify themes such as:
- Autonomy and flexibility
- Social connection
- Management support
- Work-life boundaries
This is where qualitative analysis becomes useful. It allows you to look beneath the surface of the responses.
Clean your data before analysing it
One mistake I would strongly recommend avoiding is jumping straight into analysis.
Your dataset needs to be checked first.
With questionnaire data, look for things such as:
- Missing answers
- Duplicate responses
- Incorrect entries
- Inconsistent categories
- Unusual values
- Incomplete questionnaires
Imagine that respondents were asked to rate job satisfaction from 1 to 5 and one response is recorded as 50.
That is probably an error rather than an extraordinary level of satisfaction.
You need to identify these issues before calculating averages or running statistical tests.
The same principle applies to interview research. Check transcripts carefully for obvious transcription errors and remove information that could unnecessarily identify participants.
It is also worth keeping a short record of the changes you make. If you remove incomplete responses, for example, explain how many were removed and why.
That small amount of transparency can make your methodology much stronger.
Use descriptive statistics first
If you are working with numerical data, start with the basics.
Descriptive statistics can tell you what your dataset looks like.
You might examine:
- Frequencies
- Percentages
- Mean scores
- Median scores
- Standard deviation
- Minimum and maximum values
For example, suppose you survey 150 employees and discover that employees working remotely have an average engagement score of 4.1 out of 5, while employees working mainly in the office have an average score of 3.5.
That is worth reporting.
However, you need to be careful about how you interpret it.
You cannot automatically say that remote working caused the difference.
There may be other explanations. Perhaps remote employees have greater job autonomy. Perhaps they work in different departments. Perhaps they have more experienced managers.
This distinction between association and causation is particularly important in HR research.
The CIPD’s people analytics guidance also highlights the importance of distinguishing between describing what has happened and understanding why it may have happened.
When should you use correlation?
Correlation can be useful when you want to examine whether two variables are related.
For example, you might want to investigate whether employee engagement tends to increase alongside perceived management support.
If your analysis shows a positive relationship, employees who report stronger management support may also report higher engagement.
But again, be careful.
A correlation does not prove that management support caused higher engagement.
There could be other factors influencing both variables.
This is one of those areas where a dissertation can become stronger simply by being honest about what the evidence can and cannot tell you.
What about regression analysis?
Regression is another option when you want to examine the relationship between an outcome and one or more potential explanatory variables.
For example, you could investigate whether employee engagement is associated with:
- Management support
- Workload
- Career-development opportunities
- Flexible working
- Recognition
Regression can be useful, but there is no requirement to use it simply because it sounds more advanced.
If your sample is small or your research design does not support the analysis, forcing a complex statistical technique into the dissertation may actually weaken your work.
Your method should fit your research question and data.
How to analyse interview data
Qualitative analysis needs a slightly different approach.
One of the most practical methods is thematic analysis.
The basic idea is to look through your interview material and identify recurring ideas or patterns.
You could approach it like this:
- Read through the transcripts carefully.
- Highlight sections that appear relevant to your research question.
- Give those sections short codes.
- Compare the codes across interviews.
- Look for connections between them.
- Group related codes into potential themes.
- Review the themes and remove anything that does not have enough supporting evidence.
- Define what each final theme actually means.
- Use selected quotations to demonstrate the findings.
For example, imagine several employees describe difficulty switching off after working from home.
You might initially code comments as “longer working hours”, “checking emails at night” and “difficulty switching off”.
Those codes could eventually contribute to a broader theme such as blurred work-life boundaries.
The important thing is not simply to count how many people said something.
You should consider why participants expressed that view and what it tells you about the research question.
Braun and Clarke’s work on thematic analysis is widely used in qualitative research and provides a useful academic foundation for this approach.
Don’t confuse results with discussion
There is an important difference between presenting your findings and discussing them.
The results section should tell the reader what you found.
The discussion should explain what those findings mean.
For example:
Result: Employees reporting greater access to career development also reported higher engagement.
Discussion: This could suggest that employees value opportunities to develop their careers and may feel more committed when they believe their employer is investing in their future.
The second statement is much more analytical.
You can then compare your finding with previous academic research and consider whether other explanations are possible.
This is where you demonstrate critical thinking.
Think carefully about reliability and validity
Good analysis depends on good-quality data.
If your questionnaire measures something poorly, sophisticated statistical analysis will not magically fix the problem.
Ask yourself:
- Did the questions actually measure what I intended?
- Was the sample suitable for my research?
- Were there enough responses?
- Could respondents have given socially desirable answers?
- Was the data collected consistently?
- Could my own assumptions have influenced the interpretation?
For qualitative research, you should also think about your role as the researcher.
Your background, expectations and relationship with participants can potentially influence how you interpret what people tell you.
Being open about these issues is not a weakness.
In fact, acknowledging them usually makes the research appear more credible.
Protect your participants’ information
HR research can involve sensitive information.
Employees may discuss managers, pay, workplace relationships, performance, health-related absences or other personal matters.
That means confidentiality needs to be taken seriously.
Think about:
- How participants gave consent
- Where research data is stored
- Who can access it
- How participants are anonymised
- Whether quotations could identify someone
- How organisationally sensitive information will be handled
The UK Statistics Authority provides guidance on ethical issues associated with qualitative research, including transparency, accountability and governance.
For research involving personal data in the UK, the Information Commissioner’s Office also provides guidance concerning research under UK data-protection legislation.
Do not assume that replacing someone’s name with “Participant 1” automatically makes them anonymous. In a small organisation, a combination of job title, department and personal details could still make a person identifiable.
Present your findings without overwhelming the reader
A dissertation does not become more academic because every possible result is placed into a table.
Use tables and charts when they make your findings easier to understand.
For quantitative research, you might use a bar chart to compare engagement between employee groups.
For qualitative research, you could present your themes and then support each theme with a small number of relevant participant quotations.
The key is relevance.
If a table does not help answer your research question, ask yourself whether it really needs to be there.
The CIPD’s professional standards also place importance on communicating analytical findings clearly and using appropriate forms of data visualisation.
Common mistakes in CIPD dissertation data analysis
There are several problems I see repeatedly in student research.
Using complicated statistics unnecessarily
More advanced statistics do not automatically mean better research.
If a simple method answers the question properly, use it.
Treating correlation as proof of cause
Two things can be related without one causing the other.
Reporting numbers without explaining them
Don’t leave your reader to work out why a particular percentage matters.
Explain its relevance.
Selecting only favourable findings
If your results contradict your expectations, discuss them.
Unexpected results can be some of the most interesting parts of a dissertation.
Using too many quotations
Participant quotations should support your interpretation. They should not become a substitute for analysis.
Ignoring limitations
Every research project has limitations.
Perhaps your sample was relatively small. Perhaps you used convenience sampling. Perhaps respondents were from one organisation or department.
Be honest about these limitations and explain how they affect the interpretation of your findings.
A straightforward process you can follow
If you are unsure where to begin, use this sequence:
- Read your research question again.
- Match each objective with the data you collected.
- Separate quantitative and qualitative information.
- Clean and organise the data.
- Decide which analytical method fits each objective.
- Start with descriptive analysis.
- Carry out additional statistical tests only where appropriate.
- Code interview data and develop themes where relevant.
- Interpret your findings rather than simply describing them.
- Compare the findings with existing research.
- Discuss unexpected results and limitations.
- Explain what the findings mean for HR practice.
Following this process keeps your analysis focused.
It also stops you from making the common mistake of analysing everything simply because the information happens to be available.
How to make your analysis more convincing
The strongest CIPD dissertations are not necessarily the ones with the biggest datasets or the most complicated statistical tests.
They are usually the ones where the research question, methodology, analysis and conclusions fit together.
When I approach data analysis, I would keep the question. Check your data before analysing it. Explain your decisions clearly and three questions in mind:
What does the evidence show?
What does it mean?
How confident can I be in that interpretation?
Those questions encourage you to move beyond simply reporting results.
For example, finding that employees with access to flexible working report higher satisfaction is useful. Exploring why that might be happening is more valuable. Considering alternative explanations and comparing the result with previous research takes the analysis another step further.
That is the kind of critical thinking expected in academic research and evidence-based HR practice.
If you need additional help with planning the wider assignment, methodology or academic structure, CIPD Assignment Writing can be a useful starting point. Any external support should, however, strengthen your understanding of the research rather than replace your own analysis and judgement.
Final thoughts
Analysing data for a CIPD dissertation is not about trying to make your research look complicated.
It is about making sense of the evidence you have collected and using it responsibly.
Start with the research question. Choose a method that fits the question. Check your data before analysing it. Explain your decisions clearly and avoid making claims that your evidence cannot support.
Most importantly, don’t stop once you have calculated the percentages or identified the themes.
Ask what the findings actually mean.
Compare them with previous research. Think about alternative explanations. Acknowledge the limitations. Then consider what the findings could mean for HR practice.
When those pieces come together, your data analysis becomes much more than a section full of numbers and quotations. It becomes the part of your dissertation that demonstrates your ability to think critically, work with evidence and turn research findings into meaningful professional insight.
