Data Analysis

The approach depends on the data. Quantitative data consists of numbers, such as how many pupils attend a homework club. Qualitative data provides descriptive detail, such as pupils' explanations of why they attend. Numbers help reveal the extent of a pattern; detailed accounts help explain people's experiences and meanings.

Preparing data carefully

First, researchers organise their material. Questionnaire answers can be entered into a spreadsheet, while interview recordings can be transcribed into written accounts. Responses must remain linked to the correct questions and participants, without revealing participants' identities in published findings.

Researchers then check for errors and missing information. For example, a number may have been entered incorrectly, or someone may have answered a question intended only for homework-club members. Such problems need checking before the results are summarised.

Missing answers must not automatically be treated as ‘no’. Researchers should explain how incomplete responses have been handled and how many people answered each question. They must not remove an answer simply because it contradicts their expectations. Cleaning data should improve its accuracy, not make it fit a preferred conclusion.

Coding means assigning labels or categories to data. A closed questionnaire answer might be coded numerically for processing, while an interview passage might be labelled ‘transport difficulty’. Numerical codes for categories are labels: they do not necessarily represent measurable quantities.

Finding patterns in quantitative data

A frequency is the number of times an answer occurs. Frequency tables organise these counts, while percentages help compare groups with different numbers of respondents. Always establish the total on which a percentage is based. If a question allows several answers, its percentages may add to more than 100% because one person can appear in several categories.

Suppose a researcher surveys 40 pupils in a UK secondary school about attending a homework club. The following constructed table compares two year groups:

More of the Year 10 respondents attend than the Year 11 respondents. Because the groups contain equal numbers, comparing their attendance counts is straightforward. If their totals differed, percentages would provide a fairer comparison. A table comparing responses across groups like this is called a cross-tabulation.

Bar charts can make differences between categories easier to see. Their labels must make clear what is being counted and which groups are being compared. A trend is a pattern of change over time; a difference between two year groups in a single survey is not, by itself, a trend over time.

For numerical measurements, researchers can summarise a typical value using the mean (the total divided by the number of values), median (the middle value when values are ordered) or mode (the most frequent value). These summaries can give different impressions: an unusually high value can pull the mean upwards. Looking at the spread of values as well as an average helps avoid hiding important differences.

Finding meanings in qualitative data

Qualitative analysis involves reading accounts carefully, coding relevant passages and comparing them to identify themes: recurring ideas or issues. Researchers should examine the whole set of responses rather than select only memorable quotations.

For example, suppose pupils discussing a homework club describe difficulty travelling home, help from teachers and pressure from examinations. These could become themes about transport, support and competing demands. A single response may belong to more than one theme.

Short quotations can show how participants expressed their experiences, but they should preserve the speaker's meaning and protect their identity. Less common views also matter. An account that differs from the dominant theme may reveal that pupils experience the same situation differently. The most frequently mentioned theme is not automatically the most important one.

Drawing a justified conclusion

Researchers return to their aim, the purpose of the study, or their hypothesis, a testable prediction. They ask whether the findings answer the question and whether the evidence supports or challenges the prediction. Unexpected findings should be reported, not hidden.

The homework-club table supports a conclusion about attendance among the pupils surveyed. It does not explain why the groups differ. Correlation, an association between variables, does not establish causation, where one factor produces a change in another. Differences in transport, club availability or other commitments might help explain the pattern.

A sound conclusion also considers the quality and limits of the research. Did questions accurately capture what pupils meant, giving valid findings? Were recording and coding procedures consistent, supporting reliability? Could leading questions or the researcher's expectations have biased the results?

Finally, consider how far the findings can be generalised. A small study in one school cannot automatically establish what happens among all UK pupils. A justified conclusion states what the evidence supports, acknowledges important exceptions and limitations, and identifies questions that remain unanswered.