9.1.1, 9.1.4, 1.2.3, 2.2.3, 2.2.4, D.1.15, D.1.16 - Data, aims and hypotheses
Psychological research turns a question about behaviour into evidence that can support or challenge a prediction. Learn how to distinguish types and sources of data, define measurable variables and write hypotheses whose meaning is clear before the results are known.
What the data tell us
The form of the evidence determines what a researcher can learn from it.
Quantitative data express amounts, scores or frequencies numerically. The number of words recalled, response time in seconds and a rating from 1 to 5 are examples. Numbers allow comparisons, but a score alone may not explain why someone responded that way. Qualitative data describe meanings, experiences or behaviour in words or other non-numerical material. An interview account of why someone prefers studying with friends can reveal reasons the researcher did not anticipate, although interpreting different accounts requires care.
The method does not automatically decide the data type. An interview can ask for the number of study sessions last week and then ask what made them useful: the first answer is quantitative, the second qualitative. An observation can record a count of offers of help alongside descriptive field notes about the interaction. A questionnaire is therefore not always quantitative, and an observation is not always qualitative.
Qualitative and quantitative data
Qualitative data describe qualities or meanings; quantitative data represent measured amounts, scores or counts. The distinction concerns the evidence recorded, rather than whether a researcher used an interview, questionnaire or observation.
A rating is a person's report converted into an ordered number; it is not an objective reading of an emotion. Similarly, assigning the labels 1 and 2 to two interview groups does not turn their accounts into measurements of amount. Ask what the number represents before choosing a way to analyse it.
Where the evidence comes from
Data can be classified by their origin as well as by their form.
Primary data are collected first-hand for the research being undertaken. If a researcher interviews consenting adults about their own study habits, the answers are primary data for that project. Secondary data were already collected by someone else, or for a different original purpose, and are reused to answer the present question. Reanalysing an existing survey archive uses secondary data even if the new researcher calculates original statistics from it.
These are two independent distinctions. A researcher's own interview transcript is primary qualitative data; their own recall scores are primary quantitative data. A published transcript reused in a new analysis is secondary qualitative data, while an existing table of recall scores is secondary quantitative data. Reading a journal article about a study is not the same as collecting that study's raw data.
Primary collection lets researchers choose suitable questions, measures and recruitment, but takes time and requires consent and data protection. Secondary data can make a larger or otherwise inaccessible dataset available without asking participants again. However, the new researcher did not control its original operational definitions, sampling or missing records. Before reusing it, check that the variables fit the new question and that access and reuse are permitted. Large datasets can still contain bias.
For example, a school wants to understand voluntary library use. Asking consenting students this term produces primary data fitted to the question. Anonymised attendance records from an earlier year are secondary data: they may count visits consistently, but might not record whether a visit was voluntary. The convenient source is useful only for conclusions it can actually support.
Turning accounts into counts
Qualitative material can be coded into numerical categories without losing sight of what those categories mean.
Suppose three fictional adults give the following reasons for using a quiet study room. These are illustrative accounts, not collected research:
| Account | Words supplied |
|---|---|
| A | “It helps me concentrate because nobody talks nearby.” |
| B | “I like having a desk, and I can concentrate there.” |
| C | “There is a desk available when my room is busy.” |
Define two categories before counting: concentration means an explicit mention of concentrating or avoiding conversation; workspace means an explicit mention of a desk. Set the unit as one whole account, and count each category at most once per account. A receives concentration; B receives concentration and workspace; C receives workspace. The quantitative result is two accounts mentioning concentration and two mentioning workspace. Because one account can receive both codes, the total of category counts can exceed the number of accounts.
This is a defensible conversion because the categories, unit and decision rules are explicit. A second coder can apply the same rules independently and disagreements can reveal unclear definitions. Merely counting all occurrences of the word “desk” would answer a different question: repeated words by one person would carry more weight than a single mention. Decide whether the aim concerns people, mentions or passages.
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account