data analysis - types of data

quantitative data

in numerical form, can be put into categories

strengths of quantitative data

- numerical so can easily be put into categories and summarised using graphs and tables
- averages can be calculated
- more objective - not open to interpretation and bias

limitations of quantitative data

- low external validity
- don’t say why - results limited as only provides numerical descriptions
- oversimplifies behaviour and lacks detail
- research is usually carried out in unnatural, artificial environments

qualitative data

not numerical - expressed in words, descriptive

strength of qualitative data

- rich detailed info → insights into thoughts and feelings → why
- increases validity

limitations of qualitative data

- can be subjective as it’s open to interpretation - low reliability, hard to replicate
- more time consuming
- difficult to analyse and generalise findings
- difficult to make comparisons and analyse in tables and graphs

primary data

- data collected directly by the research for the purpose of a specific investigation
- ‘first-hand’ data

strengths of primary data

- designed to extract only the data needed
- info directly relevant to the research aims
- authentic
- research has control over the design and so data is more reliable

limitations of primary data

- requires time and effort - involves planning and prep
- can be expensive

secondary data

- data collected by someone other than the person who is conducting the research
- e.g. taken from journal articles

strengths of secondary data

- inexpensive as the research already exists
- requires minimal effort
- less time consuming

weaknesses of secondary data

- quality and accuracy of data may be poor - either due to being out of date, or incomplete
- this reduces the validity of the data

discrete data

- can only take particular values - can be numerical or categorical
- one category doesn’t blend into the other on a scale
- e.g. nominal data, bar chart

continuous data

- quantitative data on a continuous scale with a 0 value
- it isn’t restricted to defined separate values, it can take on any value
- e.g. interval data, histogram