INTERPRETING DATA:
Having spent a great deal of time and effort to collect data – either by performing laboratory experiments; searching and making notes from available literature; or from surveys etc., you still need to reduce the data into a manageable form. Techniques such as tables, figures, flowcharts etc., not only help you to cut down the amount of data being processed, but will also help you form ideas and theories about any relationship between the data.
Don't ignore any relevant literature at this stage. For example, is your data similar to that which others have produced before? Does it agree or disagree with their results? Refer to previous work to support your claims.
As first noted above, it is also important to apply the relevant statistical technique, and this very much depends on the type of data being collected. As a first distinction, we can separate data into quantitative and qualitative.
|
Quantitative data
|
Qualitative data
|
| Based on meanings derived from numbers. Data* may be nominal (categorical), ordinal, interval or ratio. | Meaning is expressed in words. |
| Collection of data is numerical and in standardized form. | Collection of data is non-standardized and uses a variety of formats. |
| Analysis is by the use of tables, diagrams and statistical methods. Again, the methods used depend upon category of data*. | Analysis is via the use of descriptions and identification of concepts. |
*Data types: Click here for a brief glossary of terms
Qualitative Data Analysis
Qualitative research applied to a subject always generates lots of material, which has to be analysed if it is to be of use. This is a time-consuming process and you should not underestimate the time it will take. A practical technique that has found favour is "grounded theory" suggested by Glaser and Strauss (1967). It can be summarised as follows:
- Familiarisation with material: Read and re-read the work looking for patterns, themes, common ideas, attitudes etc.
- Reflection: Ask yourself questions – does this support or challenge existing/my theory? Does it answer unanswered questions?
- Conceptualisation: Are there any patterns or concepts that are emerging from the data.
- Cataloguing concepts: Record the ideas in a catalogue that contains full details of where the same ideas occur.
- Linking: Can all or some of the concepts or ideas be linked together to form an overall 'grounded theory'
[Glaser, B. and Strauss, A. (1967) The Discovery of Grounded Theory, Aldine, Chicago.]
Quantitative Data Analysis
Initially when collecting and preparing data for later analysis, we need to choose between the two approaches of 'descriptive' or 'inferential' statistics.
Descriptive statistics* involves describing and displaying results in the form of tables, graphs, charts etc., and/or using calculations to measure the spread or range of the data.
*Descriptive statistical analysis: Click here for a brief glossary of terms
Inferential statistics* are more mathematically demanding and involve probability and mathematical tests of significance. These techniques are beyond the scope of this CD-ROM; but the bases of some examples are noted here so that you can understand them - in case you wish to include the results of others' who have used these techniques, and which you have found in your literature searches.
*Inferential statistical analysis: Click here for a brief glossary of terms