Categorising the data allows us to compare observation in terms of relations of similarity and difference. To capture this information we need to link data as well as categorise it. Linking data involves reorganising substantive rather than formal relations between things. Formal relations are concerned with how things relate in terms of similarity and difference
- how far they do or do not share the same characteristics.

Substantive relations are concerned with how things interact (dentists-patients: not similar but interacting).
[Hyperlinks and relational databases are discussed in class.]

Like categorising, linking has a conceptual as well as a mechanical aspect. Potentially there are as many links as there are transitive verbs in the English language, connecting a subject and an object.
[ x - y ; f(x) = y ]

Linking and categorising complement each other:

     link
 
 
 
 
 
 
 
 

Issue related with linking data:

- Links must be labelled,
- Use a link list for clarity and consistency,
- Use a limited links list to reduce complexity,
- Ground links conceptually and emprically.

Our identification of links is itself influenced by the regularity with which events are associated.

Associate Events:
 
 
 

Linking Events:
 
 
 
 
 

The regular association of events provides a basis for inferring possible connections between them, but subject to conceptual confirmation through establishing some links or connecting mechanisms, which operate between them.

The conceptual identification of links provides a basis for identifying connections between events, but subject to empirical confirmation through regular association between them. Mostly establishing connections depends on elements.

The mechanical aspect refers to hyperlinks created between different databits. The conceptual aspect refers to the identification of the nature of the link between the databits - e.g. is it causal, explanatory or whatever.

Text is a useful vehicle for presenting information, but often pictures may correspond more closely to how we actually think. We can use matrices to compare information across cases. Matrices are used to complement the analytic work done through categorising the data, as well as to comprehend the results of categorisation. But if we insist on preserving full detail, our matrix is quickly overloaded with data, so we prefer to summarise the data. We can make comparisons across rows and columns and gain a better overall sense of the data.
[Evaluation of IE 404 quiz answers given as example]

Unlike matrices, maps do not conform to any particular format. They are particularly useful in analysing the connections between the categories. They are useful representations of reality. Map of relationships between concepts:
 
 
 
 

    Relationships
 
 
 
 
 
 

We can devise all sorts of mapping data; we are limited only by our imagination.
[Class example and debate about "The Little Prince", written by Antoine de Saint-Exupery]

Shapes, lines and patterns can represent different types of category.
[Economic wealth distribution after geographical locations, map legends, flow charts, pie charts, organisation charts,etc.]

The procedure for analysing categories connected through linked data puts the conceptual appraisal of connecting mechanisms first, and then requires some assessment of the regularity with which these occur in the data.
RISK: impressionistic and unsystematic procedures characteristic of
more traditional analysis.
SOLUTION: Evidence of linked data.

Evidence: What happened? What was said? What was meant?

Potential abuses: Fabricating evidence, discounting evidence,
misinterpreting evidence.

Measures:
- Enumerate the amount of data,
- Evaluate the quality of evidence,
- Assess the conceptual significance of the data,
- Look for exceptions, extreme or negative examples,
- Follow different pathways through the data.

Qualitative analysis is essentially an iterative process, involving repeated returns to earlier phases of the analysis.

Choosing between rival explanations:
 Which explanation is
- simpler?
- more credible?
- more internally coherent?
- has greater conceptual import?
- has the more acceptable practical import?

Back to home