The autonomy of students

I have reviewed two PhD theses recently. Both have been grounded theories and the lesson that I have learned from them is: Define the research area carefully since under or over defining the research area can severely compromise the quality of the study and either is likely to result in a thin theory and under-conceptualisation. These examples illustrate the issues but bear only a token resemblance to the original studies. 

The larger issue is one of student autonomy.

Defining the research area

This will work: 

Substantive area of interest: traffic systems.
Substantive population: road users.

Consider the traffic system of Birmingham, UK. The traffic mostly ebbs and flows but quite often grinds to a halt for a variety of reasons. The road architecture shapes the way that drivers navigate the traffic system as does the behaviour of other drivers. 

An aerial view such as this one illustrates that drivers exhibit patterns of behaviour as they negotiate motorways, gyratory systems, traffic lights, round-abouts, one way systems etc. etc.

If I were to study road users in Birmingham, I would probably start with car drivers and I might position myself in car parks around the city and seek to interview car drivers about their experiences of driving in Birmingham. I might also search YouTube for videos, look for relevant newspaper articles or blogs and contact car owner associations.

When I had identified a main concern, I would consider theoretically sampling comparison groups in Birmingham:

• Cyclists
• Pedestrians
• Lorry drivers
• Bus drivers
• Taxi drivers
• Motor cyclists. 

I might also theoretically sample car drivers in Paris (who drive on the opposite side of the road to drivers in Birmingham) 

and car drivers in:

• Tokyo (where the road architecture would use different written symbols)
• Italy (whose cultural attitudes to driving are arguably different to those prevailing in the UK)
• Chipping Campden (a small market town in the UK)
• Cleeve Prior (a village in rural UK).

and so forth. I would theoretically sample for as long as I had the resources and for as long as I was collecting useful data for my study.

You might like to read Robert Egan’s study: Precarious entitlement to public space and utility cycling in Dublin as a grounded theory with cyclists as the substantive population. 

Over-defining the research area

This will not work:

Substantive area: cul-de-sacs/dead end streets.
Substantive population: car drivers living in cul-de-sacs/dead end streets.

I most specifically would not study the literature and find that there was a gap in the literature about car drivers using cul-de sacs/dead end streets and choose to limit my:

• substantive area to cul-de-sacs/dead end streets; and my
• substantive population to car drivers living in those streets (because I live in one and have easy access to this population).

And then having interviewed those drivers, I would specifically not make an analytic decision to develop the theory about only those car drivers, driving up the street.

But this is what can happen when an individual from a department/organisation, used to a more traditional research design, decides to use an exploratory research design. Advice offered in all good faith constrains, stultifies and can ultimately kill the project.

The problem is that the data collected about the tiny substantive area (e.g. cul-de-sacs) is insufficient to find dominant and significant patterns. The method is unable to rescue the theory since theoretical sampling in comparison groups is not permitted. The resulting theory is necessarily thin, under-conceptualised.

Under-defining the research area

This will work to a degree

Substantive area: protests
Substantive population: protestors, government officials, local inhabitants, police officers, friends and family.

Data might be collected in the field from 5 different groups of people in different locations and settings over many years. Such data might be collected bravely but indiscriminately. The problem with collecting too much data entirely and with collecting all the study data first before analysing all the data second, is that patterns of behaviour are obscured and the data under-conceptualised. Analysis would aim for ‘complete coverage’ of the substantive area whereas the aim with grounded theory is to collect sufficient data to explain the behaviour in the substantive area from the perspective of one group of participants, and no more. Focusing on breadth comes at the cost of  developing a multivariate and conceptually integrated theory which enables understanding conceptually what the substantive population does and what varies what the substantive population does. 

Such a theory can be made good enough but will miss the opportunity to be fabulous.

Student autonomy

In the first instance the student suffers from a lack of autonomy and in the second, from a surfeit of autonomy. This paper “The Cry for Help!” offers another perspective on seeking an appropriate level of support: