Dissertation nearly there

I've been checking my work against the marking scheme. I reckon it passes, but I can't tell the quality of the pass; I don't think it's a distinction. How can you assess yourself against clearness? What is clear to me has not been clear to my supervisors. The marking scheme for a band that is 84-70% wants

"clear articulation of research questions that are derived from a critical review of the literature."

I can find problems with that - what's "clear"? What's "critical"? and have my questions come from the literature? Do I show that they have, or do they seem to come from the top of my head!

Restructuring

I've reordered material from the draft dissertation so that the analysis covers roles and application of agency theory to the data. That takes out loads of material from the chapter that has to be called interpretation. What's the difference between analysis and interpretation?

I think the analysis bits are where I've applied ideas from Miles and Huberman, used codes, rich picture (Checkland) and drawn up a role-ordered matrix.

That's quite helpful because it shows where the roles match Schein's concepts of client types. Then I put that analysis on to my agency theory diagram so I can see the chain of links, and back stitch links between various accountors.

The interpretation bit needs a bit more thought. I've linked it from the analysis by occasionally picking up ideas in the analysis and suggesting they are new to this case. For example, the analysis identified incidents of accountability despite that participant saying he/she couldn't think of one or didn't understand the question, so I thought that was unconscious acting of accountability and something to follow up in the interpretation.

Funny how writing helps you think. It seemed to me that that unconscious acting was a reaction to a situation, so the participant demonstrated reactive accountability, and that meant I realised that the other planned accountability was proactive. And that leads me to my last, not yet properly written chapter on findings.

Relooking at the literature

I've gone back over the literature review chapter- it's quite depressing because now I think I don't really have incisive research questions, and am not clear how I get any questions from the review at all. If I haven't set a question, then what is my research answering.

I've rewritten the section on accountability again. And again. My supervisors want a definition, but I'm not sure that a definition is good enough, because that makes it sound like you can get a grip on accountability but it's too nebulous to grip.

I've talked about Sinclair's discourses of accountability, and Bovens analysis of accountability in the public sector and Siegel-Jacobs and Yates research comparing outcome and procedural accountability seems important. Agency theory matters and I've drawn diagrams to show how I'm applying it to multiple accountors and accountees.





And that moves me on nicely to discussing consultants, clients and relationships.

---

Bovens, M. 2005 Public Accountability in The Oxford Handbook of Public Management, draft, L. Lynne C., Pollitt (eds.), Oxford, Oxford University Press
Siegel-Jacobs, K and Yates, J.F., (1996). "Effects of Procedural and Outcome Accountability on Judgment Quality" Organizational behavior and human decision processes 65 (1): 1–17.
Sinclair, A., 1995, The Chameleon of Accountability: forms and discourses, Accounting, Organizations and society 20(2/3.): 219-237

What do you do when..?

Suppose the consultant advises something, and it's bad advice. Who's responsible? The consultant for giving bad advice? Well, he won't get more work from that company. But it's not a company, it's a public service and it's public money that's been spent to get that bad advice. So why did someone in the service get bad advice? Through bad management. There are ways of preventing the situation arising:

• Write a brief,
  
• Get the contract right.
  
• Require progress reports
  
• Withhold payment.

Look at briefs. This is what the consultant said:

“Because you need to be very clear what does the client want actually get out of this, because you’ll never be able to give them the right recommendations if you have a misunderstanding or a misinterpretation of what their goals are.”

Get the contract right. Look at contracts:

"So there are contractual agreements between the council and the consultancy and there are certain agreed outputs."

Progress reports warn the managers before it goes wrong.

And if it still goes wrong, look at withholding payment:

“I can think of one, where one of our IT contracts to implement an upgrade to a social services system where the consultants failed to deliver what we wanted, and we got into a contractual negotiation where we withheld a considerable sum of payment.”

If the project gets to the stage where there’s a contractual negotiation and withholding of money, then the process has gone a long way down the wrong road. But who is ultimately responsible?
Some say one thing:

“The person that decides that they should be brought in and sets the brief”

Others look at something wider:

“I think the county needs to be more transparent to its council taxpayers, in terms of the use that is made of consultants.”

Some say the public elected leader of the council.

So it looks as if there is some sort of process to prevent and to deal with bad consultancy advice, and for the lines of public accountability that must answer for that poor advice.

The story

The case study story is about a small project in an English county council where consultants were brought in to develop an approach to a county council service.

For those who don't know, county councils are run by the public servants and the elected members or councils represent the electorate. County councils provide services such as

• education including schools and art galleries,


• care services such as adoption, social services and may be care homes,


• development services such as waste disposal,


• community services such as parks or libraries,


• resources such as estate management, encampments, transport and roads.

The basic brief was to consider the future of the service but was driven by budget problems.

The consultants visited all the branches of the service, listened to the service assistants and then produced an internal report. The operations managers incorporated the consultants’ advice into a long technical document that became the county council’s report. Some components of the report had repercussions, for instance one component advised closing eight branches, which would save over £200,000.

The elected members took the review to public consultation. People objected to losing the service and they found a totally different approach to the way those branches were run and consequently kept five of them.

The consultants’ recommendations were not totally implemented; the political process changed the outcome.

Last supervision meeting

Having spent an intense hour discussing my draft dissertation with my supervisors. As always, they've read my ramblings and courteously and logically point out the errors and lacunas of my writings. But this is the last formal submission so is the last time they get to comment on my work and my last chance for advice. I've got six weeks to ruin it by myself. It has to be handed in 10th September. The supervisors don't mark it, but some third party who knows nothing about it.

However, S#1 volunteered to take a look at it when he gets back from holiday, and has even arranged to meet to discuss it, just one week before it is due in. :) S#2 might have time to comment on an electronic copy, but is back from a longer holiday even later than S#1.

I've allowed husband to book us a trip away after submitting the dissertation, but then today they started talking about the viva. Viva! When? Talk to people about it? Oh dear, not another interview. But it's only with two people not a whole board of them and I suppose I might manage.

Asymmetry of information

Agency theory talks about asymmetry of information. The principal and agent have different information and different information needs. That's relevant to explaining a perspective of this project. One person saw that it 'petered out' rather than coming to a clear conclusion. At the final stage, operation managers were able to continue to call the consultants for advice, who therefore billed for extra days. If the principal /Operations Manager perceived a need for information & lacked confidence in own ability to carry on and implement the changes, but perceived the consultant / agent had that information, then contact was continued, and the billing.

Keil, Mann and Rai discuss this escalation in their paper on why software projects escalate [1]; I think it is irrelevant whether or not the project is a software one. There seems to be a lot of research on software projects and at least some of it must be transferable to non-software projects. Keil et al's quantitative study found that two constructs associated with agency theory, were associated with projects that could be predicted to escalate. The concepts are:

• goal incongruency


• information asymmetry

Goal incongruency is where there is potential conflict between the principal and the agent, such as the agent acting for his own best interests rather than that of the principal.

Information asymmetry is where
"the agent is assumed to have private information to which the principal cannot costlessly gain access." Baiman[2]

Implications for good practice were the importance of good communication and monitoring of projects to avoid information asymmetry. They also suggested management should implement early warning systems to detect escalation.

Good communication is something the participants have mentioned in my case study – they require transparency. A combination of monitoring mechanisms and good communication seem to have made this project successful from the perspective of most of the participants.

---

[1]Mark Keil; Joan Mann; Arun Rai, Why Software Projects Escalate: An Empirical Analysis and Test of Four Theoretical Models, MIS Quarterly, Vol. 24, No. 4. (Dec., 2000), pp. 631-664. http://links.jstor.org/sici?sici=0276-7783%28200012%2924%3A4%3C631%3AWSPEAE%3E2.0.CO%3B2-2
[2]Stanley Baiman, AGENCY RESEARCH IN MANAGERIAL ACCOUNTING: A SECOND LOOK, Accounting Organizations and Society Vol 15.No 4 p.p . 341-371.1 990.

Finding gaps

Gaps so far:

1. mechanisms of accountability missing towards the end of a project
2. potential lack of brief & timescales
3. lines of accountability that might involve an intermediate client
4. lines of accountability missed in a complex project than involves more than one service, (not in this project). Where a project involves two services then someone has to realise and communicate its existence to both of the relevant elected members.

Points of view

I have five different people, all talking about the same project.

They seem to have overlaps of their perspectives of accountability, but also there are two different views. One is absolutely sure accountability existed all the way through this project and can immediately indicate lines and mechanisms of accountability. Another sees gaps. How do I reconcile these two views? I suspect that they are both right; that there was accountability but that accountability was only clear for the first phases of the project and not in later parts.

Agency theory suggests that information asymmetry between agent and principal means the principal wants information that the agent possesses. So if there is a public-sector manager who doesn't yet feel confident that she/ he knows what he /she wants to know, then the temptation to prolong the contract must exist.

Interviewing

I have this wonderful digital audio recorder that the department has lent me to do my interviewing. It's an Olympus DM-10 and apparently costs a small fortune. At Amazon, it's nearly £200, no, £300. But it records well over the speaker phone - I haven't got to meet any of my interviewees yet. It's small and fits into my hand and it's unobtrusive. I like it so much that I might get myself one. I suppose I could justify it if I am going on to do the doctorate and I do interviews again. And it would also be useful for one of my hobbies- recording what my older relatives say about family history. I'm convincing myself...

Don't shoot the advisor

In the August copy of Management Today, Richard Reeves wrote this. He argues that it's up to the public-sector officials to make sure consultants' advice is of high quality. There's a balance needed between the cost of permanently hiring or training specialists that the service might need for only a few months, and bringing in the external expertise, but consultancies are profit making businesses, and must make money to survive. I like Reeves metaphor:
"If the public sector is handing out sweets, they can hardly be blamed for the resulting sugar high."
However, he does also accuse civil servants and politicians of deifying those in the private sector, and I don't know what evidence he has for that. It is not what I'm hearing from my interviewees. What I am hearing is that consultants earn their money by:

• providing expertise that is not in-house,
  
• moving organisations that wouldn't otherwise be moved and
  
• saving the organisation more than they cost.

And if the lines and mechanisms of accountability are in place, then there won't be many undeserved sweets.

Draft in

The draft is in, and I await the verdict. Nevertheless, I know there were more additions and refinements, not least the extra perspective from the interview due on Friday. But there was also the feeling that it wouldn't matter if I didn't use quite the right language, or if I experimented with diagrams, which I have done when examining the number of relationships.

Discussion for supervisors

I have a couple of issues to discuss with supervisors:
1. ways of analysing and interpreting
2. anonymity of design

Tomorrow the draft dissertation must be in. The new chapters are:
Chapter 4 - data collection and analysis
Chapter 5 - interpreting the data
Chapter 6 - findings

I have chapter 4 and some attempt at chapter 5, but would like to add more to it. I've structured it round answering the research questions, but am also thinking of an alternative structure that works round the concepts of power, autonomy and collectivity that came out of the analysis.

That is something to discuss with my supervisors.

The second issue to discuss is the one of anonymity. I now realise that having just one case study, and a small case study, although I can anonymise the source, because I'm looking at relationships, the participants can recognise each other. That makes it difficult to use all the data. For instance, if one participant, as a consequence of his or her relationship, perceives something that might be a critique from the perspective of another participant, then, there could be repercussions in the organisation if they read what I write. So I have limited my publishable data. Interesting, and something that I just hadn't worked out before the research, or hadn't worked out far enough. That was partly lack of experience, partly lack of forethought, but also I couldn't have told how many participants I was going have until I found the project. Obviously, there would be only one CEO, but I might have expected more directors, and more managers and more assistants, in which case there would have been anonymity in numbers.

An alternative design would be to have had a number of cases either within this organisation or in more than one organisation. The unit of analysis would still be the project and the relationships in it.

Supportive family

Isn't it wonderful what questions your children can ask? At lunch table my 17 year old asked me how I was analysing the data and I was explaining that I'd transcribed the interviews and then had devised some codes to put against chunks of speech, but that when I came to apply them they didn't work so I created other codes of climate, process and structure and subdivided those when I came across ideas that reminded me of something in the literature, like conflict suggesting political elements. I didn't know I could verbalise what I was doing.

Then the 17-Y-O, who is a mathematician was saying something about "but that's only opinion, not facts" and 19-year-old scientist got into debate of researching opinions and facts.

I'm so glad they're interested. They are going to have three more years of me doing this, and someone has to read my work before I give it to my supervisors.

Analysing vs interpreting

The instructions say that chapter 4 should be about analysing the data, and chapter 5 interpreting the data. Well, some might say that those words are a bit similar in meaning, so I have to work out where they separate. I'm taking the analysis to be a description in the light of the codes that I've applied. The interpretation is more about explaining what the data reveals, and I'm structuring that interpretation against the roles and what they suggest in answer to the research questions.

Nerves

Our Director of Research Studies tells me that I did really badly at the PhD interview and he had to dig me out of it. I'm not surprised because I have often have done badly at interviews despite having an excellent CV, appropriate for the position and a suitable covering letter written well enough to get on the short list. So what am I doing wrong?

DoRS suggests I'm answering too quickly without first working out what the question is getting at. He gave some suggestions:

1. take a few deep breaths - well I used to do that when gliding and ended up hyperventilating at the top of a launch and about to faint, so no, that won't work for me.
   
2. ask the questioner to repeat the question
   
3. write it down. I like that suggestion, so I was thinking about, when I had gone there prepared mentally to write down questions, why I didn't. It was because there was so much stuff in front of me, the laptop, my notes and a mug of water, that I would have had to move, but I couldn't move them because I would have knocked over the water.

Which brings me to a physical problem - why would I have knocked over the water? Because my hands shake when I'm nervous. I probably couldn't have read my handwriting if I had written the questions, because my hands would have shaken too much! My mother has trembling hands, only noticeable when she pours the tea, and tells me it is a familial tremor. Well, it hits me when I have emotional stress. In order to take up the DoRS's advice, I'm first going to have to deal with the tremor, so no caffeine, nor ginseng which also makes me shake. I do like the comment that apparently alcohol reduces the tremor, but at interview perhaps no glass of wine.

The DoRS's fourth suggestion was that I practise over the next year, that my supervisors get me to offer seminars on my research. That way I'll reduce the stress through familiarity with the process of discussing and defending it.

Writing the draft dissertation

The draft dissertation is due 1st August. I have heaps of issues to discuss in it:

• the scapegoating or buck-passing that Harmon mentions, in a public environment where risk taking is avoided, and perhaps I have an example that manifests the avoidance or the fear of blame. I must look for alternative interpretations of the data.
• reflect on limitations, what I would do differently and what I would do if I had more time
• write on the unit of analysis for this project as sets of relationships around any one issue of accountability and multiple cases of relationships - I'm not sure what I'm trying to say here.
• the psychological contract might be another perspective to emphasis
• the principal-agent relationship is informal between the parties
• reflect on the difficulties of anonymising in a single case study
• debate the term transparency - what did it mean to those participants who used it, but I didn't follow up.
• delimitation on the consulting project - when was the work completed (Werr, 2002:63)

Analysing

I thought I might analyse the discourse from interviews against Werr et al's [1] two types of discourse: bureaucratic and network. I started out with a table:

Role	Management of Consultants	Example of discourse
CEO	talked about process & strategy & outcomes	"It's the funding which is something usually discussed with the individual and the cabinet member" and "there are contractual agreements"

Then I checked Miles and Huberman [2] had a role-ordered matrix, which in 15 minutes gave me a useful one page table of roles in the job, as primary,ultimate, intermediate clients. I've still to work out anything useful about recognising the types of discourse though.

---

[1]Werr, A, Styhre, A. 2002, Management Consultants – Friend or Foe? Understanding the ambiguous client-consultant relationship, International studies of Management and Organization, vol. 32, no. 4, winter 2002-3 pp43-66
[2] Miles, Huberman, 1994. Qualitative Data Analysis: an expanded source book, 2nd edition, Sage

Doing the Masters in Research

If a reader is interested in following this full time course, there's a document at
http://intranet.open.ac.uk/studentservices/publications/documents/research-degrees/MRes.doc
that gives the details.

Ingredients collected

It seems as if I have a set of ingredients, the data I've collected from interviews, and a recipe book, the Miles & Huberman [1], but how I put them together is up to me, both in what I choose to do with the ingredients and how I mix them, so I come out with, not definitive results, but what I want to present, or rather, not necessarily what I want to present but what I'm able to present, depending on how I analyse and write.

I remember when I was little how I would have paints and paper, and want to represent a wonderful picture through my painting, but my paint would be too watery and I'd produce something horrid, not what I meant at all. I was disappointed because I couldn't show and share with other people what was in my mind.

So I must hone skills to use the tools and to communicate what I find.

Son's just been reading something from this week's New Scientist[2] on finding a mathematical proof after 7 years. Once a useful alternative view of the problem revealed a solution, it took days, hours to compress the proof to a few lines. Son said that he too could see this at his lower level of maths, where having realised the answer he wanted, he could say, "irrelevant, irrelevant, irrelevant" and strike out much, leaving the last succinct and correct two lines.

That's what I want to do too with this analysis - must be my mathematical background, but it's going through all the irrelevancies first before finding the points that you want and need.

---

[1] Miles, Huberman, 1994. Qualitative Data Analysis: an expanded source book, 2nd edition, Sage
[2] New Scientist, Proof and Beauty, page 48, 21 July 2007