Dissertation methodology

Picture by Frank Bohbot taken from ignant.de


I've started the work for the dissertation I'm going to be doing all year. I'm still in the blissful first stages: swimming around a bit and reading very broadly, all I can get my hands on. In the past two weeks I made good acquaintance with an introduction to sociolinguistics by Mesthrie et al. and now I have moved on to a stack of books on the methods and assumptions of critical discourse analysis (less fun to read, not as many cartoons).

The research question as of yet is: which linguistic analysis methods are useful in the study of language policy and planning?

This is different from what I had thought of initially. At first I wanted to research the impact of language policy and planning (LPP) on internal language change. This was my abstract:

Language planning and policy are a well-researched area of study, but the focus is strongly on the political side of things; how can we ascertain that people have the right to use their language for administrative purposes? In which language should education be provided in multilingual countries? The purely linguistic aspect of language policy has been given less attention. The question this dissertation will attempt to answer is what influence language policy has on internal change of the affected languages, thereby tying in to a bigger question of what factors trigger language change.  

In this dissertation I will first make a classification of types of language policies based on actual policy decisions with regards to language planning and regulation. For each type of policy I will investigate how this might affect internal language development and provide examples from languages affected by these policies. The conclusion will reflect on how these findings relate to the bigger debate on language change in general.

The feedback I got for my first proposal suggested that the chapter "Linguistic Analyses in Language Policies" by Ruth Wodak was a good start for such a dissertation. That chapter didn't mention my proposed idea at all though, but instead seemed to suggest that the only feasible method of linguistic analysis in LPP is critical analysis of texts proposing language policies. Hence I now feel that I ought to try this method also.

To criticise my own idea that you can evaluate language policies by analysing the way they impact on the languages they police; it goes against the principle of sign arbitrariness of De Saussure. The language is not better or worse at signifying meaning before or after a change, no matter whether that change be onset by a policy of some kind. However there are some social aspects - e.g. language attitudes and economic benefit - that do get better or worse as language policy changes.

Another assumption underlying my original proposal that I now don't agree with is that you can draw a clear line between linguistics and sociology or politics. There are more and less "social" approaches to studying language and LPP but they blur into each other and overlap very much. Moreover I implicitly made the assumption that this distinction is meaningful and important. It serves to delineate the number of topics that need to be discussed in a degree on any of the subject fields, but that does not mean that it is a good reflection of reality.

What I feel I have to do now is try out a set of different methods and approaches, preferably on and to the same case study. The choice will focus on the more "linguistic" type of investigation as that would be what I can get the most assistance with within the School of English. The purpose to my mind is to get an idea of the relative benefits and helpfulness of each of the methods and the type of information that can be acquired via them. I still want to look at language change and variation, but I also want to critically analyse texts surrounding my case study, and perhaps I can investigate language attitudes as well.

The book I've started reading at the minute is "Methods of Text and Discourse Analysis" by Titscher et al. I found it hard to read at first but then I employed the tried and trusted "line per paragraph" summarising strategy. Then I got into the flow and stopped summarising as it was a bit tedious, but by now (a few hours later) I have forgotten what the bit that I didn't take notes for said. The main point they have made so far is that all research methods are, implicitly or explicitly, related to particular theoretical assumptions. They then went on to set out a four-step research design decision procedure. The first step is to decide what the research is for; finding explanations, testing hypotheses or describing. The researcher then needs to decide how involved they have to be with the source of research data, how critical they want to be of predetermined categories, and whether they want to do a snapshot or track changes. More on this later.

iMc: Persuasion and word elements

Did anyone else notice that two of the most established brands, McDonalds and Apple, both use a recognizable element in their product names? Mc/Mac for the former i.e. McChicken, Big Mac, and a lowercase i for the latter i.e. iPhone. What I'm most impressed about is that they get people to actually call these products by the name they've given them; they're not having breakfast out, they're eating a McMuffin. You don't just have an MP3-player, you have an iPod. (I've even heard people call Samsung mp3-players "iPods" but those people were all old.)

Herstart

Het nieuwe semester is begonnen. Ik had gepland om allerlei vakken waar ik eigenlijk niet voor ingeschreven sta te auditeren en dat plichtmatig ingeroosterd, maar desondanks heb ik me gisteren achter de computer verschanst en hoegenaamd niets volbracht. Vandaag werd ik doodmoe wakker, te moe om lang genoeg rechtop te staan om me aan te kleden, zodat ik besloot om op een ander matras nog wat verder te slapen. Zo heb ik al drie lessen gemist. De scriptiebegeleider heeft (eindelijk) gemaild en laten weten mij volgende week te willen spreken, wat wil zeggen dat ik deze twee weken één officiëel ingeroosterd lesuur heb. Vandaag heb ik me in de regen gewaagd om wat boeken in te leveren - te laat volgens het systeem, omdat mijn betaling voor dit studiejaar nog niet is verwerkt geloof ik. Op de natte trappen en rondom de poorten van het hoofdgebouw zaten groepen nieuwe eerstejaars, duidelijk nog te bevangen door alle nieuwe ervaringen om te merken wat een grauwe dag het was. Ik interpreteerde hun drukke praten als nervositeit: de stiltes vullen in een eerste gesprek met potentiële nieuwe vrienden. In het gebouw waar ik de afgelopen maand gewerkt heb stond een lange rij buitenlandse studenten die de receptie wilden spreken, ordentelijk gehouden door een van mijn voorheen-collega's. Door mijn inactiviteit gisteren kon mijn a4-tje met gewerkte uren, dat ik toen had moeten inleveren, pas volgende maand verwerkt worden. In de bieb ben ik naar de kasten met taalkundeboeken gelopen, op de tweede verdieping, en heb min of meer lukraak interessant ogende boeken uit de kast getrokken. Taalkundeartikelen vind ik meestal makkelijk om te lezen, volgens mij omdat taalkunde nabij mijn normale modus van denken is, maar misschien is dat onzin - misschien zijn ze gemiddeld beter geschreven omdat men in die sector zelfgeselecteerd is op taalgevoel. Ik las een artikel over een praktisch onderzoek in Strassbourg - naar taalkeuze van klanten en verkopers in warenhuizen - en de intro van een artikel over de filosofie achter critical discourse analysis. Iets wat daarin werd gezegd maakte grote indruk, namelijk dat taal niet op zich beschouwd kan worden los van sociale status, maar dat het de sociale realiteit juist bouwt of helpt te bouwen. Daarna was ik echter te moe om door te lezen. Thuis stond ik een tijdje met mijn hoofd tegen de waslijn, maar toen vond John dat ik gek deed en maakte hij citroen-tijmchips en thee voor me.

Omdat ik volgend jaar wellicht in een Duitssprekend land verder wil studeren heb ik me samen met een vriendin ingeschreven voor een Duitse taalclub, waarvan deze avond de eerste les. Tot mijn geruststelling was het precies wat ik had gehoopt, namelijk een soort excuus om zoveel mogelijk Duits te spreken, met een lerares wiens functie niet was om ons oefeningen te laten doen of over grammatica te drammen, maar om met onderwerpen voor gesprekken te komen en desgevraagd een woord te vertalen. Op de vraag waarom ik daar was heb ik enigszins trots laten weten dat ik overwoog om in Luxemburg een Master te gaan doen. Gisteravond zag ik namelijk dat daar een drietalige studie over meertaligheid wordt aangeboden en het lijkt me een bijzonder goede optie. Claire hakkelde meer dan ik maar ik was erg blij dat ze zich ook had ingeschreven, het voelt een stuk minder dreigend tussen de andere deelnemers, die allemaal in "Ruhestand" waren - met pensioen dus. Na het eten heb ik een begin gemaakt in Frisch' Homo Faber, dat mama voor me mee heeft genomen. Het leest het beste met Google Translate ernaast om snel alle onbekende woorden op te kunnen zoeken, hoewel ik had verwacht dat de computer te afleidend zou zijn. Nu ben ik nog wakker met de bedoeling een post te schrijven over dat taalkunde niet aan zijn prescriptiviteit kan ontsnappen, maar het is meer een soort dagboekentree geworden. Wat dan ook.

It's done!

I've handed in the finished report last Friday. Didn't even have to stay up the night before, which is quite a personal achievement. I attribute it to keeping track of the work on here. Check out the conclusions!

Pictures (c) John Hill.

Interview questions

I spent the morning writing my interview questions. John helped to type it up. Look at the nice format:

       

I chose to use a lot of questions from Brady, Verba and Schlozman's article (the one I discussed here). I want to ask the interviewees about their resources and civic skills in order to identify obstacles to participation, and since the questions by Brady et al. (Herafter BVS) were designed to do exactly that it makes sense to copy them; and this way the results can be cross-compared.

There are two major complications with importing BVS' questions. First of all, I am interviewing people in a mental health service, but a lot of the questions are not entirely relevant to their day-to-day life. For instance the question "Last week, were you working full time for pay, working part time for pay, keeping house, going to school, or something else?" can already be answered to some extent for probably all of the interviewees: they visited the service where I will be talking to them. In order to fix dis disparity I have added a couple more specific questions and slightly changed others - to the question in the example I added the option "visited the service".

The other complication is that this survey was designed for people in the United States, but this is Northern Ireland. The problem is not just that I have to edit the questions so that they apply to the local situation, so for instance "secondary school" instead of "high school". More importantly, because of the political situation in Northern Ireland people may feel scrupulous or hesitant to answer certain questions such as: what language do you speak at home? Are you a UK citizen? What is your religious denomination? I don't want to ask such questions unless strictly necessary. I added a disclaimer to let the interviewee know the research is not about the sectarian division and they do not have to tell me what side they are on.

BVS designed their questions for a large-scale quantitative interview. I am going to interview too few people to meaningfully compare my results, but still their results could serve as a way of placing mine in context. For instance, if a participant turns out to have very little time, we can refer to BVS who found that people with little time generally do not participate as much. The ease of integration of their questions in my interviews shows that a question is not inherently qualitative or quantitative; it depends on the number of research participants and on how the acquired information is processed.

P.S. I updated my consent form so that it is relevant to the actual questions. Here

Suggested search gives an idea of people's image of mental illness.

Research (now with vulnerable adults)

My practical research (as opposed to the literature review) consists of two major parts, one of which I've already completed and the other I haven't made an appoitnment for yet. I'm going to have to hurry as I've a month left to get EVERYTHING done so interviewing, writing, editing, cleaning up the references, getting the whole thing printed and ring-bound somehow.


Research design

Part one is an experiment, to do with the All-Party Group I described in the previous section. For one of the meetings we invited a service user to come along to give a talk about her experiences with the particular piece of legislation that the meeting was about. After her talk I handed out surveys for present MLAs and MLA representatives to fill out, about what they thought of the talk and if they had had contact with service users before, that sort of thing. Beside the surveys the oral comments from those present, my own observation, and importantly the conversations I had with the person who gave the speech were very useful sources of information. Conclusions from the experiment relate to how a service user can be accommodated to give such a talk and how it can have the most impact on the decision-makers it is meant to influence. Here is the questionnaire I used.

Part two of the practical research, the part that I haven't done yet, would be interviews with service users, nurses and officials about their experiences with politics and the obstacles and barriers to their participation. The questions I had for this are actually, in hindsight, pretty bad; I'm basically just asking them what programme they think should be ran. That's alright as a side question but not as a main question I don't think. Having done the research that I've done now I think there should be:

1. Profiling questions about their type of illness and what class they would say they were in; middle, working, upper?
2. Identifying barriers or obstacles to participation
1. A considerable section on civic skills -- maybe I could ask some of Brady, Verba & Schlozman's original questions. This is to test my thesis that this is one of the main obstacles. Includes questions about education, work experience and confidence.
2. Questions about their opinions of MLAs and of Stormont to see if they think politics is really impenetrable, plus a couple of knowledge questions about Politics to see if they need informed about politics, as this is one of the Rethink Politics programme's main pivoting points (besides skill building.)
3. Searching solutions: would they participate in 
1. a programme to help them build civic skills 
2. a committee or board 
3. a user-run service... and which do they think would help (and why?) (and which programme do they think should be ran, if any)

Research ethics

As much as I like this whole blogging thing I won't be able to post any of the evidence or the results on here though. That's part university etiquette and part common decency, succinctly summarised as research ethics. When talking about research ethics, people suffering from mental illness are generally classed as vulnerable adults, along with people with Down's syndrome and learning disabilities. The main ethical priorities when doing research with this "group" are informed consent, privacy and confidentiality. I will now explain what these mean in general and for my research in particular, among other issues.

Informed consent means the participants in the research know what the research is about and what the risks and benefits for them could be. With people with learning disabilities the risk is that the researcher might make them sign a form to say they have read and understood the risks of the research, when in fact they have no idea what's going on. One time as a kid I was playing with marbles with a girl my age who had Down's syndrome. I asked did she want to play "keepsies", i.e. whoever won got to keep the marbles, and she said yes; but then when I won and wanted to keep her marbles, a classmate told me off. His argument was that whatever I'd ask her, she would probably say yes despite not understanding. Here we could say there was no informed consent despite her "signing my form". I don't think this scenario applies to mental health service users as much as it applies to people with Down's syndrome. In fact the whole point of my research is that mental health service users may have lost control over some parts of their lives, but this does not mean that they are unable to make sensible decisions about any part of their life. (What do the experts say?) My consent forms for the interviews: 1, 2

Another problem of informed consent is that in some researches it is crucial that the participant does not know exactly why they are being asked to do something, as it could influence the way that they do it. This is not a problem in my research; for the interviews I will tell the participants in advance what my research is about and why I'm asking certain questions, and for the experiment I went to great lengths to make sure that the speecher didn't mind me handing out questionnaires and I had her read the questionnaires and told her I could edit them if she wanted me to.

My main concern was that she'd feel like she was being judged for the speech - which would be understandable as the questionnaires are basically asking MLAs "what did you think of this speech". I think this demonstrates the tension between getting research results and taking care of your participants. If I'd made her feel too bad about herself to give the speech then I couldn't have gone on with the experiment despite how many positive implications it might have had for mental healthcare in general and for my own academic credentials. This is now drifiting away from informed consent in the direction of psychological harm, which is something that you want to avoid as a researcher. I think psychological harm is a bigger risk when doing research with mentally ill people than lack of informed consent is, as service users may already lack self-esteem. If they are portrayed negatively in the research report or if they get the idea they did something "wrong" e.g. they gave the wrong answer to a knowledge question, this might further undermine their self-esteem.

Mental health service users also have a lot to win or lose when it comes to forced treatment and imprisonment and rights to self-determination. This was also noted by Barnes et al. while they were trying to find user-run groups to participate in their research:

Users of services have objected to being subjected to the professional gaze of the researchers as they are also objecting to being constructed as passive recipients of welfare services. In some instances they have refused to participate in researach projects. ... Considerable discussion was required before one particular group finally agreed to be involved. ... Researchers working in this area have to develop new ways of working with users as participants rather than subjects of the research process. (p.10-13)

Privacy and confidentiality together consist of the considerations: what can I ask, what can I write about and what do I do with the data afterwards? It's a bit like with the NSA. People may not be comfortable with their data being published all over the world, especially if it could still be connected to them in specific, and especially if it concerns incriminating, embarassing or very personal information such as experiences with drug use, memories of having been bullied or how long a person has been out of work for. Data can be anonymised, but the more detailed and specific the research, the more risk there is that participants can be identified by their remarks anyway. If there is one case study or one small group of interviewees, like in my case study, the risk of identification is quite high. Likewise it is higher in a qualitative study where participants are quoted at length than in a quantitative study where they represent a dot in a scatter plot.

To an extent risk of identification via quotes is inevitable especially if the point of the research is to ask people for their specific experiences, but the researcher should use their own good judgment in what information to quote and what to leave out. Specific placenames shouldn't be mentioned and names can be changed or left out of the report. In my particular case, I will be asking people for their work experience, but I'll avoid talking about that in any recognisable way. Once again, Barnes et al. came across a similar problem in their research:

Interviewees in one group had found it uncomfortable to have their words reflected back to them in the case study report. They were concerned that other group members would be able to recognise who had said what, even though individuals had not been identified. (p.14)

After I finish the research I'm going to have to destroy the records and transcripts of the interviews. It's a form of good practice even if there is no realistic risk that someone would find them and hound the interviewee for its contents. Really there shouldn't be anything that I could possibly still need it for and if I would use it in another way than for the reports, that would be something that the interviewees did not give me explicit permission for.


In conclusion, the risk that mental health service users would not be able to actually consent to the research is not very big and there is no deception involved. Real risks to studies in this area are negative portrayal of participants, such as making them seem like passive recipients of mental healthcare; causing psychological damage with unwarranted remarks or interview questions, and putting identifiable details from the interviews in the final reports. For each of these risks there are few absolute solutions other than the researcher being aware of this, using their good judgment and taking advice from others, including the participants themselves.