Analysing private hospital websites from a critical perspective: Potential issues of - PowerPoint Presentation

Analysing private hospital websites from a critical perspective: Potential issues of methodology, analysis and interpretation of findings Jonathan Rante Carreon King Mongkut’s University of Technology Thonburi Macquarie University

Background MEDICAL CARE PROFIT MAKING

BackgroundCRITICAL DISCOURSE ANALYSIS (CDA) 3

Purposes and advantages of CDAUseful in examining how events, practices and texts arise and ideologically shaped by power relations and power struggles Useful in exploring opaque language uses that are assumed by writers as part of the background knowledge (BGK) ( Fairclough , 1985) and how the opacity of relations between discourse and society is itself a factor in securing power and hegemony ( Fairclough, 1995; Van Dijk, 2008) 4

Purposes and advantages of CDAHelps in examining how our social practice is bound up with causes and effects that may be not at all clear to us ( Fairclough , 1995; Van Dijk, 2008) It is a flexible theoretical and methodological framework (Chouliaraki & Fairclough, 2010)Contributes to social emancipation (Fairclough, 2003) etc.5

6BUT…

Criticisms of CDA Several occasional (Hammersley, 1997 & Stubbs, 1997) and persistent (Collins & Jones, 2006; Jones, 2004; Jones, 2007; Jones & Collins, 2006; Widdowson , 1995; Widdowson , 1996; Widdowson, 1998) critiques of CDA strongly argue that is filled with the researcher’s political biases. 7

Key methodological issuesData selection (ideologic) Methods of analysis (analyst’s choice; mostly not pilot tested) Interpretation of findings (potentially biased, possibly pre-determined) 8

9SO…

Addressing methodological issues Data selection (Criteria) -representativeness of selections -relative importance of the selections -purposes of the study Awareness of potential source of bias in data selection10

11Sample Private hospitals are torn between two conflicting purposes: providing medical care and profit making. Suppose I want to look at these conflicting priorities of hospitals from a critical perspective by examining their websites, how will I choose my data?

SampleProviding medical care *pages for the different centres and medical services of the hospital Profit making*Investor relations pages and pages for non-medical services 12

BUT… 13

14SO…

Addressing data selection issues -representativeness of selections -relative importance of the selections -purposes of the study 15

16Data Selection Representativeness

17 Data Selection Relative importance of pages R A N D O M

Why homepage?Reception area (portal of entry) (Knox 2007, Nielsen, 2001)Provides an overview of the website’s contents (highlights are brought to the homepage) (Knox, 2007) First and most commonly visited page (Nielsen, 2001)18

19Data Selection Purposes of the study Examples: Analysis of linguistic features Semiotic analysis Analysis of interpersonal/ideational/textual meaningsetc.

20Addressing methodological issues B. Data analysis - Selection of appropriate methods *productivity in a pilot study e.g. in the main project ONLY 4 out of 35 were chosen - Clear and doable guidelines of methods prepared PRIOR to main analysis (pilot-testing)

Selection of productive methodsWhen to exclude/include a method ? *To illustrate how methods are included/excluded Metaphor analysisKeyword analysis 21

22Metaphor Analysis 1. Identify the metaphors from the text *metaphorical uses of lexis can be identified where literal meanings differed from intended meanings e.g. (from BIH website) located in the heart of Bangkok (can be interpreted that the writer is making a medical analogy for the role of BIH in Thailand and thus promoting medical issues)

Metaphor analysis The purpose of using metaphors is to present lexical choices that match the purposes of a writer. However , there is the danger of over-interpretation of findings whereby commonly used metaphors are assigned a specific critical interpretation within a given text even though the choice of metaphor reflects general language use rather than specific purposes of the writer (O'Halloran, 2007). 23

24Metaphor Analysis 2. Identified metaphors were searched for in the British National Corpus (BNC) e.g. located in the heart of *occurred 13 times in BNC *this is a 5-gram (high frequency) **suggests common/general language use

25Metaphor Analysis 3. Decide on the productivity of the method *Metaphor analysis was not productive **Including it in the analysis may potentially predispose the analyst to over-interpret findings and eventually fall into the trap of finding whatever s/he wants to find.

26Keyword Analysis * Problematise the analysis. Decide on the parameters of the analysis and follow these decisions all throughout. 1. Analysis e.g. What is a word? *Are variations in spelling (e.g. American v British English) analysed separately or as one? *On websites, are words in dropdown menus taken as separate words or only the topics considered?

27Keyword Analysis *If proper nouns are taken as upper case words, are they analysed as one name or separated into components? *Are words with the same spelling but different meanings separated or analysed as one? *Are words in pictures counted as words or included as part of the pictures? *Will proper nouns be treated as regular words and analysed as lower case words?

28Keyword Analysis **Different answers to these questions have implications to the findings of any research employing keyword analysis as a method . Word = Any single unit of language in a text functions independently in relation to other words in the same text; part of speech classification of words are not based on their function in a text. Implication = A word is taken at face value and the analyst does not consider the specific function of words in a text.

29Keyword Analysis e.g. international Bumrungrad International international patients both counted as adjective ; falsely elevates the frequency of international as an adjective

Keyword AnalysisBumrungrad [PN] International [PN] International [ADJ] patients[N] PN= Proper Noun ADJ = Adjective N = Noun 30

31Keyword Analysis 2. Interpretation of findings e.g. * Are keywords interpreted as they are or are the co-texts considered? *If co-texts are considered, how many words to the left or right are called co-texts? *Are keywords thematised or analysed as separate entities? *If thematised, what is the basis of thematisation and how is it done? *What is the basis of the interpretations?

32Keyword Analysis *Are absolute or relative frequencies considered in the interpretations? *If relative frequencies are selected, to what extent are these frequencies relative to a benchmark corpus? *What is the role of socio-political context in the analysis? *Is the socio-political context part of the data to be analysed or are they baselines where findings are compared?

33Keyword Analysis Are absolute or relative frequencies considered in the interpretations? Absolute frequency and the centertoBumrungrad Relative frequency center bumrungrad international services surgery

34Keyword Analysis 3. Statistical measures *Compare absolute frequency against a benchmark corpus (e.g. BNC) and compute log-likelihood (LL ) = relativ e frequency **What absolute frequency is considered? *How to identify/choose keywords?*Measures of reliability (iterative thematic analysis of keywords )(e.g. Cohen’s kappa)

Summary: Responds to the critiques1. Data selection 2. Analysis of selected text 3. Interpretations 4. Reliability checks 35

Key pointDecisions are made from pilot testing, creating guidelines for analysis, which govern the analysis all throughout and they are followed irrespective of what findings they produce. This restricts the impact of researcher’s bias in the study being conducted. 36

CDA: Process and ProductProcess: Critique of the methodology (methodology should be transparent and doable) PLUS 2. Product: Critique of the data/discursive practices (classic CDA) 37

Thank you!Q & AJonathan Rante Carreon milojohn23@gmail.com 38