11.02.2013 Views

PDF (Whole Thesis) - USQ ePrints - University of Southern ...

PDF (Whole Thesis) - USQ ePrints - University of Southern ...

PDF (Whole Thesis) - USQ ePrints - University of Southern ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

creating a glossed-over and incomplete understanding for the reader <strong>of</strong> the discursive<br />

formations <strong>of</strong> the particular text section being analysed. Doing so potentially also ignores the<br />

complexities <strong>of</strong> categories and discourses that emerge from thorough analysis. With a variety<br />

<strong>of</strong> texts selected from each era investigated, this provides less opportunity for separating text<br />

from its context and original purpose.<br />

Bias in reporting data in some ways can be linked with under-analysis. Under-analysing data<br />

can occur by only selecting an isolated segment <strong>of</strong> text, rather than being analysed within its<br />

larger context. This isolated text, then, could be selected as it fits within a particular<br />

perspective a researcher has, or fits within an argument a researcher is trying to make relating<br />

to the data. Therefore, it is important to put measures in place to both avoid taking sides in<br />

the analysis <strong>of</strong> data and to avoid being seen to take sides. This is a difficult aspect <strong>of</strong> analysis<br />

to overcome, because just as no text is neutral, the perspective <strong>of</strong> the researcher analysing the<br />

text can never be neutral either. Conscious <strong>of</strong> this and one way to avoid accusations <strong>of</strong> bias is<br />

for the analyst to be explicit about their perspectives. What is particularly difficult is to avoid<br />

taking sides when discussing issues or events where a particular group has been discriminated<br />

unfairly against (Antaki et al., 2003), partly because it is easy for readers to also have<br />

sympathy for the oppressed group and quality <strong>of</strong> analysis can be overshadowed as a result <strong>of</strong><br />

this sympathy (Widdowson, 1998; Blommaert, 2001). As part <strong>of</strong> a larger criticism <strong>of</strong> CDA,<br />

Widdowson writes,<br />

If you can persuade people by an appeal to moral conscience, you do not need good<br />

arguments. But such persuasion deflects attention from questions <strong>of</strong> validity. It thus<br />

inhibits intellectual enquiry and ultimately undermines its integrity… (1998, p. 150).<br />

This is especially the case, if the discrimination represented in the text being analysed is part<br />

<strong>of</strong> a larger discrimination in society.<br />

Steps taken to avoid this, other than through disclosures, are to ensure that multiple data<br />

sources are used to support claims made in the analysis <strong>of</strong> representations <strong>of</strong> groups. As<br />

explained by Antaki et al., taking sides is a form <strong>of</strong> under-analysis and “…can produce a<br />

flattening <strong>of</strong> discursive complexity, as the analyst selects quotations for the rhetorical effect<br />

<strong>of</strong> appealing to the readers as co-sympathisers or co-scolders. The result is enlistment, not<br />

analysis” (2003, p. 10). This sentiment is aligned with historical research practices, as it<br />

95

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!