Content analysis
Traditional approaches to content analysis have often involved attempts to measure
bias in, for instance, election coverage or to study the negative representation
of various ethnic groups as in the racist descriptions of ethnic minorities.
The methods employed usually involved the ‘breaking up’ of texts so that frequency
counts could be made of specific words that were used or where the
amount of newspaper coverage or air-time given to a specific issue could be calculated.
Conclusions could then be drawn about the amount of time given to
different candidates or the number of occasions on which certain pejorative
words were used about one side rather than another. The problem with such an
approach is that when texts are broken up in this way, the context in which
words are used becomes blurred and their actual meaning and the manner in
which they are being employed can be obscured. The ‘meaning’ of the measurements
that are made tends to be added by the investigator. So, for example,
if one candidate receives more media coverage than another in an election, it
might be concluded that this is beneficial. But supposing that coverage included
images of the candidate falling over or mistakes being made when making
speeches, then it becomes less clear which ‘box’ the coverage should be put
into. A much more sophisticated analysis is required of how meanings are
established and how audiences receive and understand them