The analysis in the present study is divided into four sections according to what features of a
communicative purpose and the formal features of structure and language. The data for the analysis of discourse communities, communicative purpose and structural features was obtained through content analysis of the blogs as a whole, including boilerplate texts and comment sections. The linguistic features were analysed through a statistical corpus study of the text bodies of each blog entry.
In order to situate the genre of academic research blogs in a discourse community, I analysed the demographic characteristics of the blog authors. These include the sex, country, title and field of research of the author, as well as the number of authors in each blog and whether they use their real name or a pseudonym. In addition, audience and reader characteristics were determined. These characteristics were collected by analysing the blog entries according to what kind of audience the author has intended the texts for, and looking at the comments section of the entries to see the characteristics of the readers. Also analysed were the amount and form of interaction between the blog author and the audience. This was done by determining if the authors refer to their audience in the blog entries and if there is any kind of dialogue between the author and the audience in the comments sections of the blogs.
Next, the overall communicative purposes of the blogs were coded. Each blog was assigned to one of the following four categories of communicative purposes according to what categories were most strongly represented in the blog: invisible college (i. e. blogs written for fellow academics in the field on field-related topics), diary about academic life (i.e.
blogs about the academic life and work of the researcher), popularising science (i.e. blogs on field-related topics written with lay people in mind) and soap box (i.e. blogs where the academic blogger shares their opinions on controversial, field-related topics. These categories were synthesised from earlier similar studies of Herring et al. (2004), Halavais (2006), Walker (2006) and Kjellberg (2010). Below is an example of the analysis of the communicative purposes of the blogs.
Table 3: Example of the analysis of communicative purposes
Purpose Respectful
Insolence
Profgrrrrl Aetiology All That
Matters
Invisible college X
Diary About Academic Life X
Popularising Science X
Soap Box X
The structural analysis of the blogs was carried out by selecting appropriate structural features to be studied and counting their occurrence in the blogs. The selection of the features was based on earlier studies on Web genres (e.g. Herring et al. 2004). These features include entry length, images, videos, advertisements, software used, comments section, about page, archives and links. In addition to counting the number of links in each blog, the links were divided into four categories according to what kind of resource they were linking to. The link categories were formed based on the analysis of the most common functions of the links in the data. These categories are reference, self-reference, explanation and entertainment.
Temporal features of the blogs were also measured, such as the frequency of posting and the age of the blog. All the analysed structural features are listed below:
About page
The final stage of the genre analysis of academic research blogs was the linguistic analysis. The linguistic analysis was carried out on the lexico-grammatical level, and the analysis was limited as the words in the corpus were not tagged according to the parts of speech they represented.
Therefore, the linguistic data could be collected with simple commands using standard GNU/Linux corpus tools. For example, first person pronouns were quantified with the following command line:
cat allblogs.txt | tr '[:upper:]' '[:lower:]' |
egrep –wio 'I|we|me|us|my|our|mine|ours|myself|ourselves'|
sort -nr | wc –l
For some features, simple calculations of word frequencies were not sufficient, and they had to be analysed in their context. For example, when counting the frequency of the emphatic do + verb one had to look at the context where the word do appeared to determine whether it was followed by a verb to see if it was used as an emphatic. If one would only count the instances of the word do one would get wrong results, as other constructions, like questions,
their context to see if it was followed by a verb. In order to see the words in their context, a following command was formed:
cat allblogs.txt | tr '\n' ' ' |
egrep -io '[[:print:]]{30}do [[:print:]]{30}'
This command produced a concordance list of the word do with 30 characters before and after it. Below is an example of the concordance list formed by the command:
a private activity. But if we do decide to share our data, how no reason why I would want to do that as an individual, but for tty much exhausted what I can do that as an individual, but for a little bit longer, though I do cover a few of the topics I ta ts that followed, I decided to do a thought experiment. Imagine the brain. To begin with, how do we expect our best games to
psychophysiological response. Do not change too many variables they might want to stay on to do a Masters programme or seek
The feature frequencies were then normalised and analysed and compared to the results of earlier studies.
4 RESULTS AND ANALYSIS
In this section, I will present the results of this study. The findings of the research are shown in quantitative summaries, and these summaries are followed by more in-depth analysis and interpretation of the results. I will begin this section with presenting the results of the analysis of the discourse communities that use and produce academic research blogs. In the second section of this chapter I will look at what the results tell about the communicative purposes of the blogs. The final two sections are dedicated to the formal features of the blogs, that is, their structural and linguistic characteristics of academic research blogs.