Study Design and Methodology
This study takes a two pronged approach by drawing upon data from a previous survey on Facebook and blending it with new specific content analysis. This combinational approach was selected in part because of its convenience but also to help establish both a broad and focused understanding of the events that have transpired on Facebook. Unfortunately time limitations prevented in-depth interviews or true digital ethnography work so the findings are constrained to be mostly surface-level and quantitative-feeling in nature.
The Facebook Project 2007 Survey
The second in a series of yearly surveys on Facebook, the survey used for this paper was sent out over the summer of 2007 to a formal randomly selected portion of the undergraduate student population. All respondents were full-time degree-seeking students over the age of 18. The decision was made to exclude part-time and non-degree seeking students after it was determined they were statistically more likely to be of a significantly older age and only comprise a minimal, outlier population at UIUC. Students under the age of 18 could not be included for ethical reasons. In total the official university statistics department, the Division of Management Information1, pulled an 1100 person sample randomly from the entire undergraduate student population. The response rate to this survey was very poor due to a survey response limit mistake2 as well as the sheer length of the survey and technological limitations that prevented collection of partial or specific responses.3 All told only 75 students (a pitiful 7%) fully completed the survey, which effectively means the data is not generalizable to the overall student population to a statistically significant degree. Despite these shortcomings the survey, when paired together with content analysis, presents a number of interesting findings about the student populace that can be easily confirmed via more in depth qualitative study. It is best to consider it a sort of scout work to inform future investigations and inquiries.
This study employs a rather exploratory method of content analysis, being that the researcher is entirely new to the method in general. Pertinent data was organized and reduced to uncover patterns of human activity, action and meaning through both simple conceptual and introductory discourse analysis. Though deductive in nature the methods were not purely social-anthropological, as the researcher does command a certain native perspective but cannot claim to extensive specific ethnographic experience. The analytic task as always, however, remains to both identify and explain the ways people perform in the Facebook setting; how they’ve come to understand things, account for, take action, and generally relate to the pro-Chief Facebook cause in their day-to-day life. In order to avoid misunderstandings and ensure reliability only two small and particularly volatile Facebook groups, “Don’t Like the Chief? Go Somewhere Else… fuckin Idiots!” and “Pro-Chief People Wouldn’t Know Racism if it Bit Them on the A$$” were closely coded and studied, but in total 17 Chief-related groups were overviewed and scouted out. The validity of the text observed from these groups is obviously bias but aptly demonstrates the emotional fury surrounding the issue amongst students. This study does not seek to suggest that these groups are by any means representative of the whole student populace, the same way the survey is only a sampling. Quotes were not taken out of context and the research attempted to best consider all relevant aspects of messages in question, which necessitated the extension of analysis to include latent content beyond the obvious manifested content.
Two levels of content analysis were conducted for the extent of this study. The first dealt only with high level, non-complex units of analysis consisting primarily of obvious features gathered about the 17 groups selected for overview. These original 17 were subjectively selected based on a query return of several hundred Chief-related groups. The criteria for consideration consisted of both group size and prospective emotional engagement. Non-serious groups, like being pro-Master Chief from Halo, were disregarded, as were most smaller groups. The exception came in regards to anti-Chief groups, as their relative scale was minute in comparison to the pro-chief giants. A 70-some person pro-chief group is tiny but pretty darn big for anti-chief groups. All in all only 4 anti-chief groups were picked (there were so few and this number was about proportionate to their rate of occurrence) as compared to 13 pro-chief groups. Sadly this analysis did not include several 500-800 person pro-Chief groups, which shows the sheer severity of the cut-off line for selection. The information cataloged for the high level analysis included basic identification information such as the group name, URL, the date observed, classification category, and other inconsistent information such as website or location. A summary of the groups’ written purpose was recorded for reference purposes. Special attention was paid to group composition4, which included the number of members, officers, administrators, and listings of relevant related groups. The relevance of related groups was once again a subjective judgment as massive joke groups like “When I was your age Pluto was a planet” were disregarded but political, sporting, and race-related cultural groups were noted. This set the stage for the second level of content analysis.
To effectively uproot some of the strong attitudes surrounding the issue of the Chief research took to a intense study of two of the most volatile pro and anti chief groups on all of Facebook. Other large groups would have also been an excellent field for observation but unfortunately would have taken too much time to adequately dismantle. Instead sociological-style coding was applied to the wall and discussion topic posts to capture rudimentary understandings of topic, tone, issues of identity, and potential connection to social capital. Due to the subjective nature of this method and inexperience of the researcher, however, this analysis is really best considered exploratory.
 An ambiguous category for the number of responses was embedded amongst questions pertaining to per respondent limitations – I initially mistook it to be the number of times a single respondent could fill out the survey.
 The DMI requires the use of the University built survey builder application which does not allow for skip logic or multiple user pathways, nor does it capture responses of partially filled out surveys.
 This originally included group racial composition, but due to limitations in the reliability of collection was abandoned in favor of a dependable simple analysis. Instead it paved the way for suggestions pertaining to future research.