Blogging Ferguson in Black and White

Douglas L. Mendenhall, PhD (bio)
Abilene Christian University
Instructor/Journalist in Residence
Department of Journalism and Mass Communication
Abilene Christian University

Author Note

Suggesting this study and assisting with its data collection were 47 undergraduate students enrolled in Media Issues, the introductory course within the ACU Department of Journalism and Mass Communication. They were supervised by Jessica Clark.


Blog posts and appended comments about the shooting of a black teen in Ferguson, Missouri, are analyzed for differences based on the race of the authors. Using Diction 7.0, a common word-counting program, seven differences are seen, with black-authored posts higher in commonality, cognition, hardship, human-interest, satisfaction, and self-reference, while white-authored posts are higher in use of collectives. From a social identity perspective, tonal differences do not appear to constitute differing levels of incivility.

Keywords: Ferguson, race, blogs, Diction 7.0, incivility, social identity theory

Blogging Ferguson in Black and White

Online responses to the August 2014 shooting death of Michael Brown in Ferguson, Missouri, can be seen as a virtual continuation of the racial divisions and tensions that erupted after this African-American teenager was shot by a white police officer. Eclipsing even “The Ice Bucket Challenge,” Ferguson immediately became the most discussed news story in the United States (Halimi, 2014), as protests and other developments filled newspapers, television networks, online sites and social media with facts, speculation, analysis, and comments. As with earlier race-based news juggernauts in the United States, such as the cross-racial homicides involving O.J. Simpson in 1994 and George Zimmerman in 2012 (Blow, 2012), whites and blacks quickly developed “a war of competing narratives” (McClune, 2014) around news from Ferguson.

The present study seeks to examine how black and white writers expressed their opinions about Ferguson on the Internet’s most popular political blogs, specifically looking for differences in the tone of their messages. The study examines the tone not only of blog posts but also of the comments appended to them to see whether black and white authors elicited differently toned responses. These analyses are made with the use of Diction 7.0, a widely applied software package that offers a quantitative method for assessing the tone of any written message by searching for and measuring dozens of semantic features. While a simple tally would be sufficient to test whether whites and blacks lined up on different sides in discussing the altercation between Michael Brown and Officer Darren Wilson, this study seeks to go deeper and probe what differences can be seen in the tone of the messages with which they expressed their positions. Also, with roughly two-thirds of each of the past four generations—Millenials, Gen Xers, Baby Boomers, and the Silent Generation—agreeing that “Incivility in America has risen to crisis levels” (Weber Shandwick, Powell Tate, 2014, p. 3) and even greater numbers agreeing that online social media including blogs encourage incivility (p. 9), this study will ask whether racial differences emerge in the levels of uncivil language used by black and white bloggers.

Literature review

Cyberspace represents a roughly level playing field for studying differences in message tone along racial lines. The most recent Pew survey finds that in most categories Internet use among African-Americans is equivalent to that of whites, with black representation higher than white in certain social networking platforms such as Twitter (Smith, 2014) and with the continuation of a trend first identified in 2009 of blacks using smart phones at a higher rate than whites (Smith, 2010). Across a spectrum of online activities that includes blogging, blacks and other minority groups were found to be more active in the creation of content, with the enhancement of social identity identified as a primary motivation for such creation (Correa & Jeong, 2011). Furthermore, Daniels (2013) notes that blogging by African-Americans and other minorities can allow for the building of “diasporic communities” in which the members shape one another and their scattered audiences (p. 9).

Such findings pave the way for the use of social identity theory as a perspective from which to consider the Ferguson blogs. This explanation rises from a general assumption permeating social psychology that while individual behavior is wrapped up in the human search for personal meaning, this often happens within group settings (Tajfel & Turner, 1979). People who satisfy one another’s needs and who expect to achieve satisfactions from their association quickly develop feelings of mutual attraction and hence become a group (Turner et al., 1987).

Refinement of this theory by Abram and Hogg (1988) led to the eventual suggestion that, “Social identity and intergroup behavior is guided by the pursuit of evaluative positive social identity through positive intergroup distinctiveness, which in turn is motivated by the need for positive self-esteem” (Hogg & Terry, 2000, p. 124). In other words, individuals prefer to view themselves as part of a group that represents the most personally advantageous association they could have made. They gain self-esteem from group esteem and will therefore work harder to see their chosen group succeed (Derks, van Laar, & Ellemers, 2009).

One way to increase such gratification is to maximize the distance between the “in” (or self-identified) and “out” groups (Tajfel & Turner, 1979). That is, we enhance our own group to increase our own self-worth and inherently seem to believe that pushing other groups down elevates our own (Platow et al., 1999). Furthermore, an individual tends to identify with a group and judge others by their group traits rather than by their individual traits more during a time of threat to the in-group’s social identity (Branscombe & Wann, 1994). In some cases, outright racism can result from such social identification, as entire sectors of a society are judged as a group based on skin color or ethnicity (Tajfel & Turner, 1979). Yet a clear marker such as skin color is not required: Early experiments (Tajfel, 1970, 1981, 1982; Tajfel & Turner, 1979) suggest that group identification and preference can be triggered even by something as minimal as giving a common but meaningless label to a subset of a large group. The salience and strength of this self-identification with a group is seen to rise and fall depending on external pressures (Huddy, 2001).

One stream of research using social identity theory has focused on how individual and group identities are created, manipulated, and used in online communication given the ease with which users can shield their personal information. Internet participation also brings into play certain commonalities, breaking participants into groups along lines of gender, age, race, and differentiations in the amount and intensity of their online use (King, 2001) that allow for similar affective responses throughout the Internet audience, and creating different types of public space depending on the group identities employed there. Visual anonymity, rather than interfering with the processes of SIT, has been found to directly increase an online participant’s attraction to a group and indirectly increases that participant’s positive stereotyping of the group (Lea, Spears, & de Groot, 2001). Consequent research suggests that this tugging toward a group caused by online anonymity can even overcome individual personality traits (Lee, 2006). Predating social identity theory—and the flaming Internet by almost a century—is a strangely relevant perception about ethnocentrism actually being an expression of self-love at a group level: “The insiders in a we-group are in a relation of peace, order, law, government, and industry to each other. Their relation to all outsiders, or others-groups, is one of war and plunder, except in so far as agreements have modified it” (Sumner, 1906, p. 12).

Consistent with this line of thinking, in times of war, plunder, and other major disturbances SIT would predict clearly defined differences between two groups of bloggers—and evidence to support such a perspective is indeed available along racial dividing lines. For example, one researcher cites major themes of racial inequity and white privilege within the stream of blogs and other Internet messages from African-American authors in the disturbing wake of Hurricane Katrina, exemplified by a post headlined, “Nobody gets out of here who’s not white” (Brock, 2008, p. 14), that spoke to concerns about inequities in the evacuation of New Orleans. Demonstrating a different type of racial difference during a different type of crisis, authors within an African-American blog community were found to utilize within their posts about the effects of the AIDS epidemic at least three traditional black communication conventions: interaction, communal support, and kinesics—body movements including dancing and singing (Kvasny & Igny, 2008, p. 584). Another researcher found that black-authored blogs represent a unique opportunity to enhance discussions about niche topics such as health concerns that primarily affect African-Americans (Della, Griffin, Eroğlu, Bernhardt, & Wells, 2013). Thus, the literature suggests that times of stress often bring out racially focused differences in the blogosphere. Furthermore, when such online differences are expressed with uncivil language on blog sites, it can have a “nasty effect” that results in increased polarization of groups within the audiences (Anderson, Brossard, Scheufele, Xenos, & Ladwig, 2014).

As for what is achieved by such racially distinctive message creation, many researchers have concluded that refining social identity through Internet participation can for African-Americans be a positive step in dealing with societal inequities. “Much like the world offline, participating in online cultural communities will help them to develop a healthy sense of racial identity, what psychologists argue is necessary to resist the pernicious effects of racism” (Byrne, 2008, p. 33). This group social identity also can be seen in the finding that black readers browsing a news site were more likely to select both positively and negatively valenced stories about their own race, but only negatively valenced stories about their racial outgroup—a habit not shared by white readers in the study (Appiah, Knobloch-Westerwick, & Alter, 2013).

Thus, SIT suggests that in the wake of the Michael Brown shooting in Missouri, we can expect to see increased salience for group identities built along racial lines. The present study chooses to focus on the blogosphere to study possible differences in how these racial groups express themselves because the diffusion of the Internet offers a potential for minority authors to speak as often as they wish that can be lacking in the more traditional media realm, where for example only 4% of the books reviewed by a recognized progressive outlet such as the New York Times are written by Africans or African-Americans (Gay, 2012). Online, such a lag for black voices is less of a problem. Schradie (2011) reports that lower educational levels still serve as a barrier to the ability of African-Americans to blog, but notes that once they overcome this barrier and gain online access, blacks blog at a higher rate than whites with access. Admittedly, even if, as Schradie suggests, this difference is based on socio-economic factors rather than race, it may bring with it the problem of African-American bloggers being limited in their ability to reach beyond their silos and speak to working-class blacks or to broader white audiences. Because of that limitation, comparing the sizes of the actual audiences reached by black and by white bloggers is mostly beyond the scope of this study, which will be limited to comparing the tone of the messages offered for consumption.

Thus, the current study of Ferguson-related online messages attempts to shed light on three main questions:

RQ1: In what ways if any does the tone of messages created by white authors differ from the tone of messages created by black authors as they comment on a sensitive, race-related current event?

RQ2: In what ways if any does the tone of reader responses to white authors differ from the tone of reader responses to black authors?

RQ3: Are different levels of incivility apparent in the tone of messages created by black and by white authors?


Diction 7.0 provides a stable, quantitative look at what is usually a more nuanced, qualitative undertaking: understanding the tone of a mass media message that builds from the connotations of each word selected and positioned by an author. An established research guideline is that the use of computers as an aid for content analysis should be limited to cases in which this will produce valid measures, reduce the study’s cost, and improve reliability (Riffe, Lacy, & Fico, 2005). Diction 7.0 software offers a scientific method for quickly measuring the tone of a written message by searching for and measuring dozens of semantic features. The software employs 31 non-overlapping, internal dictionaries that define the conceptual categories. The software computes quantitative scores for these categories based on the content properties of the texts in question. Moreover, scores for these categories are combined and subjected to mathematical functions to yield composite scores for some variables of interest, including the five “Master Variables” of activity, optimism, certainty, realism, and commonality. Diction 7.0 relies on a 10,000-word search routine that can be used with any kind of English-language texts.  Scholars have employed Diction to examine a variety of topics across numerous communication contexts (Young & Soroka, 2012). A recent example is Political Tone: How Leaders Speak and Why (Hart, Childers, & Lind, 2013), co-authored by one of the developers of Diction. Political Tone uses the word-counting software to reveal the overall tone of contemporary political messages, asserting that tone consists of “(1) individual word choices that (2) cumulatively build up (3) to produce patterned expectations (4) telling an audience something important (5) about the author’s outlook on things” (p. 12) and using that template to calculate and differentiate between what constitutes, for instance, an overall “Democratic” tone or “Republican” tone (p. 71). Other examples of researchers applying Diction include analyses of the tone of campaign speeches by winning U.S. presidential candidates (Lowry & Naser, 2010), of differences between the rhetoric at local and national Promise Keepers events (Eidenmuller, 2002), of media coverage of the 9/11 terror attacks (Cho et al., 2003), of the comparative optimism of governors and presidents (Crew and Lewis, 2011), and of the use of power in the speeches of Adolf Hitler and Martin Luther King Jr. (Robinson & Topping, 2012). These studies and others demonstrate the broad acceptance of Diction within the communication discipline as an appropriate tool for comparison of tone within categories of messages such as those to be gathered from popular blog sites for the current study.

At a small southwestern college, 47 students in a beginning mass media course worked in three-person teams to collect Ferguson-related blogs published from August 9 to August 23, representing the first two weeks following the shooting death of Michael Brown. A team was assigned to each of the 15 most popular political websites in the United States (, October 2014) and instructed to save up to 100 random, Ferguson-related blog posts as well as the first 20 comments appended by readers of each blog post. Sites included in the study are, in order of their popularity: Huffington Post, The Blaze, Drudge Report, News Max, Politico, Salon, Info Wars, Brietbart, Daily Caller, Washington Times, Christian Science Monitor, WND, Daily Kos, Think Progress, and Town Hall. Because of the immensity of the Huffington Post site and to increase the number of black-authored posts, collection there was limited to the “Black Voices” section. Posts were converted to text files and efforts were made to categorize their authors by race using an accompanying photograph or biography. Several more authors were categorized after employing a search engine to seek more personal information about them. Blog posts for which the author’s racial category could not be determined were eliminated from the sample. Finally, all remaining files were vetted to eliminate those that consisted of traditional news stories, leaving only those that consisted of commentary or opinion. This resulted in a sample frame of 157 blog posts, of which 119 were identified as having white authors and 38 as having black authors. Also collected were 114 sets of up to the first 20 comments appended to these blog posts, including 86 sets from white-authored posts and 28 sets from black-authored posts. The standard message length for Diction analysis is 500 words, so the first 20 comments on a post were considered as a single message to avoid the inability to individually assess the many brief comments. Several blog sites did not allow comments or had removed them from their site by the time collection was attempted in September and October. Each of the 271 collected files was analyzed using Diction 7.0 and transferred to SPSS.


Splitting the 157 blog posts by the race of the authors, one-way ANOVAs were used to compare mean scores for each of the 35 Diction variables, the five Diction Master Variables calculated using combinations of those variables, and a variable constructed to indicate the use of 20 vulgar or profane words commonly found on Internet sites according to two studies (Kirk, 2013; GFK Group, 2013).

Answering RQ1 in the affirmative, significant differences in the tone of white- and black-authored messages were found in seven variables at the level of p < .05, as displayed in Table 1 and Table 2. The measured variable was significantly higher for black-authored messages in these six instances: cognition (p = .01), commonality (p = .02), hardship (p = .02), human interest (p = .002), satisfaction (p = .03), and self-reference (p = .001). Only the variable of collectives (p = .02) was significantly higher in white-authored messages. Definitions of these seven significant variables are provided in Table 3. Commonality is one of Diction’s five Master Variables, calculated by combining the variables of centrality, cooperation, and rapport, and subtracting from this subtotal the variables of diversity, exclusion, and liberation. Despite a clear black/white difference in commonality at the level of p < .05, however, significant black/white differences were not recorded in any of the six components of this Master Variable, suggesting a small but consistent trend rather than a single spike within one of the six components. Definitions of the six components of commonality are included in Table 3.

Answering RQ2 in the negative, a similar analysis of the 114 sets of comments appended to these blog posts found no variables fluctuating significantly based on the race of the author of the message on which readers were commenting. No effort was made to determine the race of each commenter because of a clear trend of anonymity for such blog comments.

Providing an answer to RQ3, about incivility, requires a discussion of what it means that black authors reflected higher levels of cognition, commonality, hardship, human interest, satisfaction, and self-reference, as well as a lower level of collectives.


Commonality: The higher mean score for the Diction master variable of commonality for black-authored blog posts than for white-authored ones is indicative of increased use of words that express agreement on core values (centrality), working together as a group (cooperation), and sharing attitudinal similarities as a group (rapport). It also indicates decreased use of words that describe individuals who differ from the group norm (diversity), social isolation (exclusion), and the importance of individual freedom above social convention (liberation). Within the sample of blog posts about the death of Michael Brown and subsequent events in Ferguson, the highest commonality score was 55.39 for a black-authored Huffington Post article outlining five ways for educators to teach students about Michael Brown in the coming year (Emdin, 2014), and the top four scores for commonality were all for black-authored messages. Among white-authored posts, the highest score for commonality was 52.45 for an assertion that widespread grassroots opposition to the militarization of local police departments was unlikely to lead to bipartisan legislation to stop it (Donovan, 2014). The high black commonality score can be seen as evidence of what social identity predicts, that an online African-American community would coalesce around the discussion of issues common to their experience.

Cognition: The heightened variable of cognitive terms in the writings of black bloggers suggests that overall they made greater use than white bloggers of words expressing both rationalistic (perception, speculation, judging) and intuitional (knowledge, thinking, inquiry) thought processes. The highest cognition score from a black-authored post was 35.57 on the aforementioned article on how to teach students about Michael Brown (Emdin, 2014). The second highest score was 23.97 on a brief exhortation by MLK niece Dr. Alveda C. King on to “think rationally, honestly, and morally right to accord with the thinking of the creator” (King, 2014). The highest white-authored cognition score of 40.21 was on a Think Progress explanation of why an aircraft “no-fly zone” had been imposed on Ferguson (Strasser, 2014). The higher black score on this variable suggests assertions by African-American authors that more attention needs to be focused on Ferguson, with greater efforts made to understand events and their broader implications. Again, this can be seen as consistent with SIT expectations of the emergence of a black community coalescing via blog messages about how to deal with common issues and implications.

Hardship: A higher mean score for the variable of hardship within black-authored blog posts is indicative of a variety of negative actions, from natural disasters (starvation, tornado, pollution) to hostile (rebellion, bankruptcy, infidelity) or even illegal human behavior (slavery, exploitation, killers). The highest hardship score from a black-authored post was 22.35 for a Salon essay that drew examples from a century of mistreatment by whites in defending black rage:

Nothing makes white people more uncomfortable than black anger. But nothing is more threatening to black people on a systemic level than white anger. It won’t show up in mass killings. It will show up in over-policing, mass incarceration, the gutting of the social safety net, and the occasional dead black kid. Of late, though, these killings have been far more than occasional. We should sit up and pay attention to where this trail of black bodies leads us.  They are a compass pointing us to a raging fire just beneath the surface of our national consciousness. We feel it. We hear it. Our nostrils flare with the smell of it (Cooper, 2014).

Scoring 17.67 for its hardship tone, the next highest black-authored message was the aforementioned piece by Alveda C. King, who also drew on historic examples from her uncle’s struggle for civil rights in the 1960s to explain contemporary black rage (King, 2014). The highest score for hardship in the entire sample was 29.66 for a brief, white-authored pondering of how the events in Ferguson would have been covered by news media if they happened in China or Russia (Byers, 2014). The higher black score on this variable suggests that black authors more frequently tied events in Ferguson into a broader narrative of past and present grievances shared by the community, again consistent with SIT expectations.

Human-interest: The human-interest variable attempts to measure the extent to which human beings and their activities are central to the discourse. Its dictionary includes personal pronouns, words describing family members and generic human terms. The highest human-interest score was 70 for a Huffington Post “open letter to the people of Ferguson” whose author referred to himself as an African-American male and reflected on the pain felt by Michael Brown’s grandmother as well as by wider categories of black mothers and fathers (Sanders, 2014). The highest white human-interest score was 68 for a vote of confidence in the investigation of U.S. Attorney General Eric Holder, which dealt with the autopsy performed on Michael Brown’s body (Salzillo, 2014). The higher black mean on this variable suggests that black authors treated the subject from a more personal viewpoint, discussing its effects on individual human beings or on specific groups of people. Once again, this can be seen as consistent with the predictions of SIT.

Satisfaction: The variable of satisfaction measures the presence of terms associated with joy, pleasure, triumph and other positive affective states. A black-authored Huffington Post essay praising the Internet for giving African-Americans a better way to rally support and form communities earned the highest score for satisfaction, at 19.86 (Cyril, 2014). The highest satisfaction score for a white-authored post was 17.62 for the Salzillo (2014) essay mentioned above that compliments the investigation by the U.S. attorney general. Again, the higher mean score for black-authored messages is consistent with an SIT expectation that membership in any in-group is likely to bring with it an optimized attitude about the benefits of belonging to that group.

Self-reference: The variable of self-reference indexes all first-person references—such as I, I’m, I’ve, me, mine, myself—to show instances in which “the locus of action appears to reside in the speaker and not in the world at large.” The top score for black-authored self-reference is 27.28 on a mother’s first-person narrative that begins:

I don’t want to talk about the boy and the sneakers peeking out from the sheet crudely draped over his corpse in the street, because I have been happy this month and it is so rare that I'm happy and that you, at age four, don’t have to touch my knee or shoulder or face and say, “What's wrong, Mama? You sad?” (Brown, 2014)

The highest self-reference score for a message by a white author is 36.25, again for the Salzillo attorney general essay. Next behind that is a score of 16.87 for a point-by-point analysis of events in Ferguson in which the author offers his personal expertise on a number of topics, including concussions, police radio technology, and the Trayvon Martin case (Hayden, 2014). Again, the higher black-authored mean for this variable is consistent with SIT predictions that personal identity will be closely tied to in-group membership.

Collectives: Only on the variable of collectives was the mean score for black-authored messages significantly lower than for white-authored ones. In a way, this variable can be seen as the opposite of self-reference, in that it is built on terms that describe groups in mostly non-specific ways, such as team, crowd, staff, county, republic. A higher mean score for white-authored messages can be read as the consequence of less personal involvement with the issue being discussed than is exhibited by black bloggers. As in all of the six above variables exhibiting a significant difference along racial lines, this finding appears to be consistent with SIT expectations.

More study is needed to clarify the processes by which online communities emerge and operate, especially in times of crisis, but the present study does suggest that among black authors writing about a news event of especially high salience for African-Americans, the ways in which their messages had a different tone than those written by whites on the same topic were consistent with qualities that social identity theory would expect to see. It is just as important to note that these seven differences, the only ones measured by Diction’s spectrum of variables, do not seem consistent with heightened expressions of incivility. Certainly it is possible that tones indicating hostility, anger, resentment, and even calls for violence are heightened by either black or white bloggers, but if that is the case, future researchers may need to employ different tools to measure it.

Table 1: ANOVAs of Diction variables that differ significantly by bloggers’ race

Cognition: Black-authored posts are higher than white at the .05 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model            229.595b                      1                 229.595                     6.682                        .011                .041

Intercept                           8017.873                      1              8017.873                 233.344                        .000                .601

Author race                         229.595                      1                 229.595                     6.682                        .011                .041

Error                                    5325.921                 155                   34.361                                                                                       

Total                                  14657.155                 157                                                                                                                       

Corrected Total               5555.516                 156                                                                                                                       

R Squared = .041 (Adjusted R Squared = .035)

Commonality: Black-authored posts are higher than white at the .05 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model               20.978b                      1                   20.978                     5.211                        .024                .033

Intercept                       279071.871                      1          279071.871            69327.730                        .000                .998

Author race                           20.978                      1                   20.978                     5.211                        .024                .033

Error                                       623.937                 155                     4.025                                                                                       

Total                                377548.997                 157                                                                                                                       

Corrected Total                 644.915                 156                                                                                                                       

R Squared = .033 (Adjusted R Squared = .026)


Hardship: Black-authored posts are higher than white at the .05 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model            127.440b                      1                 127.440                     5.660                        .019                .035

Intercept                           7839.688                      1              7839.688                 348.197                        .000                .692

Author race                         127.440                      1                 127.440                     5.660                        .019                .035

Error                                    3489.843                 155                   22.515                                                                                       

Total                                  12941.373                 157                                                                                                                       

Corrected Total               3617.283                 156                                                                                                                       

R Squared = .035 (Adjusted R Squared = .029)


Human Interest: Black-authored posts are higher than white at the .005 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model          1287.043b                      1              1287.043                   10.370                        .002                .063

Intercept                         67712.709                      1            67712.709                 545.596                        .000                .779

Author race                      1287.043                      1              1287.043                   10.370                        .002                .063

Error                                  19236.709                 155                 124.108                                                                                       

Total                                100137.807                 157                                                                                                                       

Corrected Total             20523.752                 156                                                                                                                       

R Squared = .063 (Adjusted R Squared = .057)


Satisfaction: Black-authored posts are higher than white at the .05 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model               43.307b                      1                   43.307                     4.715                        .031                .030

Intercept                           1344.893                      1              1344.893                 146.427                        .000                .486

Author race                           43.307                      1                   43.307                     4.715                        .031                .030

Error                                    1423.637                 155                     9.185                                                                                       

Total                                    2976.025                 157                                                                                                                       

Corrected Total               1466.944                 156                                                                                                                       

R Squared = .030 (Adjusted R Squared = .023)


Self-reference: Black-authored posts are higher than white at the .005 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model            253.025b                      1                 253.025                   11.505                        .001                .069

Intercept                           1517.574                      1              1517.574                   69.003                        .000                .308

Author race                         253.025                      1                 253.025                   11.505                        .001                .069

Error                                    3408.916                 155                   21.993                                                                                       

Total                                    4950.432                 157                                                                                                                       

Corrected Total               3661.941                 156                                                                                                                       

R Squared = .069 (Adjusted R Squared = .063)

Collectives: White-authored posts are higher than black at the .05 level

Source                         Type III Sum                    df                    Mean                              F                         Sig.   Partial Eta

                                          of Squares                                          Square                                                                       Squared

Corrected Model            189.028b                      1                 189.028                     5.098                        .025                .032

Intercept                         12662.760                      1            12662.760                 341.498                        .000                .688

Author race                         189.028                      1                 189.028                     5.098                        .025                .032

Error                                    5747.398                 155                   37.080                                                                                       

Total                                  25436.331                 157                                                                                                                       

Corrected Total               5936.426                 156                                                                                                                       

R Squared = .032 (Adjusted R Squared = .026)

Table 2: Descriptives of Diction variables that differ significantly by blogger race

95% CI for Mean

                                 N            Mean          Std. Dev.     Std. Error           Lower               Upper          Min.              Max.


White                 119      6.930588        5.5283777      .5067856      5.927015          7.934162         .0000         40.2100

Black                     38      9.753947        6.8170920    1.1058784      7.513225        11.994670       1.2300         35.5700

Total                   157      7.613949        5.9676028      .4762666      6.673186          8.554712         .0000         40.2100


White                 119         48.7900             1.90391           .17453         48.4444             49.1356         41.86             52.45

Black                     38         49.6434             2.30278           .37356         48.8865             50.4003         44.92             55.39

Total                   157         48.9966             2.03324           .16227         48.6760             49.3171         41.86             55.39


White                 119      7.197311        4.7466293      .4351228      6.335649          8.058973         .0000         29.6600

Black                     38      9.300789        4.7398394      .7689035      7.742843        10.858736         .5000         22.3500

Total                   157      7.706433        4.8153620      .3843077      6.947315          8.465551         .0000         29.6600

Human Interest

White                 119    20.900840      10.6559219      .9768268    18.966457        22.835223       1.7600         68.0000

Black                     38    27.585526      12.5611653    2.0376901    23.456774        31.714279       8.9000         70.0000

Total                   157    22.518790      11.4700703      .9154113    20.710589        24.326990       1.7600         70.0000


White                 119      2.803529        2.5626293      .2349158      2.338332          3.268727         .0000         17.6200

Black                     38      4.029737        4.1872499      .6792616      2.653422          5.406052         .0000         19.8600

Total                   157      3.100318        3.0665111      .2447342      2.616898          3.583739         .0000         19.8600


White                 119      2.147395        4.2646838      .3909429      1.373222          2.921568         .0000         36.2500

Black                     38      5.111316        5.8420369      .9477035      3.191086          7.031546         .0000         27.2800

Total                   157      2.864777        4.8449956      .3866727      2.100987          3.628567         .0000         36.2500


White                 119    11.764706        6.2408577      .5720985    10.631795        12.897617       2.5400         30.0800

Black                     38      9.202895        5.5786712      .9049800      7.369231        11.036558       1.3200         22.7700

Total                   157    11.144650        6.1687933      .4923233    10.172170        12.117130       1.3200         30.0800

Table 3: Definitions of the 7 Diction variables found to be significant

Cognitive terms:Words referring to cerebral processes, both functional and imaginative. Included are modes of discovery (learn, deliberate, consider, compare) and domains of study (biology, psychology, logic, economics). The dictionary includes mental challenges (question, forget, re-examine, paradoxes), institutional learning practices (graduation, teaching, classrooms), as well as three forms of intellection: intuitional (invent, perceive, speculate, interpret), rationalistic (estimate, examine, reasonable, strategies), and calculative (diagnose, analyze, software, fact-finding).

Commonality (1 of 5 Diction master variables): Language highlighting the agreed-upon values of a group and rejecting idiosyncratic modes of engagement. Commonality Formula built by converting these variables to Z-scores, then concatenating them as: [Centrality + Cooperation + Rapport] – [Diversity + Exclusion + Liberation]. These six components are defined as:

CENTRALITY: Terms denoting institutional regularities and/or substantive agreement on core values. Included are indigenous terms (native, basic, innate) and designations of legitimacy (orthodox, decorum, constitutional, ratified), systematicity (paradigm, bureaucratic, ritualistic), and typicality (standardized, matter-of-fact, regularity). Also included are terms of congruence (conformity, mandate, unanimous), predictability (expected, continuity, reliable), and universality (womankind, perennial, landmarks).

COOPERATION: Terms designating behavioral interactions among people that often result in a group product. Included are designations of formal work relations (unions, schoolmates, caucus) and informal associations (chum, partner, cronies) to more intimate interactions (sisterhood, friendship, comrade). Also included are neutral interactions (consolidate, mediate, alignment), job-related tasks (network, detente, exchange), personal involvement (teamwork, sharing, contribute), and self-denial (public- spirited, care-taking, self-sacrifice).

RAPPORT: This dictionary describes attitudinal similarities among groups of people. Included are terms of affinity (congenial, camaraderie, companion), assent (approve, vouched, warrants), deference (tolerant, willing, permission), and id entity (equivalent, resemble, consensus).

DIVERSITY: Words describing individuals or groups of individuals differing from the norm. Such distinctiveness may be comparatively neutral (inconsistent, contrasting, non-conformist) but it can also be positive (exceptional, unique, individualistic) and negative (illegitimate, rabble-rouser, extremist). Functionally, heterogeneity may be an asset (far-flung, dispersed, diffuse) or a liability (factionalism, deviancy, quirky) as can its characterizations: rare vs. queer, variety vs. jumble, distinctive vs. disobedient.

EXCLUSION: A dictionary describing the sources and effects of social isolation. Such seclusion can be phrased passively (displaced, sequestered) as well as positively (self-contained, self-sufficient) and negatively (outlaws, repudiated). Moreover, it can result from voluntary forces (secede, privacy) and involuntary forces (ostracize, forsake, discriminate) and from both personality factors (smallmindedness, loneliness) and political factors (right-wingers, nihilism). Exclusion is often a dialectical concept: hermit vs. derelict, refugee vs. pariah, discard vs. spurn).

LIBERATION: Terms describing the maximizing of individual choice (autonomous, open-minded, options) and the rejection of social conventions (unencumbered, radical, released). Liberation is motivated by both personality factors (eccentric, impetuous, flighty) and political forces (suffrage, liberty, freedom, emancipation) and may produce dramatic outcomes (exodus, riotous, deliverance) or subdued effects (loosen, disentangle, outpouring). Liberatory terms also admit to rival characterizations: exemption vs. loophole, elope vs. abscond, uninhibited vs. outlandish.

Hardship: This dictionary contains natural disasters (earthquake, starvation, tornado, pollution), hostile actions (killers, bankruptcy, enemies, vices) and censurable human behavior (infidelity, despots, betrayal). It also includes unsavory political outcomes (injustice, slavery, exploitation, rebellion) as well as normal human fears (grief, unemployment, died, apprehension) and incapacities (error, cop-outs, weakness).

Human interest: An adaptation of Rudolf Flesch’s notion that concentrating on people and their activities gives discourse a life-like quality. Included are standard personal pronouns (he, his, ourselves, them), family members and relations (cousin, wife, grandchild, uncle), and generic terms (friend, baby, human, persons).

Satisfaction: Terms associated with positive affective states (cheerful, passionate, happiness), with moments of undiminished joy (thanks, smile, welcome) and pleasurable diversion (excited, fun, lucky), or with moments of triumph (celebrating, pride, auspicious). Also included are words of nurturance: healing, encourage, secure, relieved.

Self-reference: All first-person references, including I, I’d, I’ll, I’m, I’ve, me, mine, my, myself. Self- references are treated as acts of indexing whereby the locus of action appears to reside in the speaker and not in the world at large thereby implicitly acknowledging the speaker’s limited vision.

Collectives: Singular nouns connoting plurality that function to decrease specificity. These words reflect a dependence on categorical modes of thought. Included are social groupings (crowd, choir, team, humanity), task groups (army, congress, legislature, staff) and geographical entities (county, world, kingdom, republic).

Author Bio

Doug Mendenhall is an assistant professor in the Department of Journalism and Mass Communication at Abilene Christian University, where he is journalist in residence, teaches courses related to writing, history, race and religion, and oversees the department’s internship program. In 2013 he was instrumental in developing the department’s guest lecture series on race and media.

Mendenhall joined the ACU faculty in 2008 after a 26-year career with daily newspapers in Texas, Tennessee and Alabama, with an emphasis on design and graphics. Since 2000 he has written a weekly newspaper column on matters of faith and values.

Mendenhall earned his master’s degree from Middle Tennessee State University in 2003, with a dissertation that used live newspaper distribution to compare the audience’s perceptions of two different design styles on a single front page.  He earned his doctorate in 2014 from Texas Tech University with a dissertation titled, “Comparing Levels of Incivility Across Political and Religious Blog Posts.”


Abrams, D., & Hogg, M. A. (1988). Comments on the motivational status of self-esteem in social identity and intergroup discrimination. European Journal Of Social Psychology, 18(4), 317-334.

Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A. and Ladwig, P. (2014), The “Nasty Effect:” Online Incivility and Risk Perceptions of Emerging Technologies. Journal of Computer-Mediated Communication, 19: 373–387. doi: 10.1111/jcc4.12009

Appiah, O., Knobloch-Westerwick, S., & Alter, S. (2013). Ingroup Favoritism and Outgroup Derogation: Effects of News Valence, Character Race, and Recipient Race on Selective News Reading. Journal Of Communication, 63(3), 517-534. doi:10.1111/jcom.12032

Blow, C. M. (2012, April 6). From O.J. to Trayvon. The New York Times. Retrieved from

Branscombe, N. R, & Wann, D. A. (1994) Collective self-esteem consequences of outgroup derogation when a valued social identity is on trial. European Journal of Social Psychology, 24, 641-657.

Brock, A. (2008). Race matters: African Americans on the Web following hurricane Katrina. In F. Sudweeks, H. Hrachovec, and C. Ess, (Eds.) Proceedings of Cultural Attitudes toward Communication and Technology (pp. 91-105). Accessed at

Brown, S. L. (2014, August 11). When Parenting Feels Like a Fool's Errand: On the Death of Michael Brown.

Byers, D. (2014, August 18). Ferguson from 30,000 feet.

Byrne, D. N. (2008). The Future of (the) “Race”: Identity, Discourse, and the Rise of Computer-mediated Public Spheres. In A. Everett (Ed.) Learning Race and Ethnicity: Youth and Digital Media. (pp.15-38). Cambridge, MA: The MIT Press. doi: 10.1162/dmal.9780262550673.015

Cho, J., Boyle, M. P., Keum, H., Shevy, M. D., Mcleod, D. M., Shah, D. V., & Pan, Z. (2003). Media, Terrorism, and Emotionality: Emotional Differences in Media Context and Public Reactions to the September 11th Terrorist Attacks. Journal Of Broadcasting & Electronic Media, 47(3), 309-327.

Cooper, B. (2014, August 12). In defense of black rage: Michael Brown, police and the American dream. michael_brown_police_and_the_american_dream/

Correa, T., and Jeong, S. H. (2011). Race and Online Content Creation. Information, Communication & Society, 14(5), 638-659. doi:10.1080/1369118X.2010.514355

Crew, R. E., & Lewis, C. (2011). Verbal Style, Gubernatorial Strategies, and Legislative Success. Political Psychology, 32(4). doi: 10.1111/j.1467-9221.2011.00832.x

Cyril, M. A. (2014, August 15). Thank You, Black Internet, For Bringing #Ferguson to Me.

Daniels, J. (2013). Race and racism in Internet Studies: A review and critique. New Media & Society, 15(5), 695-719. doi:10.1177/1461444812462849

Della, L. J., Griffin, D. B., Eroğlu, D., Bernhardt, J. M., & Wells, R. R. (2013). Is There Health Out There in the Afrosphere? An Analysis of Health-Related Content Posted by Black Bloggers. Health Marketing Quarterly, 30(1), 1-18. doi:10.1080/07359683.2013.758012

Derks, B., van Laar, C., & Ellemers, N. (2009). Working for the Self or Working for the Group: How Self- Versus Group Affirmation Affects Collective Behavior in Low-Status Groups, Journal of Personality and Social Psychology, 96, 183-202.

Donovan, T. (2014, August 18). Spineless frauds exposed by Ferguson: Where pols really stand on militarized cops. (October 2014). 15 most popular political websites | October 2014.

Eidenmuller, M. E. (2002). American Evangelicalism, Democracy, and Civic Piety: A Computer-Based Stylistic Analysis of Promise Keepers’ Stadium Event and Washington D.C. Rally Discourses. Journal of Communication & Religion 25(1). 64-85.

Emdin, C. (2014, August 20). 5 Ways to Teach About Michael Brown and Ferguson in the New School Year.

Gay, R. (2012, June 6). Where Things Stand.  Accessed at

GFK Group (2013). The Digital Abuse Study: A survey from MTV & The Associated Press-NORC Center for Public Affairs. Discrimination/AP-NORC-MTV%202013%20Topline_FINAL_ALL%20Qs.pdf

Halimi, N. (2014, September 30). Top Twitter Hashtags of the Past 30 Days. Retrieved from

Hart, R. P., Childers, J. P., & Lind, C. J. (2013). Political Tone: Why Leaders Talk and Why. Chicago: University of Chicago Press.

Hayden, J. (2014, August 21). The case of Michael Brown: Missing police reports, anonymous sources and shoddy journalism.

Hogg, M. A, & Abrams, D. (1988). Social Identifications: A Social Psychology of Intergroup Relations and Group Processes.

Hogg, M. A., & Terry, D. J. (2000). Social identity and self-categorization processes in organizational contexts, Academy of Management Review (25)1, 121-140.

Huddy, L. (2001). From Social to Political Identity: A Critical Examination of Social Identity Theory. Political Psychology 22(1). 127-156.

King, A. B. (2001). Affective Dimensions of Internet Culture. Social Science Computer Review, 19(4), 414-430.

King, A. C (2014, August 18). Pray for Justice in Ferguson.

Kvasny, L., and Igwe, C. F. (2008), An African American Weblog Community’s Reading of AIDS in Black America. Journal of Computer-Mediated Communication, 13: 569–592. doi: 10.1111/j.1083-6101.2008.00411.x

Kirk, C. (2013, September 11). The most popular swear words on Facebook. Lexicon Valley: A Blog About Language. Retrieved from

Lea, M., Spears, R., & de Groot, D. (2001). Knowing me, knowing you: Anonymity effects on social identity processes within groups. Personality And Social Psychology Bulletin, 27(5), 526-537. doi:10.1177/0146167201275002

Lee, E. (2006). When and How Does Depersonalization Increase Conformity to Group Norms in Computer-Mediated Communication?. Communication Research, 33(6), 423-447.

Lowry, D. T., & Naser, M. A. (2010). From Eisenhower to Obama: Lexical Characteristics of Winning Versus Losing Presidential Campaign Commercials. Journalism & Mass Communication Quarterly 87(3/4). 530-547.

McClune, B. (2014, August 29). Ferguson: A Tale of Two Narratives. Retrieved from

Platow, M. J., Durante, M., Williams, N., Garrett, M., Walshe, J., Cincotta, S., ... & Barutchu, A. (1999). The contribution of sport fan social identity to the production of prosocial behavior. Group Dynamics: Theory, Research, and Practice, 3(2), 161-169.

Riffe, D. D., Lacy, S. R., & Fico, F.(2005). Analyzing Media Messages: Using Quantitative Content Analysis in Research. New York: Routledge.

Robinson, J. L., & Topping, D. (2012). The Rhetoric of Power: A Comparison of Hitler and Martin Luther King Jr. Journal of Management Inquiry 21(3). 1-17.

Salzillo, L. (2014, August 18). U.S. Attorney Eric Holder Releases New Statement on Ferguson Investigation Update.

Sanders, J. (2014, August 11). A Letter to the People of Ferguson, Missouri.

Schradie, J. (2011). The digital production gap: The digital divide and Web 2.0 collide. Poetics, 39(2), 145-168. doi:10.1016/j.poetic.2011.02.003

Smith, A. (2010). Mobile Access 2010, Pew Internet and American Life Project. Washington D.C. Accessed at:

Smith, A. (2014). African Americans and Technology Use, Pew Internet and American Life Project. Accessed at:

Strasser, A. R. (2014, August 18). Why There’s a No-Fly Zone Over Ferguson, Missouri.

Sumner, W. G. (1906). Folkways: A Study of the Sociological Importance of Usages, Manners, Customs, Mores, and Morals. Boston: The Athenaeum Press.

Tajfel, H. (1970). Experiments in intergroup discrimination. Scientific American, 223, 96-102

Tajfel, H. (1981). Human Groups and Social Categories: Studies in Social Psychology. Cambridge: Cambridge University Press.

Tajfel, H. (1982). Socal Psychology of Intergroup Relations. Annual Review Of Psychology, 33(1), 1-39.

Tajfel, H. & Turner, J. C. (1979). An Integrative Theory of Intergroup Conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations. Monterey, CA: Brooks-Cole.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987). Rediscovering the Social Group: A self-categorization theory, Oxford and New York: Blackwell.

Weber Shandwick, Powell Tate & KRC Research (2014). Civility in America. Retrieved from

Young, L., & Soroka, S. (2012). Affective News: The Automated Coding of Sentiment in Political Texts. Political Communication, 29(2), 205-231. doi:10.1080/10584609.2012.671234.

About | Issues
© NMEDIAC & individual NMEDIAC authors, editors, and programmers
home issues