Website content persistence and change: Longitudinal analysis of pro-white group identity

Abstract

Despite the ability of websites to quickly evolve, little attention has been paid to persistence and change in site content. Longitudinal examination of 163 pro-white advocacy group websites, in which establishing a core group identity is a critical strategic goal, showed a half-life of 2.40 years and 34% remained active after five years. Analysis of text content from 28 sites collected annually from 2007 to 2012 (n = 1947) showed that persistence was more likely for advocacy group identity, while examples of group goals were transient. Content persistence trends reflect broader phenomena of ideologically oriented website persuasive material.

Keywords

advocacy group content analysis group identity persistence pro-white website

1. Introduction

One advantage of websites as a persuasive and informative communication tactic lies in their adaptability to changing circumstances, incorporating new ideas and examples. Yet few scholars have looked at change in content that is central to the site’s purpose, instead focusing on the survival rate of specific web pages [1 –3] or structural attributes like internal links and applications [4]. The adaptability of web content can be particularly useful for ideologically oriented advocacy groups to mobilize others or sway public opinion. Computer-mediated communication has low costs [5] and potentially wide reach. Websites may hold particular value for activists pursuing goals outside the mainstream, providing ‘free space’ for open communication [6].

Considering the importance of websites in building online communities [7, 8] and the utility of content in strategic framing processes [9] for advocacy groups to mobilize others, examining content change is important in understanding persuasive communication strategies. Sites that are static – that is, always feature the same content – offer adherents little reason to visit a site once they have absorbed the content. By contrast, sites that regularly change may spur subsequent visits, further triggering mobilization by offering new opportunities to participate, information on the latest activities and examples of what advocates stand for or attempt to change.

In an effort to understand web persistence, this study adopts a longitudinal approach focused on site persistence, content persistence and change. The context is pro-white advocacy groups, which have a long history of promoting racial intolerance and mobilizing others to their cause, including a tradition of using computer-mediated communication. Websites of pro-white groups in the USA that were active in 2007 (n = 163) formed the population, with visits in the subsequent five years quantifying site persistence. Content persistence and change were analysed by examining text content of a sample (28 sites, 1947 articles) over those six time periods to see what content remained on sites, what was removed and what was added.

2. Conceptual framework

Little scholarly attention has gone into explaining the ephemeral nature of websites and web content. Researchers are still gathering evidence about web and content persistence and this analysis typifies web persistence and change by examining a particular web sphere [10] of pro-white activist sites over time. Evidence on site and content persistence, continuity and change is examined, followed by literature on the contextual elements of advocacy groups in general and pro-white groups in particular.

2.1. Website persistence and change

Most scholarship on web page persistence and change comes from studies using web crawlers to examine URLs. While these findings are useful in speaking about overall online patterns, few examine specific web spheres, nor do they analyse changes in content. Some focus upon a specific realm of content, defined as synchronic studies [11], while others are diachronic, for studies of random websites. Sites change in several basic ways – sites disappear altogether or reappear at a different address, pages are deleted, new pages are created and existing pages are revised. Thus, longitudinal studies are important for understanding web processes [12].

Studies of website persistence illustrate the dynamic nature of computer-mediated communication. Scholars use terms like ‘link rot’, ‘half-life’ and ‘decay’ to describe the ephemeral nature of websites and pages that disappear or persist [3]. Several studies have looked at web pages over time, including a diachronic study of 360 webpages [3], which found that 33.8% of URLs selected in 1996 remained active after more than six years. Additionally, 31.8% of the web pages failed to respond after one year [13] and Koehler [14] calculated that, every two years, half of the pages being tracked could not be accessed (that is, a half-life of two years). A longitudinal diachronic study [4] compared 921 ‘high quality’ websites over seven years, finding that 19.8% disappeared, 46.3% changed addresses and the remaining 33.7% remained constant. Additionally, analysis of web pages devoted to informetrics over an eight-year period found that 36.3% of the 866 URLs persisted after five years and 19.1% still existed after eight years [12, 15].

A common type of synchronic investigation examines citations in academic journals, which is especially important considering that contemporary scholarship may be built upon transient information sources. Dimitrova and Bugeja examined citations published between 2000 and 2003 in five communication journals [2] and the Journal of Computer-Mediated Communication [1]. The study of five journals traced 1126 URL addresses in citations, finding that 36% worked as promised in 2004, a steady decline in functioning URLs over time, and calculated the average half-life at 3.17 years. The Journal of Computer-Mediated Communication study estimated that 36.4% of the URLs cited in the journal were inaccessible by 2005. Citations in other academic areas were explored as well, including scientific and medical journals [16], library and/or information science journals [17, 18], health management journals [19] and academic conference articles [20]. Overall, those studies found reason for concern in continuity of scholarship, as many URLs were no longer accessible shortly after publication and the number grew over time. The first research question addresses site persistence:

RQ1: How many pro-white websites active in 2007 remained active one, two, three, four and five years later?

2.2. Content persistence and change

While a focus of this study is on change in content, most previous studies have dealt with whether pages changed or with general types of content. For example, a six-year investigation of 360 web pages [3] found that just 3% of pages had not changed after the first year, but the longer a page remained, the less likely it was that it had changed. An 11-week study of 151 million web pages [21] found that 88% of pages were available during the final crawl, 63% had not changed and large pages were more likely to change than smaller ones. Analysis of content change found a sizeable increase in internal links, applications (files, documents, Java applets) and images, and a slower growth in pages, multi-media content and external links [4]. The second research question addresses site content persistence and change:

RQ2: How much of pro-white website content remained on the site in subsequent years?

2.3. Advocacy groups and websites

Websites serve multiple strategic purposes for advocacy groups, including mobilizing members, attracting potential adherents and communicating beliefs to outsiders. Web activism includes three basic components [22]: e-mobilization, for the ability to mobilize others through motivational calls to action, logistics and information on the organizers; e-movements, for movements that exist primarily online; and e-tactics, for types of activism that are accomplished online, such as petitions. The lower costs and obstacles of computer-mediated communication are particularly beneficial for ideologically oriented activists outside the boundaries of traditional institutions [5]. Websites allow groups to have a public presence – professional-looking sites are easy to construct and make the groups appear legitimate – while also remaining private, as potential recruits can scan material at home. Pro-white groups use the ‘free space’ [6] on websites to advocate intolerance against minorities, potentially attracting a large audience.

Web presence can build a community of like-minded individuals. Online social networks eliminate spatial constraints and re-define community in social rather than spatial terms [23]. Similarly, online groups sharing ideological interests are closely connected and result in a cyber community [7]. The major objective of pro-white and other extremist websites is to share ideology, thereby creating a socially constructed community [24]. Online communities provide members with a sense of a common, collective purpose.

Web presence allows ideas outside the mainstream to flourish and facilitates the free expression of ideas [25]. Extremist views can find a safe place for expression on the Internet [26, p. 465] by developing a ‘network of web users with a broadly similar, yet highly unconventional and controversial, shared outlook on the topic’. Pro-white groups use the Internet to unite individuals with similar beliefs and spread controversial and socially constrained ideologies [8]. Pro-white groups use websites to express adversarial and controversial ideas (see [25, 27, 28]).

Advocacy groups can use websites to communicate with potential adherents through framing processes (e.g. [9, 29]). Groups engage in strategic efforts to communicate common understandings of the world, including identification of a problem or threat, and their ideology as steps to help create a collective identity. Schurman and Munro [30, p. 4] explain: ‘Indeed, for all social movements, but particularly for those that can be characterized as “quality-of-life”-type movements, grievances have to be not only identified, but analytically constructed and articulated before collective action is imaginable.’

2.4. Pro-white groups and website content

The growing realm of pro-white groups (activists that typically push an agenda favouring ‘majority’ rights or opposing racial minorities) has received public attention in the USA, especially since the election of President Barack Obama in 2008. Organizations like the Southern Poverty Law Center and individuals (e.g. [31]) monitor pro-white websites and other forms of computer-mediated communication, with estimates that there are at least 1000 sites [31]. Critics worry that such sites raise intolerance and threaten violence, and represent a form of domestic terrorism. For instance, shortly after Obama was inaugurated as President in 2009, race-motivated incidents of violence were reported [32, 33], with some perpetrators linked to pro-white websites [34]. Pro-white groups have been classified as domestic terrorist organizations by the Department of Homeland Security [35] for their potential to trigger violence.

Although scholars since the mid-1990s have studied pro-white websites, most examined site content at one time, thus not considering change (an exception was [36], which downloaded material over one year). At least 16 articles published between 1997 and 2011 focused on the content of websites created by such groups. The studies used a wide range of approaches, including social identity theory [37], the nature of the content and site characteristics [38 –40], discursive/rhetorical analyses [36, 41 –43], communication techniques [44, 45], sale of merchandise [46] and examination of hyperlinks [7, 24, 47].

While such sites are considered threats to the social order by watchdog groups and potential terrorist threats by the US government [35], sites that are ephemeral or static would seem to be less of a threat than sites that persist and/or change frequently. A site that cannot be accessed at its traditional URL may have gone or may be difficult to locate at a new address. Sites that consist of the same content will be of less interest to adherents over time than a site that is updated.

2.5. Themes in website content

Initial examination of pro-white web content revealed two broad themes: (1) primarily about the group and (2) examples of what the group supports or opposes. These broad themes may be shared by other advocacy groups, which seek to establish legitimacy and credibility, to demonstrate that the group can effectively address social ills. Evidence of legitimacy and credibility may be found in themes that establish a group identity, including beliefs, history and membership. These themes appear to be fairly common among a wide range of advocacy groups on websites; examination by the author of other types of ideological groups (including environmentalists, groups fighting indecency in entertainment media and pro-anorexia groups) found similar content about the groups’ beliefs, history and membership. Beliefs demonstrate a core ideology, giving visitors a sense of the group’s vision, perhaps suggesting that the beliefs are built upon reason and logic. History adds to legitimacy by showing that the organization or its goals have roots. Items about membership show that the group is more than just one person, suggesting it has broad appeal and also may include opportunities for others to join.

Examples of what a group supports or opposes are also shared by other types of advocacy groups, although the specific targets vary by group ideology. Advocacy groups often attribute threats to an external entity, which may be contained within grievances, described as a ‘strain’ triggering psychological disturbances leading to action (for an overview, see [48], pp. 6–11). Strains include identification of a societal problem, attributing that problem to someone or something, and evidence of a continued threat. Threats or grievances have been found in a wide range of ideologically movements, including civil rights [49] and women’s rights [50]. Further, relative deprivation theory proposes that identification with a group mobilizes people because they experience discontent when they perceive their own group to be disadvantaged relative to another [51].

Within the realm of pro-white groups, analysis of print and video documents featuring white supremacists [52, p. 187] found victimization to be an enduring theme of the material and suggested that it connected with ‘a broader sentiment of the victimized white male to which white supremacists adhere and add, but which they also utilize and exploit’. Similar themes reverberate throughout the history of groups like the Ku Klux Klan. For instance, the early history of the Klan shows that they lashed out against Carpetbaggers in the South. The Klan singled out ‘inferior’ immigrants as responsible for economic problems and portrayed Catholics as threatening morality and engaging in evil acts [53]. Although previous studies did not explicitly apply an attribution label, evidence of such themes is present. For instance, Zickmund [43] described how targets, chiefly African-Americans, homosexuals and Jews, were identified and described as threats and Bebrier [52] found that white supremacists fear that survival of the white race is threatened by social factors, including a relatively low birth rate and massive non-white immigration.

Thus, grievances, threats, repression and victimization can be communicated in group actions and examples of persecution. Based upon the expectation that content themes like beliefs, history and membership, as part of the group’s core identity, are relatively stable, while content providing examples of what the group supports or opposes will be more transient, the following hypothesis is presented:

H1: Web site themes that reflect the group’s core identity are more likely to persist over time than themes that reflect examples of what the group stands for or opposes.

3. Method

Two layers of material from pro-white websites were analysed. Site persistence was measured by comparing the number of pro-white sites active in 2007 with those still active one, two, three, four and five years later. Continuity and change of web content were measured by analysing text articles from 28 pro-white sites in 2007 with content on the same sites in 2008, 2009, 2010, 2011 and 2012.

3.1. Website persistence

The population consisted of 163 pro-white websites active on 15 February 2007. The population was developed by examining two directories [31, 54] that published URLs and following those links to determine active sites. Sites open to the public and containing pro-white material were considered active; sites opening to a discussion board were excluded. Four types of groups identified by the SPLC were analysed and the population consisted of 51 Ku Klux Klan sites (31%), 65 Neo-Nazi sites (40%), 30 Neo-Confederate sites (18%) and 17 Skinhead sites (10%). A second coder visiting the sites two weeks later found the same distribution of active sites, except that one site that was active on the first visit was no longer active two weeks later.

Subsequent visits to each site active in 2007 took place on the same date (or within one day) in 2008, 2009, 2010, 2011 and 2012. Sites that contained pro-white material or automatically forwarded visitors to another URL with pro-white content were considered active. URLs that did not link to pro-white content were considered inactive and inactive sites were checked again within a week. A second coder, also checking on 15 February 2008, found the same distribution of active sites as the first coder in 2008. Considering the nearly perfect confirmation in 2007 and 2008, no second coder was used in 2009, 2010, 2011 and 2012.

3.2. Content analysis

Among 163 sites active in February 2007, 28 were randomly selected for analysis of text content. The sample was representative of the population, with eight KKK sites (29%), 10 Neo-Nazi sites (36%), six Neo-Confederate sites (21%) and four Skinhead sites (14%). Group names and URLs are given in the Appendix.

The unit of analysis was the text article, defined as a discreet body of information with a clear beginning and ending that was five or more sentences long, typically with larger text (a headline) at the beginning. Cursory examination of several sites revealed that some contained hundreds of articles. To control the number of units from a site, the following rubric was used: all text articles on the home page and any page with a link off the home page (second pages) were coded. Articles on third pages (links off second pages) were coded until the sample from that site reached 100. This represents what a casual visitor would be likely to encounter. The coders worked top-left to bottom-right to ensure they were using the same procedure to access content. Articles were copied into text documents for later analysis. Chat, bulletin board and discussion room posts were excluded, primarily because those required a password to enter. The sample was confined to material within that specific web site; hyperlinks that left the site, such as to a news article, were excluded.

Content was coded by two trained researchers, who entered a brief description of the main article theme, which was then used to create specific categories. As anticipated, many of the articles fell into categories of group identity (beliefs, membership and history). The remaining descriptions were condensed into themes that reflect examples or comments about topics that reflect group goals: actions, politics, persecution/repression and other/miscellaneous. A third coder examined 148 articles to assess the reliability of the quantitative article theme variable and the measure was reliable (Krippendorf’s α = 0.84).

4. Results

In response to RQ1, the persistence of pro-white websites was examined by checking the 163 URLs active in 2007 again in 2008, 2009, 2010, 2011 and 2012. The number of active sites steadily declined: 109 (66.8%) were active and contained pro-white content when accessed in 2008, 95 (58.3%) in 2009, 76 (46.6%) in 2010, 65 (39.9%) in 2011 and 56 (34.4%) in 2012. The pro-white sites had a half-life of 2.40 years and 48 sites (29.4%) were active on each yearly check. After five years, 28% of Ku Klux Klan sites were active, 45% of Neo-Nazi sites, 30% of Neo-Confederate sites and 24% of Skinhead sites.

Sites that survived from 2007 to 2008 were somewhat more likely to persist after that, similar to what Koehler [3] found. Although just two-thirds survived year one to year two, the subsequent rates were much higher: 87.2% year two to year three; 80% from year three to year four; 85.6% from year four to year five; and 86.2% from year five to year six. Once sites became inactive, they tended to stay inactive; of 104 sites inactive at any check, 92 remained inactive at each subsequent visit (88.5%). Several went against that trend: seven of 54 sites that were no longer active in 2008 were active in 2009, five of 62 inactive sites in 2009 were active in 2010, and one site inactive in 2010 was operational in 2011. In 2012, no sites previously inactive had become active.

Content persistence (RQ2) was examined by looking at change in content in five periods, each one year apart. Overall, 1947 text articles were analysed over six time periods, which at the last check included 410 present, 864 that had been removed and 673 from inactive sites. On the initial check in 2007, 937 articles were found on 28 active sites (M = 33.5). In 2008, 617 articles (358 from 2007 and 259 new articles) were found on 19 active sites (M = 32.5). In 2009, 531 articles (271 from 2007, 80 from 2008 and 180 new articles) were found on 15 active sites (M = 35.4). In 2010, 353 articles (110 from 2007, 59 from 2008, 29 from 2009 and 157 new articles) were found on 12 sites (M = 29.4). In 2011, 398 articles (92 from 2007, 53 from 2008, 18 from 2009, 33 from 2010 and 202 new articles) were found on nine sites (M = 44.2). In 2012, 410 articles were found on eight sites (97 from 2007, 2 from 2008, 4 from 2009, 13 from 2010, 82 from 2011 and 212 new articles; M = 51.3).

Articles progressively disappeared over time: the 937 articles found in 2007 dropped steadily to 358, 271, 110 and 92 in subsequent years, then up slightly to 97 in the final year; 259 new articles were found in 2008, decreasing to 80, 59, 53 and 2; 180 new articles found in 2009 dropped to 29, 18 and 4. Combining those three starting years, just 33.9% of the articles were accessible after one year, 25.3% after two years and 13.6% after three years. The articles available on the initial check in 2007 – which could have been on the site for several years – showed somewhat more persistence than those added in subsequent years.

Analysis by site showed distinct patterns. First, among the eight sites active in 2012, three were static and five varied considerably. Among the static sites, New Christian Crusade Church had the same 77 articles present on each yearly check, Haralson’s Invincibles Sons of Confederate Veterans had three of its four articles available at each check (the other was present just in 2009) and Connecticut White Power Skinheads had five of its seven articles present each year (the other two were present in 2007 and 2008 only).

At the other extreme were the three most prolific groups: Nationalist Movement (549 articles), National Socialist Movement (306 articles) and National Vanguard (253 articles). Each of those groups added new content on a regular basis, had at least 100 articles in 2012 and retained little (just two articles for National Socialist Movement and three for National Vanguard survived since 2007). Among active articles in 2012, 98% for Nationalist Movement, 97% for National Vanguard and 83% for National Socialist Movement first appeared in either 2011 or 2012. A fourth group with variability was Empire Knights of KKK, one of the few in which the site went defunct (in 2009) yet returned. Empire Knights of KKK had 26 articles present in 2007, of which seven were active in 2012; two were present in 2008 and 2012, and 16 first appeared in 2012.

To examine change in themes, the content present in one year was compared with its status in the subsequent year. This rolling approach to content persistence evaluated theme survivability over time. Five separate comparisons of survivability were analysed: from 2007 to 2008, from 2008 to 2009, from 2009 to 2010, from 2010 to 2011 and from 2011 to 2012. Persistence was calculated two different ways: (1) comparing themes representing the group’s core identity (beliefs, history and membership) with examples of group goals or efforts (actions, politics, persecution/repression and other) and (2) within each specific theme.

Both approaches overall supported H1. Themes reflecting group identity were significantly more likely to persist than themes reflecting examples of what the group stands for or opposes in each of the year-to-year comparisons (see Table 1). The year-to-year persistence of core identity ranged from 63 to 82%, while the examples theme ranged from 25 to 51%.

Table 1.

Frequency of theme persistence

	Core identity	Examples	Totals	χ ²
2007–2008	76.9% (n = 333)	35.5% (n = 287)	57.7% (n = 620)	107.95***
2008–2009	73.9% (n = 310)	51.1% (n = 225)	64.3% (n = 535)	29.42***
2009–2010	82.1% (n = 162)	41.1% (n = 158)	61.9% (n = 320)	56.89***
2010–2011	71.6% (n = 169)	41.2% (n = 170)	56.3% (n = 339)	31.89***
2011–2012	63.4% (n = 186)	24.8% (n = 206)	43.1% (n = 392)	59.64***

***

p ≤ 0.001.

Figures in each cell represent the frequency of that type of article that persisted from previous year and the number of articles of that type from previous year. Thus, figures in ‘2007 to 2008’ are percentages of articles present in 2008 among those present in 2007.

Analyses by specific themes showed that persistence came primarily from articles about group beliefs, which persisted significantly more often in each of the five time periods (see Table 2), ranging as high as 83%. Of the other core identity themes, history persisted more often in two comparisons and was not significant in the other three, while membership persisted more often in one comparison, less often in two and was not significant the other two. Among the examples, the persistence rate was significantly lower than average. Persecution persisted less often in four of the five comparisons, and actions, politics and other/miscellaneous in two each. Politics persisted more often in one comparison and all other comparisons were not significant. Of the two most prolific sites, one largely matched the overall trend. For National Socialist Movement, analysis showed more persistence for core beliefs than examples in three of the five year-to-year comparisons (the other two were not significant). By contrast, Nationalist Movement had more persistence for examples in three of the comparisons, with the other two not significant.

Table 2.

Frequency of theme persistence, by specific theme

	2007–2008	2008–2009	2009–2010	2010–2011	2011–2012
Beliefs	74.2% (n = 221)	72.4% (n = 192)	83.4% (n = 135)	78.5% (n = 135)	69.3% (n = 153)
	38.16***	8.55**	50.42***	44.85***	70.07***
Members	73.1% (n = 26)	37.5% (n = 24)	66.7% (n = 9)	27.8% (n = 18)	37.5% (n = 8)
	2.62†	7.86**	0.09	6.31**	0.11
History	84.9% (n = 86)	86.2% (n = 94)	72.2% (n = 18)	62.5% (n = 16)	36.0% (n = 25)
	30.15***	23.76***	0.87	0.26	0.55
Actions	64.3% (n = 70)	60.3% (n = 63)	31.3% (n = 32)	17.1% (n = 35)	32.0% (n = 25)
	1.39	0.49	14.14***	24.38***	1.34
Politics	51.3% (n = 39)	36.0% (n = 25)	9.4% (n = 32)	62.5% (n = 8)	62.5% (n = 16)
	0.71	9.15**	41.54***	0.13	2.56†
Persecution	19.4% (n = 160)	52.1% (n = 119)	56.6% (n = 83)	47.3% (n = 112)	16.9% (n = 142)
	130.10***	9.92***	1.31	5.53*	62.37***
Other	33.3% (n = 18)	33.3% (n = 18)	45.5% (n = 11)	40.0% (n = 15)	39.1% (n = 23)
	4.53*	7.78**	1.30	1.70	0.16
Totals	57.7% (n = 620)	64.3% (n = 535)	61.9% (n = 320)	56.3% (n = 339)	43.1% (n = 392)

***

p ≤ 0.001; ** p ≤ 0.01; * p ≤ 0.05; † p ≤ 0.10.

Figures in each cell represent: (1) the frequency of that type of article theme that persisted from previous year; (2) the number of articles of that type from previous year; and (3) the χ² figure for that theme, compared with all other types.

For example, the cell ‘beliefs’ in ‘2007–2008’ indicates that 74.2% of the 221 articles about beliefs that were present in 2007 were still present in 2008 for a χ² figure of 38.16, which was significant at p ≤ 0.001. The ‘totals’ figures on the bottom row show that 57.7% of 2007 articles survived to 2008, so the ‘beliefs’ frequency was significantly higher when compared with all other themes combined.

5. Discussion

Analysis demonstrated that websites of pro-white groups and content varied considerably in persistence and change. After five years, 34.4% of the sites remained active as pro-white sites. Text content on the sites changed as well, with just 10.4% of content present in 2007 still available in 2012, either because the site was defunct or the article had been removed. Analysis of article themes showed that material primarily about the core group identity persisted, while examples of what the group supports or opposes were more transient.

Longitudinal analyses reveal useful patterns among advocacy group websites. Perhaps most importantly, analyses showed that group identity helps explain article persistence, a potential building block towards identifying patterns of web persistence and change among ideologically oriented material. These findings also are interesting from a societal standpoint, especially those concerned that pro-white websites pose a threat of violence or terrorism.

5.1. Longitudinal analysis

Over-time analysis revealed volatility among pro-white sites. After one year, nearly one-third of the sites were inactive or had switched form, and after five years, about one-third remained active. Pro-white sites (2.40 half-life) persisted at a higher rate than random sites [3, 14], in which half of all URLs disappeared after two years, but at a lower rate than academic URLs [2], which had a half-life of 3.17 years. It is not surprisingly that the persistence rate of non-mainstream advocacy group sites was lower than those in academic citations; more surprising, perhaps, is that pro-white sites persisted at a rate slightly higher than random sites. With pro-white groups categorized by US Homeland Security as potential terrorist threats and claims that pro-white sites fuel racial intolerance and have experienced more visitor traffic since Obama became President, the persistence rate of such sites provides useful context. That is, calculating the sheer number of pro-white sites at any one time may be less important than understanding which sites persist, for sites that remain intact over time offer a larger potential threat to the social order than sites that disappear. Future research can delve into characteristics that help explain site persistence.

The 28 sites closely analysed revealed interesting overall patterns. First, sites that started with more text articles were somewhat more likely to persist than sites with fewer text articles. Based upon the number of articles on the first check, the average number per group climbed from 33.5 in 2007 to 58.6 in 2012. However, the five sites that remained intact throughout the five-year period fell into two distinct camps – two low-content sites probably operated by one person and three with extensive content from many contributors. A low-volume site is Haralson’s Invincibles Sons of Confederate Veterans, in which three of its four articles were available at each check. The high-volume category is typified by The Nationalist Movement, in which hundreds of articles were archived and the home-page content showed regular updates.

Paying attention to content that persists as compared with new content is important in assessing the potential impact of advocacy group websites. It is difficult to expect visitors to return to sites in which content never changes. By contrast, sites that change regularly offer new ideas and examples that potentially mobilize adherents. Sites that consistently update content may be a more fruitful target for those worried about the potential threat of pro-white sites. Some groups simply have greater capacity to change – that is, such sites may have the personnel (members, webmasters, contributors) to create or find new content and/or greater server space. While findings here are limited because content was drawn from just a portion of pro-white sites, future research could capture content from a larger set of sites to more precisely quantify these trends.

5.2. Persistence of advocacy group identity

Analysis demonstrated that core group identity was much more likely to persist than were examples of group goals. The persistence of group identity offers insights to a wide range of advocacy sites, as ideologically driven groups may share similar patterns of identity and goals in web content. This makes sense from an advocacy group persuasion perspective, as content reflecting group legitimacy persists while new examples replace old examples to give visitors new grievances to ponder. Material about group beliefs demonstrates an ideology based upon reason. Such material would rarely change, and thus can offer a constant presence on the site. By continually adding new examples, pro-white groups give regular visitors a reason to return to the site and offer fresh instances of why people should care, perhaps be outraged and maybe act. Similar reasoning should apply to other advocacy groups as well.

The patterns seem applicable for a wide range of activists. The balance between content that is relatively permanent and that which is fleeting appears to be a good fit for advocacy groups in which ideology and group identity are critical. While this seems a particularly good fit for pro-whites (e.g. [7, 8]), such trends may be applicable to a wider population of activists in which ideology and identity matter, especially those outside the mainstream [55]. Exemplars used to defend and illustrate group goals are more transitory. The permanence of the identity and ideology can be aimed at those interested in what the group stands for and secondarily as a reminder to long-term members. Transitory material reminds adherents of issues posing a continued threat to their goals. Advocacy organizations need a fresh supply of new examples to give adherents both motivation to return to the site and a new reason to get concerned. Future research could confirm these patterns by examining other realms of advocacy group persuasive content.

In conclusion, the patterns found in this analysis form useful building blocks in understanding web persistence. By focusing upon a single web sphere [10], these findings reflect important characteristics for understanding strategic use of web content among like-minded advocates. Two characteristics of site persistence are worth attention: (1) websites with more content are somewhat more likely to persist; and (2) rather than showing a steady decline, the rate of persistence increases over time. Perhaps more critical, however, is analysis of content persistence: material about core group identity is more likely to persist than examples of what stands for or opposes. This latter point is useful evidence toward identifying patterns of web persistence.

Footnotes

Appendix

Appendix:

Group name, URL and status as of 15 February 2012 of sample (28 pro-white websites) used in content analysis

Group name, by group type	URL (active in 2007)	Status in 2012
Ku Klux Klan
American White Knights	http://www.awkkkk.org/	Inactive
Brotherhood of Klans International	http://www.ku-klux-klan.org/	Inactive
Church of the National Knights of KKK	http://www.cnkkkk.com/	Inactive
Knights of KKK, Florida	http://www.pagerealm.com/florida/KnightsFla.htm	Inactive
Louisiana White Knights of the KKK	http://www.lwkkkk.org/	Inactive
Empire Knights of KKK	http://www.knights311.org/	Active
Knights of KKK, Tennessee	http://www.newporttennessee.net/	Inactive
White Country Rebel	http://www.rebelarmy.com/	Active
Neo-Nazi
Church of Jesus Christ Christian	http://twelvearyannations.com/	Inactive
Cosmotheism	http://cosmotheism.net/	Inactive
Final Solution	http://www.finalsolution88.com	Inactive
National Socialist Movement	http://www.nsm88.com/	Active
National Vanguard	http://nationalvanguard.org/	Active
American National Socialist Workers’ Party	http://www.nazisozi.com/	Inactive
Aryan Confederation	http://www.geocities.com/arcon_wp/	Inactive
New Christian Crusade Church	http://www.newchristiancrusadechurch.com/	Active
Wake Up or Die	http://www.wakeupordie.com/	Inactive
National Socialist Movement NW	http://www.nukeisrael.com/	Inactive
Neo-Confederate
Southern Party of North Carolina	http://www.southernpartync.com/entry.php?SessID=11791	Inactive
The Heritage Project	http://home.comcast.net/~historyproject/	Inactive
The Red Shirts	http://www.redshirts.org/	Inactive
Florida League of South/Suwannee Valley	http://cloud.prohosting.com/fllos/	Inactive
Haralson’s Invincibles Sons of Confederate Veterans	http://cbearhamrick.tripod.com/	Active
Palmetto Studies	http://www.palmetto.org/	Inactive
Skinheads
Nationalist Movement (Crosstar)	http://www.nationalist.org	Active
Warriors of the White World	http://www.warriorsofthewhiteworld.com/	Inactive
Confederation of Racialist Working Class Skinheads	http://www.crwskinheads.com/	Inactive
Connecticut White Power Skinheads	http://www.angelfire.com/ct3/arise14/	Active

References

Dimitrova

Bugeja

. Exploring the use of online citations in an online-only journal: A case study of the Journal of Computer-mediated Communication. Journal of Computer-mediated Communication 2009: 19, http://www.cios.org/ejcpublic/019/1/019126.HTML

Dimitrova

Bugeja

. The half-life of internet references cited in communication journals. New Media and Society 2007: 9: 811–826.

Koehler

. A longitudinal study of web pages continued: A report after six years. Information Research 2004: 9: paper 174.

Ortega

Aguillo

Prieto

. Longitudinal study of content and elements in the scientific web environment. Journal of Information Science 2006: 32: 344–351.

Van Aelst

Walgrave

. New media, new movements? The role of the Internet in shaping the ‘anti-globalization’ movement. Information, Communication and Society 2002: 5: 465–493.

Simi

Futrell

. Cyberculture and the endurance of white power activism. Journal of Political and Military Sociology 2006: 34: 115–142.

Burris

Smith

Strahm

. White supremacist networks on the internet. Sociological Focus 2000: 33: 215–235.

Chau

. Mining communities and their relationships in blogs: A study of online hate groups. International Journal of Human–Computer Studies 2007: 65: 57–70.

Snow

Burke Rochford

Jr Worden

Benford

. Frame alignment processes, micromobilization and movement participation. American Sociological Review 1986: 51: 464–481.

10.

Schneider

Foot

. Web sphere analysis: An approach to studying online action. In: Hine

(ed.) Virtual methods: Issues in social research on the internet. Oxford: Berg Publishers, 2005, pp. 157–170.

11.

Tyler

McNeil

. Librarians and link rot: A comparative analysis with some methodological considerations. Portal: Libraries and the Academy 2003: 3: 615–632.

12.

Bar-Ilan

Peritz

. Evolution, continuity, and disappearance of documents on a specific topic on the web: A longitudinal study of ‘informetrics’. Journal of the American Society for Information Science and Technology 2004: 55: 980–990.

13.

Koehler

. An analysis of web page and web site constancy and permanence. Journal of the American Society for Information Science 1999: 50: 162–180.

14.

Koehler

. Web page change and persistence – a four-year longitudinal study. Journal of the American Society for Information Science and Technology 2002: 53: 162–171.

15.

Bar-Ilan

Peritz

. The lifespan of ‘informetrics’ on the web: An eight-year study (1998–2006). Scientometrics 2009: 79: 7–25.

16.

Dellavalle

Hester

Heilig

Drake

Kuntzman

Graber

Schilling

. Going, going, gone: Lost Internet references. Science 2003: 302: 787–788.

17.

Casserly

Bird

. (2008). Web citation availability: A follow-up study. Library Resources and Technical Services 2008: 52: 42–53.

18.

Goh

DHL

. Link decay in leading information science journals. American Society for Information Science and Technology 2007: 58: 15–24.

19.

Wagner

Gebremichael

Taylor

Soltys

. Disappearing act: Decay of uniform resource locators in health care management journals. Journal of Medical Library Association 2009: 97: 122–130.

20.

Sellitto

. The impact of impermanent web-located citations: A study of 123 scholarly conference publications. Journal of the American Society for Information Science and Technology 2005: 56: 695–703.

21.

Fetterly

Manasse

Najork

Wiener

. A largescale study of the evolution of web pages. Procedures of the 12th International WWW Conference, 2003.

22.

Earl

Kimport

. Digitally enabled social change: Activism in the Internet age. Cambridge, MA: MIT Press, 2011.

23.

Wellman

. Community: From neighborhood to network. Communications of the ACM 2005: 48: 53–55.

24.

Zhou

Reid

Qin

Chen

Lai

. US domestic extremist groups on the web: Link and content analysis. IEEE Intelligent Systems 2005: 20: 44–51.

25.

Glaser

Dixit

Green

. Studying hate crime with the Internet: What makes racists advocate racial violence? Journal of Social Issues 2002: 58: 177–193.

26.

Giles

. Constructing identities in cyberspace: The case of eating disorders. British Journal of Social Psychology 2006: 45: 463–477.

27.

Lee

Leets

. Persuasive storytelling by hate groups online: Examining its effects on adolescents. American Behavioral Scientist 2002: 45: 927–957.

28.

Levin

McDevitt

. Hate crimes: The rising tide of bigotry and bloodshed. New York: Plenum, 1993.

29.

Snow

D A

Benford

. Master frames and cycles of protest. In: Morris

Mueller

(eds) Frontiers in social movement theory. New Haven, CT: Yale University Press, 1992, pp. 133–155.

30.

Schurman

Munro

. Ideas, thinkers, and social networks: The process of grievance construction in the anti-genetic engineering movement. Theory and Society 2006: 35: 1–38.

31.

Franklin

. The Hate Directory, release 10.2, http://www.bcpl.net/~rfrankli/hatedir.htm (2006, accessed January 2007).

32.

Asbury

. Attacks linked to election of Obama, police say. The Press Enterprise, 18 January 2009, C1.

33.

Pugh

. In autumn of America’s discontent, Obama’s race emerges as an issue. McClatchy-Tribune News Service, 18 September 2009.

34.

White supremacists: Dangers of hate. Omaha World-Herald, 16 May 2009, 4B.

35.

Tough times may be right for surge in extremists; Homeland Security warns right-wing radicals may be seeking more recruits. Grand Rapid Press, 15 April 2009, A1.

36.

Bostdorff

. The internet rhetoric of the Ku Klux Klan: A case study in website community building run amok. Communication Studies 2004: 55: 340–361.

37.

Douglas

McGarty

Bliuc

Lala

. Understanding cyberhate: Social competition and social creativity in online white supremacist groups. Social Science Computer Review 2005: 23: 68–76.

38.

Adams

Roscigno

. White supremacists, oppositional culture and the World Wide Web. Social Forces 2005: 84:759–778.

39.

McDonald

. CyberHate: Extending persuasive techniques of low credibility sources to the World Wide Web. In: Schumann

Thorson

(eds) Advertising and the World Wide Web. Mahwah, N.J.: Lawrence Erlbaum Associates, 1999, pp. 149–157.

40.

Meddaugh

Kay

. Hate speech of ‘reasonable racism?’ The Other in Stormfront. Journal of Mass Media Ethics 2009: 24: 251–268.

41.

Duffy

. Web of hate: A fantasy theme analysis of the rhetorical vision of hate groups online. Journal of Communication Inquiry 2003: 27: 291–312.

42.

McNamee

Peterson

Pena

. A call to educate, participate, invoke and indict: Understanding the communication of online hate groups. Communication Monographs 2010: 77: 257–280.

43.

Zickmund

. Approaching the radical other: The discursive culture of cyberhate. In: Jones

(ed.), Virtual culture: Identity and communication in cybersociety. London: Sage Publications, 1997, pp. 185–205.

44.

Schafer

. Spinning the web of hate: Web-based hate propagation by extremist organizations. Journal of Criminal Justice and Popular Culture 2002: 9: 69–88.

45.

Weatherby

Scoggins

. A content analysis of persuasion techniques used on white supremacist websites. Journal of Hate Studies 2005-06: 4: 9–31.

46.

Gerstenfeld

Grant

Chiang

. Hate online: A content analysis of extremist Internet sites. Analyses of Social Issues and Public Policy 2003: 3: 29–44.

47.

Reid

Chen

. and Internet-savvy U.S. and Middle Eastern extremist groups. Mobilization: An International Quarterly Review 2007: 12: 177–192.

48.

McAdam

. Political process and the development of black insurgency, 1930–1970, 2nd edn. Chicago, IL: University of Chicago Press, 1999.

49.

Morris

. The origins of the civil rights movement: Black communities organizing for change. New York: Free Press, 1984.

50.

Taylor

. Social movement continuity: The women’s movement in abeyance. American Sociological Review 1989: 54: 761–775.

51.

Veenstra

Haslam

. Willingness to participate in industrial protest: Exploring social identification in context. British Journal of Social Psychology 2000: 39: 153–172.

52.

Berbrier

. The victim ideology of white supremacists and white separatists in the United States. Sociological Focus 2000: 33: 175–191.

53.

McVeigh

Myers

Sikkink

. Corn, Klansmen, and Coolidge: Structure and framing in social movements. Social Forces 2004: 83: 653–690.

54.

Active hate websites. Intelligence Report 2005: 117: 59–65.

55.

Kriesi

Koopmans

Duyvendak

Giugni

. The politics of new social movements in Western Europe. Minneapolis, MN: University of Minnesota Press, 1995.