Normalizing or equalizing? Characterizing Facebook campaigning

Abstract

The article analyzes whether Facebook campaigning is consistent with the Normalization or Equalization hypothesis, drawing on data from the election campaigns for the 20th Israeli Parliament in 2015. We looked at six indicators of Facebook activity (number of fans, number of posts, and scope of engagement [likes, comments, shares, and overall engagement]) of all parties running for the Knesset as well as candidates with realistic electability rankings. We found that a comparison between dominant and peripheral parties across all indicators is consistent with the Normalization hypothesis, but when it is framed in terms of expectations and is forward-looking rather than backward-looking, that is, the difference in Facebook performance is between parties that expect to gain a significant number of seats in the parliament, and those that do not anticipate significant parliamentary achievements.

Keywords

Campaigning equalization Facebook social media

Introduction

One of the key research questions of online politics is whether Internet usage favors established/dominant or peripheral/marginal parties and candidates (Farrell, 2012). To further examine this question, this study analyzes Facebook campaigning by parties and contenders during the 2015 Israeli parliament (Knesset) election campaigns.

The Normalization and Equalization hypotheses are two competing hypotheses which address the impact of the online realm on the political arena (Gibson and McAllister, 2015; Gueorguieva, 2008; Larsson and Moe, 2014; Lilleker and Vedel, 2013; Lilleker et al., 2011; Margolis and Resnick, 2000; Small, 2008; Southern, 2015). The Equalization hypothesis claims that Internet platforms predominantly aid peripheral and marginal players. This hypothesis is equally applicable during the campaign season. Since mass media exposure can be prohibitively expensive, online platforms may function as an optimal solution for peripheral and marginal players, due to the relatively low expense of establishing and maintaining a digital presence, potentially enabling new contenders to reach larger, as well as more targeted, audiences.

Conversely, the Normalization hypothesis suggests that online activity will eventually reward established and dominant players at the expense of their peripheral and marginal counterparts. Due to availability of resources, dominant players will outperform contenders on more digital channels, generate more content, attract more traffic, and engage more audiences. Also, it turns out that substantial resources are necessary to maintain a constantly updated professional presence on multiple online platforms, puncturing zero cost romanticism. Consequently, the gap between established and peripheral players in terms of establishing digital presence, reaching audiences, and engaging them will actually expand over time.

The corpus of studies that were published over the years demonstrate that over time the use of political websites aligns with the Normalization hypothesis (Lilleker and Vedel, 2013). The growing use of Facebook as a new arena for online political activity has established new hopes for online political equalization (Gibson and McAllister, 2015; Gueorguieva, 2008). Establishing and maintaining a political Facebook presence is less expensive and more intuitive than starting and operating a website. Audience-wise, accessing a Facebook page is more intertwined with daily usage habits; users do not need to go out of their way to access political websites. As such, there is widespread hope that online equalization, unobtainable using websites, might become viable due to Facebook. A few studies conducted in various countries have indeed demonstrated that Facebook is used first and foremost by peripheral contenders (see Schweitzer, 2011; Southern, 2015).

But does Facebook really offer a new hope for the Equalization hypothesis? We add to the existing literature in the following ways: first, our sample consists of Israeli users and Israel is a country saturated by Facebook (see data later). In light of the continuing penetration and establishment of Facebook as a political tool, data from high saturation locations will provide useful insights into trends and potential developments for political Facebook usage in lower penetration locations. Second, even though the significance of Facebook in political campaigns is incontestable, there is a lack of standardization concerning activity measurements. Scholars regularly use the number of fans, number of posts, or level of public engagement as indexes. This study, however, addresses all three parameters, offering wider and more generalizable conclusions that address Facebook activities by both contenders and audiences.

Third, to the best of our knowledge, this study is the first to analyze both the activities and engagement of the parties on Facebook as well as those of candidates in realistic rankings. Fourth, this study is among the first to use computer-automated tools to analyze campaigning activities. These tools greatly expand the scope of research while significantly shortening the time required to carry it out (Lazer et al., 2009).

Theoretical background

The fast penetration of the Internet, and its availability in many spheres of daily life, has generated significant interest, both public and academic, in its social and political impact. From its early days, many believed that the Internet would change the rules of the political game. The unique characteristics of the Internet, such as its potential interactivity, wide reach, the reduced entry barriers for new players, the relatively low usage costs, and the abilities to convey messages to large audiences in a relatively short time span, enable candidates and parties unfamiliar to the public at large to expose themselves to wide audiences and disseminate their messages (Bonchek, 1997).

In terms of impact of the Internet and social media on the political arena, two key hypotheses exist: the Equalization hypothesis and the Normalization hypothesis (Gibson and McAllister, 2015; Gueorguieva, 2008; Larsson and Moe, 2014; Lilleker and Vedel, 2013; Lilleker et al., 2011; Margolis and Resnick, 2000; Small, 2008; Southern, 2015).

The Equalization hypothesis

The Equalization hypothesis according to Larsson and Moe (2014) suggests that the Internet will promote equality of use between dominant and peripheral political players and their audiences. It further claims that the “less fortunate” players will use various online techniques in higher degrees compared to their competitors so as to compensate for their structural disadvantages.

The Equalization hypothesis did not gain much empirical support. Studies supporting it were of two types: first, they were predominately early studies that compared the activities of all parties that had competed in a specific campaign (Corrado and Firestone, 1996; Gibson and Ward, 1998). Second, case studies were used in order to analyze the digital uses of certain political actors. These specifically focused on cases of “outsider” parties and candidates, where the center of gravity of their campaign was conducted, and gained significant success, online, even though their digital success did not always carry over to the ballots. Studies refer to cases such as the digital campaign of Howard Dean in the United States (Hindman, 2005), the Podemos party in Spain (Tremlett, 2015), and Beppe Grillo and the five star movement in Italy (De Rosa, 2013). In Israel, parties such as the “Green movement-Meimad” and “Eretz Hadasha” similarly demonstrate that online salience does not necessarily translate into votes (Lev-On, 2011).

The Normalization hypothesis

The Normalization hypothesis, on the other hand, has gained substantially greater support in comparison to the Equalization hypothesis, at least in reference to website usage. According to this hypothesis, gaps in resources and presence would be reproduced and even enlarged in new media platforms. This is because dominant players are better equipped and financed, with peripheral parties lacking substantive resources and abilities to maintain a significant scope of activities over sophisticated digital platforms. The authors most identified with this hypothesis are Margolis and Resnick (2000), who claim that

As the Internet develops, patterns of socioeconomic and political relationships on-line come to resemble those of the real world. Applied to political parties, this hypothesis implies that just as the major parties dominate the sphere of everyday domestic politics, so they come to dominate cyberspace. (p. 26)

When the dominant parties actualize their resources for a comprehensive and sophisticated digital presence, they will not only be more active online but their websites will be more functional; in turn, this will attract more traffic (Foot and Schneider, 2006; Noam, 2005).

Many of the studies regarding website usage by political players are consistent with the Normalization hypothesis. Most have focused on specific election campaigns within a single state (Gibson, 2012; Lilleker and Vedel, 2013). This can be illustrated by Small’s (2008) analysis of website activity for Canada’s nine leading political parties, demonstrating extensive differences in terms of scope of activity between websites of dominant and peripheral parties. Likewise, Strandberg (2009) analyzed website usage throughout the general elections in Finland in 2003 and found that the probability of having a website was 10 times greater for candidates of large parties than those of fringe parties.

Since these studies are specific to context of place and time, it is difficult to generalize from them. Still, a similar picture emerges from longitudinal studies that examine website usage in the same country over a few consecutive election cycles. Among the relevant longitudinal studies are those by Schweitzer (2011), who compared digital presences (websites and social media) of parties competing in three federal election campaigns in Germany (2002, 2005, and 2009). The study demonstrates that at first parties that were not represented in the Parliament were innovative in their use of digital tools. However, over time the parties represented in Parliament understood the importance of maintaining fully functional websites and invested significantly more resources in them compared to parties not represented in Parliament (see also Schweitzer, 2008).

Normalization versus equalization on Facebook

The majority of studies that support the Equalization hypothesis elicit their data from websites. Yet the center of gravity for online politics is clearly moving to Facebook (Lev-On, Under Review). Political Facebook campaigning is introducing new perspectives for the study of equalization. On one hand, it arguably offers a friendlier environment for promoting equalization. The establishment and maintenance of websites demand resources and financial commitment; however, opening a Facebook page requires only very basic skills. In addition, Facebook is far more accessible for users to reach than politician websites. Consequently, some argue that the emergence of Facebook as a primary online political arena generates new hope in the direction of the Equalization hypothesis (Gueorguieva, 2008).

Nevertheless, one can argue that Facebook, even more than personal websites, benefits the established and resourceful political players, for two main reasons (Jackson, 2011). First, incumbents can develop dedicated communities of followers over time. Comparatively speaking, newcomers can only publicize their identities shortly before elections, limiting their time to develop public support. Second, they have more content to upload to their political page, content produced during routine political activity. New candidates may find it hard to generate comparable volume and salience of content (with the exception of public figures with pre-existing followers who enter politics).

In addition, although the elections in Israel are parliamentary (and not presidential), candidates still promote their party and party leaders on Facebook during campaigns. Thus, parties benefit from Facebook activity not only on the party’s own page, but also on that of the candidate. Parties with many active Members of Parliament (MPs) can benefit from their Facebook campaign activities, in addition to the advantages gained from the prominence of its own Facebook page.

Studies that support the Equalization hypothesis on social media

In a similar vein to websites, initial Facebook studies on equalization versus normalization supported the Equalization hypothesis. Before the dominant political players entered the stage, peripheral ones understood the advantages of social media and were the first to exploit them (Gueorguieva, 2008; Kalnes, 2009). Thus, studies revealed a picture ranging from modest to more significant equalization effects. Kalnes (2009) studied the presence of Norwegian politicians on Facebook and found that it mirrored the distribution of seats in the Parliament. Hansen and Kosiara-Pedersen (2014) studied the digital campaigns of the general elections in Denmark in 2011. They found that the dominant parties use social media for political purposes in a similar scope to that of their peripheral contenders. Likewise, the study by Larsson and Moe (2014) in Norway focused on Twitter and demonstrated that although dominant parties were better presented, peripheral parties had better ways of making their voices heard than on other media platforms.

Gibson and McAllister (2011) studied online campaigning in the general elections for the Australian Parliament in 2007. They found that social media use during the campaign provided real advantages to candidates from peripheral parties. In 2014, they conducted a longitudinal study based on data collected from 2001 up until, and inclusive of, 2010. This study found that major parties dominate in the adoption of personal websites, although minor parties are stronger users of social media. Chen and Smith (2010) comparatively analyzed online campaigning in Canada, Australia, and New Zealand. Their findings illustrated that Facebook use is more prevalent with peripheral parties whereas dominant parties are better represented on websites.

Perhaps the strongest support for the Equalization hypothesis comes from a study by Koc-Michalska et al. (2014), who analyzed online campaigning in regards to the presidency of France between 2007 and 2012. They confirmed that 2007 saw no significant differences in online activities of dominant as compared to peripheral parties. However, in 2012, the peripheral parties were much more active on Facebook than their dominant counterparts. Southern (2015) studied websites and social media campaigning in England in 2010, and conclusively found that websites presented strong differences between dominant and peripheral parties, supporting the Normalization hypothesis. Nonetheless, when looking at social media, peripheral parties were much more dominant, thus supporting the Equalization hypothesis. Southern (2015) argues that

Website adoption among candidates do support normalization overall, adoption and use of certain types of social media refute normalization on some measures, some to a strong degree. That is that—patterns of “Web 1.0” (Web sites) adoption supporting normalization, but “Web 2.0” (social media) adoption and use offering some evidence to refute normalization. (p. 1)

Studies that demonstrate normalization in social media

The studies presented so far demonstrate that at least in the earlier stages of Facebook’s diffusion in a given country, peripheral parties use it more than dominant parties even if the dominant parties are prevailing on websites that preceded the use of Facebook. Still, some recent studies present a more nuanced picture according to which the Normalization hypothesis actually gains hold in terms of Facebook usage as well. This may be the result of Facebook maturing into an established and leading platform used by the general public for political and non-political purposes. As such, dominant political parties have started to flock to this platform as well.

Lilleker et al. (2011) were among the first scholars to demonstrate this phenomenon. Their study of Facebook usage by candidates in six countries during the election campaigns for the European Parliament in 2009 revealed a picture consistent with the predictions of the Normalization hypothesis. Similarly, Vergeer et al. (2013) analyzed personal Twitter campaigns of candidates for European Parliament in 2009. Additionally, Chen and Smith (2011) looked at online campaigns in Canada in 2008, while Klinger (2013) focused on online campaigns for the 2011 Swiss general elections. All of these studies demonstrated signs of normalization in Facebook campaigning. Notably, Strandberg (2013) studied the uses of social media by candidates and the public during the elections for the Finnish Parliament in 2011. Their campaign characteristics also fall under the umbrella of the normalization approach in terms of usage by politicians on Facebook. Finland is one of the most online savvy countries in the world, and boasts one of the highest rates of Facebook penetration.

A similar picture emerges in longitudinal studies, for example by Schweitzer (2011) who found support for the Normalization hypothesis in online campaigning in Germany between 2002 and 2009 in terms of both websites and social media. Schweitzer claimed that the “normalization thesis holds true in all its dimensions over time and in the Web 2.0 era … the gap is wider in Web 2.0 than in Web 1.0” (p. 310). Baxter and Marcella (2013) studied Facebook use in two rounds of the Scottish general elections held in close proximity in 2010 and 2011. They concluded that more political Facebook activity occurred in the second election, but the increase was attributable to the more dominant political parties whereas the peripheral ones maintained similar levels of Facebook activity.

These studies do not present a single clear picture, and still many of them support the Normalization hypothesis. This support can be classified as mainly deriving from the Web 1.0 environment and less so from Web 2.0. In addition, it seems that the Normalization hypothesis holds better in environments in which users have acclimatized over time to a given technology. In the early stages, when a new technology appears, peripheral players tend to be early adopters. When the technology is popularized, more dominant parties start experimenting with it. They scale up and professionalize the technology by comparison to peripheral players. Obviously, the predictions of the Equalization hypothesis are no longer sustainable and must give way to a picture that better aligns with the Normalization hypothesis.

Still, there are a few difficulties in generalizing from the results presented above, for a few reasons. First, data were collected in different regions with varying Facebook penetration rates and Internet usage levels. Second, the research comprises an umbrella of eclectic representative political systems, some presidential and some parliamentary. Some studies focus on the use of Facebook and social media by individual candidates and some by political parties.

It is especially interesting to study Facebook usage by candidates and parties in locations where usage is very high and political activism is very strong. Indeed, Israel is a fertile ground for studies relating to Facebook users. Israel is one of the world leaders in Facebook usage in terms of population percentage of social network users and time allocated on Facebook (ComScore, 2011). In 2015, 6.3 million Israelis (about 77% of the Israeli population) were online, whereas 4.4 million users were on Facebook. Facebook also functions as a central news hub, mainly for users between the ages 13–34 years (68%), and slightly less so for users who are 35 years old and above (50%) (Mann and Lev-On, 2016). Not only is the Israeli public flocking to Facebook in large numbers, but Israeli political figures are also quite active users. As of January 2015, 108 out of 120 MPs were active on Facebook in varying scopes and frequencies (Steinfeld and Lev-On, Under Review).

This study adds to the existing literature also by analyzing normalization versus equalization patterns at the intra-party level. To the best of our knowledge, only two studies analyzed these questions at the intra-party level. Strandberg (2009) concluded that high-ranked candidates were more likely to have a web presence than low-ranked candidates in the Finnish parliamentary elections. Vergeer et al. (2013) found that mostly highly ranked candidates used Twitter in the elections for the European parliament. This study adds to the small literature on intra-party effects on candidates’ digital presence.

The Israeli political context

Israel’s electoral system is based on nation-wide single district proportional representation, where citizens vote for their preferred party and not for any individual candidates. Prior to the elections, each party submits its list of candidates for the Israeli parliament, the Knesset (in order of precedence). The parties select their candidates for the Knesset in primaries or by other procedures. After the elections, the 120 seats in the Knesset are then assigned proportionally to each party that received votes, provided that the vote share that the party gained exceeded the electoral threshold (currently 3.25%).

Two studies in Israel have looked at the normalization and equalization question regarding the Knesset elections of 2009. Atmor and Siani (2011) analyzed all the websites of the parties running for the 18th Knesset (Parliament), using the methodological scheme developed by Gibson and Ward (2000). They found notable differences between parties represented in the Knesset and those that failed to cross the electability threshold. In all four dimensions characteristics of websites (information provision, resource generation, interactivity, and connectivity), the peripheral parties scored lower than the dominant parties. Caspi and Lev (2009) calculated an index for new media usage for each of the 33 parties that ran in the elections of 2009 based on six media channels (emails, blogs, websites, YouTube, Facebook, and cellular text messages). One point was assigned for every usage of each of these tools. The average grade of the 15 parties that crossed the threshold was 3.9, whereas the average grade of the 18 parties that did not cross the threshold was 2.7.

This article analyzes normalization versus equalization in political Facebook campaigns through the case study of the 20th Knesset elections of March 2015. Although the electoral system in Israel is parliamentary, the personalization process in Israeli politics (Rahat and Sheafer, 2007) is well documented. One aspect of this is that media attention is directed at the candidates and not at the parties (Livak et al., 2011). In addition to tracking party activity on Facebook, we also addressed Facebook activities of candidates in realistic rankings on the party lists.

Research questions

Our research addresses four facets of the Normalization versus Equalization question:

RQ1. Is there a difference in the mere Facebook presence of dominant and peripheral parties as well as in presence of candidates ranked in realistic places?

RQ2. Is there a difference in the scope of Facebook activity between dominant and peripheral parties?

RQ3. Is there a difference in the scope of Facebook activity between the candidates in realistic rankings in the dominant and peripheral parties?

RQ4. In the intra-party facet, are there differences between the scope of Facebook activity between high-ranking and low-ranking candidates?

Methodology

This study analyses the Facebook campaigning of the parties that ran for the 20th Israeli Parliament (Knesset) and of candidates with realistic rankings. A total of 25 parties competed in the elections on 17 March 2015. On 30 January 2015, all parties that had decided to run for Knesset submitted the required forms to the Central Election Committee (henceforth, submission day). Only from that point on did it become clear which parties were going to compete and how the candidates were positioned. Consequently, we decided to cover the time period between submission and election days. The data collection period took place between 2 February 2015 (the first full week after the lists of candidates were finalized and submitted) and Election Day on 17 March 2015.

We faced a dilemma as to the number of candidates from each party who should be monitored. Our assumption was that candidates with a realistic chance for being elected into the Parliament would be more inclined to be politically active on Facebook. Candidates in non-realistic rankings would be correspondingly less active or even inactive. To reach an operational definition as to which of the candidates are ranked realistically, we examined all the polls within the study’s time-frame using the Facebook page of project 61.¹ This page is a public initiative unaffiliated with any official body that incorporates all the polls published during the campaign season. We used this page to find the maximal number of seats predicted for each party according to the most optimistic poll.

In addition, starting from the current elections, Knesset entry threshold was pushed up to 3.25% which equals four seats. Hence, for the parties that were not predicted by any poll to cross this threshold, we examined the activity of the first four candidates. For the parties that were fluctuating or predicted to cross the threshold, we checked the optimal number of expected seats plus 4 seats for purposes of standardization. To illustrate, if a party was predicted 20 seats at best, we monitored the Facebook performance of the leading 24 candidates.

We obtained the names of the parties and their lists of candidates from the Central Election Committee’s website. We located the Facebook pages of the parties and the candidates with the help of three research assistants who searched for them on party websites, on Facebook and on Google. The study population included 265 pages: 25 party pages and 240 candidate pages, out of which we located 23 party Facebook pages and 132 candidate pages. Only pages (and not personal profiles) were analyzed, since according to Facebook’s terms of use, political activity is allowed only in the format of designated political pages and not through personal profiles. Violation of these terms can lead to a profile being shut down overnight.

To collect party and candidate activity data, we used Netvizz, an application that scrapes page data from Facebook’s Application Programming Interface (API). The number of page fans was collected manually on Election Day by the researchers. Overall, the study analyzed six indicators of Facebook campaigning: number of fans on Election Day; number of posts published throughout the data collection period (including posts by page and by users); overall number of likes, comments, and shares of posts during the data collection period; and, finally, the overall engagement that summed up all engagement dimensions with contents uploaded to the pages over the data collection period (number of likes, comments, comments–likes, and shares) for each page.

To address the final research question (RQ4) regarding difference in scope of Facebook activity in the intra-party level between high-ranked and low-ranked candidates, we divided the lists of candidates of the five largest parties into groups of eight according to their position on the party list: places 1–8, 9–16, 17–24. Parties for which the Facebook activity of 24 candidates or more were recorded, namely Likud and the Zionist Camp, were divided into three groups. The rest of the parties, for which the Facebook activity of 16–23 candidates or more was recorded, which consists of the Arab Joint List, Yesh Atid, and the Jewish Home, was divided into two groups.

Findings

For purposes of analysis, we have divided all parties into two categories: dominant and peripheral parties. The dominant parties include 11 parties that had a realistic chance of electability, and that eventually attained significantly more votes than the peripheral parties (see Table 1). A total of 10 of these parties entered the Knesset, and only one party, “Yachad,” came close to the threshold but failed to cross it. The dominant parties are Likud, Zionist Union, Jewish Home, Yesh Atid, Kulanu, Meretz, Israel Beitenu, Arab Joint List (i.e. Alliance of four Arab-dominated parties), Shas, Yachad, and Yahadut HaTorah (Ultra-orthodox party). The peripheral parties are U’Bizchutan, Democratura, Kulanu Haverim NaNach, Social Leadership, Calcala, Ale Yarok, The Pirates Party, Orr, Israeli Green Party, Nivcheret Ha’Am, Rent with Dignity + Direct Party, Hope for Change, Arab List, and Perach Netz.

Table 1.

Party and candidate Facebook (FB) activity indicators.

Party	Voters	Seats	Maximal expected seats + 4	No. of FB pages	Fans	Posts	Likes	CMT	SHR	ENG	Fans	Posts	Likes	CMT	SHR	ENG
			Maximal expected seats + 4	No. of FB pages	Activity—Party page						Activity—Candidate pages
Likud	985,408	30	31	25	26,886	1835	164,331	22,309	20,208	206,848	67,353	50.3	123,269.2	15,376	11,945.1	150,590.3
Zionist Union	786,313	24	30	25	27,047	1970	159,552	31,365	25,834	216,751	21,588	144.4	39,591.6	10,089.3	4418.6	54,099.5
Arabic Joint List	446,583	13	17	11	21,984	509	172,126	6044	14,258	192,428	6435	51.7	20,758.8	1800.1	1211.5	23,850.4
Yesh Atid	371,602	11	18	18	12,449	85	27,682	3910	5056	36,648	31,880	544.3	28,375.4	6743.4	2132.7	37,251.6
Kulanu	315,360	10	15	9	11,543	1179	34,403	6241	3006	43,650	9227	178.3	24,180.9	4221.7	2056.8	30,459.3
Jewish Home	283,910	8	18	15	30,826	1372	26,304	3276	2457	32,037	39,270	60.2	114,086.8	11,065.9	9630.5	134,783.2
Shas	241,613	7	13	3	4685	337	11,102	1601	1032	13,735	4445	45.2	4260.9	593.9	214.5	5069.3
Israel Beitenu	214,906	6	11	10	46,257	398	61,658	5257	3281	70,196	27,284	123.8	16,726.9	2115.3	1006.5	19,848.6
Yahadut HaTorah	210,143	6	12	3	2538	88	10,260	1144	1470	12,874	67.9	2.3	7.4	3.7	0.8	11.9
Meretz	165,529	5	10	8	70,967	1110	213,970	28,932	28,516	271,418	19,675	153.9	39,190.7	5610	4314.1	49,114.8
Yachad	125,158	0	9	3	1849	234	7361	446	473	8280	8107	177.5	43,134.4	3430.1	2436.4	49,000.9

Ale Yarok	47,180	0	4	0	82,773	1046	137,209	12,461	19,083	168,753	0	0	0	0	0	0
Arab List	4301	0	4	0	0	0	0	0	0	0	0	0	0	0	0	0
Israeli Green Party	2992	0	4	0	20,655	123	1051	182	314	1547	0	0	0	0	0	0
Kulanu Haverim NaNach	2493	0	4	0	2085	121	2074	234	533	2841	0	0	0	0	0	0
U’Bizchutan	1802	0	4	2	3881	338	16,257	1203	1604	19,064	23.3	22.3	32.8	1.8	1.8	36.3
Hope for Change	1385	0	4	0	2523	73	1578	228	33	1839	0	0	0	0	0	0
The Pirates Party	895	0	4	0	1155	23	90	13	20	123	0	0	0	0	0	0
Perach Netz	823	0	4	0	0	0	0	0	0	0	0	0	0	0	0	0
Nivcheret Ha’Am	761	0	4	0	1277	305	5704	960	6307	12,971	0	0	0	0	0	0
Orr	502	0	4	0	4797	47	758	306	166	1230	0	0	0	0	0	0
Rent with Dignity + Direct Party	423	0	4	0	1360	48	201	112	139	452	0	0	0	0	0	0
Calcala	337	0	4	0	29,855	23	33	14	2	49	0	0	0	0	0	0
Democratura	242	0	4	0	97	19	46	19	7	72	0	0	0	0	0	0
Social Leadership	223	0	4	0	87	56	47	107	28	182	0	0	0	0	0	0

Interestingly, it appears that the demarcation line between dominant and peripheral parties is not the presence of the party in the current parliament, but rather the expectation to enter the next parliament. An interesting case is Kulanu, a party established only a few months prior to the elections. Still, polls predicted between 8 and 10 seats, and its patterns of Facebook usage resembled those of the other dominant parties. Another interesting case is the Yachad party that failed to cross the threshold although it was borderline. Their online activity could have been one of the contributing factors deciding success or failure, and hence also resembled those of the other dominant parties.

To answer the research questions, we compared the various indicators of Facebook campaigning of both peripheral and dominant parties.

Research question 1

Is there a difference in the mere Facebook presence of dominant and peripheral parties as well as in presence of candidates ranked realistically? Altogether, for the 25 parties, 23 Facebook pages were located. Only 2 parties, both peripheral, did not maintain a Facebook page. Hence, both dominant and peripheral parties are nearly universally using Facebook.

Things are different in reference to the activities of candidates with realistic rankings. A total of 130 of the 184 dominant party candidates ranked realistically (70.7%) had Facebook pages. However, only 2 of 56 candidates from peripheral parties with realistic rankings (3.6%) had Facebook pages. Altogether, 132 out of 240 candidates ranked realistically (55%) had a Facebook page.

Research question 2

Is there a difference in the scope of Facebook activity between dominant and peripheral parties? Table 1 presents the Facebook indicators for the different parties. Due to the high variance within the groups, Mann–Whitney tests were used to find whether there are differences in the scope of Facebook activity between the dominant and peripheral parties. Table 2 presents the average Facebook indicators of dominant and peripheral parties. Across all Facebook indicators, the pages of the dominant parties were significantly more active than the pages of the peripheral parties: more fans (Z = –2.46, p < .05), posts (Z = –3.28, p < .01), likes (Z = –3.66, p < .01), comments (Z = –3.61, p < .01), shares (Z = –3.17, p < .01), and greater overall engagement (Z = –3.55, p < .01). According to all Facebook indicators, the dominant parties are more active than the peripheral parties.

Table 2.

Facebook activity indicators of dominant and peripheral parties.

Fans		Engagement		Shares		Comments		Likes		Posts
SD	M	SD	M	SD	M	SD	M	SD	M	SD	M
21,557	21,935	99,556	100,442	10,646	9599	11,581	10,048	79,241	80,795	693	829	Dominant parties
22,539	10,753	44,636	14,937	5189	2017	3281	1131	36,359	11,789	276	159	Peripheral parties

SD: standard deviation.

Research question 3

In order to analyze the differences in the scope of Facebook activity between candidates with realistic rankings in dominant and peripheral parties, we used Mann–Whitney tests (again because of the high variance within the groups). Table 3 presents the average of the Facebook indicators of candidates from dominant and peripheral parties. Again, it is apparent that pages of candidates from dominant parties had significantly more Facebook activity than the pages of candidates from peripheral parties, that is, more fans (Z = –8.08, p < .01), posts (Z = –8.04, p < .01), likes (Z = –8.16, p < .01), comments (Z = –7.98, p < .01), shares (Z = –7.96, p < .01), and greater overall engagement (Z = –8.22, p < .01). In response to the third research question, candidates with realistic rankings on the lists of dominant parties are more active than those on the lists of peripheral parties.

Table 3.

Facebook activity indicators of dominant and peripheral parties’ candidates.

Fans		Engagement		Shares		Comments		Likes		Posts
SD	M	SD	M	SD	M	SD	M	SD	M	SD	M
107,714	28,233	297,007	62,392	24,808	4604	310,504	7157	242,317	50,631	586	138	Dominant parties
9	2	15	3	1	0	1	0	13	2	8	2	Peripheral parties

SD: standard deviation.

Research question 4

In order to examine the differences in the scope of activity, we ran Kruskal–Wallis and Mann–Whitney tests. According to the Kruskal–Wallis test, there is a significant difference in all the indicators for the Likud. According to the Mann–Whitney test, the source of difference is between the first and the third groups of eight. All six tests were significant, at least in significance levels of p < .05.

According to the Kruskal–Wallis test, there is a significant difference in all the indicators within the group division for the Zionist Camp. According to Mann–Whitney, the source of differences is again between the first and the third groups of eight, and also between the second and third groups of eight—11 out of the 12 tests were significant in the significance level of p < .05.

According to the Mann–Whitney test, there were no significant differences in all the indicators between the candidates of the Jewish Home party. According to the Mann–Whitney test, there were significant differences between the first and the second group of contenders in Yesh Atid in all the indicators that were checked in the level p < .05, with an exception of the number of fans. Finally, according to the Mann–Whitney test, there were no significant differences in any indicators that were checked, among the members of the Arab Joint List.

Henceforth, in response to research question 4, in most but not all cases, and especially within the dominant parties, the pages of high-ranking candidates are significantly more active on Facebook than those of the low-ranking candidates (Table 4). Even when differences between groups of eight were not significant, they were still large and indicative.

Table 4.

Facebook activity indicators of high-rank versus low-rank candidates.

Party	Fans	Posts	Likes	CMT	SHR	ENG	Fans	Posts	Likes	CMT	SHR	ENG	Fans	Posts	Likes	CMT	SHR	ENG
Activity indicators per candidates 1–8 (average)							Activity indicators per candidates 9–16 (average)						Activity indicators per candidates 17–24 (average)
Likud	207,593	69	428,520	53,744.6	42,919.6	525,184.3	42,227	84.3	40,953.6	4734.4	2712	48,400	8945	20.8	7182.8	978.6	498.6	8660
Zionist Union	52,797	384.6	118,307.3	30,687.6	13,233.5	162,228.4	20,477	96.6	26,743.8	6699.3	2968.4	36,411.4	5687	38.8	2071.4	281.5	176.9	2529.8
Arabic Joint List	9040	74.5	38,460.8	3813	2426.6	44,700.4	4635	35.4	5651.8	182.1	147.9	5981.8	–	–	–	–	–	–
Yesh Atid	60,152	1082.3	57,830.9	13,964.8	4479.8	76,275.4	11,579	129.1	5414.1	1066.6	288.5	6769.3	–	–	–	–	–	–
Jewish Home	73,274	82	238,451.5	23,537.5	20,117.8	282,106.8	14,893	52.6	18,174	1358.9	1549.5	21,082.4	–	–	–	–	–	–

Discussions and conclusion

The impact of social media on equalization versus normalization is one of the central questions in the study of the Internet in general and social media in particular. Does the Internet assist to even the playing field by lowering entrance costs and opening up more opportunities for peripheral players? Or does it solidify the established and well-endowed players, entrenching and even exacerbating the status-quo?

The literature review demonstrates that when a new and relatively inexpensive technology such as websites enters the playing field, it is used first by peripheral and smaller players that lack resources and are motivated to find cost-effective channels for conveying their messages. Hence, they become the early adopters of these new technologies and in such cases it seems as if the predictions of the Equalization hypothesis do indeed hold. Yet over time, the more established and resourceful players identify the potential of these new technologies and begin to make use of them. Since their resources are significantly greater than those of weaker and peripheral players, their presence on new platforms overshadows that of the latter, and their websites are more functional and continuously updated. The same holds true for social media, and over time the supremacy of established over marginal players is reproduced. This dynamic resembles the one portrayed by Seymour-Ure (1977) according to which established political parties adjust their media strategies to take advantage of the dominant media channels at an opportunistic point in time. Weaker players are earlier adopters of new technologies that get extensively used by more established players only when the technology is popularized.

The growing dominance of Facebook in online campaigning has sparked a new interest in the question of equalization. Compared to websites, it is simpler to establish and maintain a Facebook presence for new contenders in desperate need of inexpensive channels for public recognition and exposure. Audience-wise, accessing a Facebook page is more intertwined with daily usage habits; users do not need to go out of their way to access political websites. On the other hand, Facebook may be friendlier for incumbents who have already established significant groups of online supporters over time. These incumbents can also publicize their political track records on Facebook, which new contenders cannot. Some evidence demonstrates that when Facebook was first recruited for political ends it was indeed more hospitable to peripheral players; but this might be changing as the presence of dominant players on social media continues to grow.

This study strongly supports the predictions of the Normalization hypothesis, and is innovative for a number of reasons. First, it examines a number of Facebook activity indicators (namely, presence, fans, posts, and engagement). Second, it analyzes pages of both parties and candidates, based on the perception that the campaign is spread over multiple Facebook pages and (at least) the candidates ranked realistically have clear incentives to assist the party campaign and simultaneously increase their electoral chances. Third, the study was conducted in Israel where there is an overarching usage of Facebook. We assume that countries in which Facebook has deep penetration, such as Finland (Strandberg, 2009) and Israel, present a picture that will be reflected in future studies conducted in countries in which current Facebook penetration is still in its infancy. Hence, these findings can act predictively for other countries when Facebook is going to become a more popular, sophisticated and effective political tool.

The findings demonstrate that operating a Facebook page is now imperative in the political arena, as 23 out of 25 parties who competed in the elections had a Facebook page. Even among the peripheral parties, a substantive majority (12 out of 14 parties) had operated a Facebook page. But while a Facebook presence per se is required, the scope of Facebook activity significantly varies between dominant and peripheral parties, across all the indicators examined.

We found dramatic differences between activities in the pages of the dominant and peripheral parties, where the demarcation line between the two types of parties is the expectation to enter parliament. An interesting case is the Yachad party that failed to cross the threshold although it was borderline. Their online activity could have been one of the contributing factors deciding success or failure. In situations where election results determine if the party enters parliament and how many seats it would receive, the party and the high-ranking contenders have stronger incentives to be active on Facebook, as opposed to parties with slimmer chances. Hence, we can think of the scope of Facebook-related activity on party and candidate pages as a barometer which is influenced by the polls.

Additional evidence that the Facebook-related activity of parties and candidates, as a rule of thumb, is forward-looking (driven by expectations derived from polls outcomes) and not backward-looking (driven by number of seats in the existing parliament), is the case of Kulanu, a party established only a few months prior to the elections. Still, polls predicted between 8 and 10 seats, and its patterns of Facebook usage resembled those of the dominant parties. This claim has two exceptions: first, there were a number of peripheral parties that did not expect to be elected to the Knesset, and were far from passing the threshold, but still had some level of Facebook activity. These parties, namely the Orthodox women’s party U’Bizchutan (promoting ultra-orthodox women’s rights) and Aleh Yarok (calling for drug legalization), had modest levels of Facebook-related activity which was needed to convey their messages to the public (probably the reason they decided to run for parliament in the first place). Congruently, Ultra-orthodox parties such as Shas and Yahadut HaTorah barely used Facebook ranking among the dominant parties, due to cultural and religious reasons (Lev-On, and Neriya-Ben Shahar, 2011). Likewise, most of their contenders did not have Facebook pages.

The fact that, on one hand, the majority of parties were present on Facebook, and, on the other hand, a significant difference existed in their activity indicators, demonstrates that future studies should not be content with inspecting mere presence on Facebook but should also examine scope of Facebook-related activities.

We also examined the differences in the scope of Facebook-related activity between high-ranked and low-ranked candidates. We found significant differences between the different groups in each party. Although in some cases these differences were not significant, they were still indicative. As a rule of thumb, the pages of candidates that had more to gain from being elected to the parliament, such as party leaders and candidates who anticipated ministerial appointments, were very active on Facebook. As candidates were ranked lower on party lists, their level of page activity dropped.

Almost all of the candidates in rankings one to eight, in all five dominant parties (Likud, Zionist Camp, Arabic Joint List, Yesh Atid, and the Jewish Home) that were examined, had a Facebook page. Nonetheless, the scope of page activities was quite different between the groups of eight. It seems that the differences in the scope of activities were not a result of the fact that the pages of all candidates in the first group of eight were very active, the pages of candidates in the second group were less active, and the pages of those in the third group of eight were even less active or not active at all. Instead, the differences derived from the fact that the active candidates in the first group of eight were much more active than the active candidates in the second group of eight, and so forth.

Additionally, the differences in the scope of candidate page activity within parties sometimes seemed sensitive to particular candidate selection methods. For example, the Likud party holds internal primaries. Up until the 20th place on the list, candidates are ranked by their relative place in the primaries, while from the 21st place onward, the party uses regional representation. In this system, the incumbent MPs are generally more highly ranked. This means that those ranked higher on the list tend to be active incumbents or new candidates who are previously known from within the public scene. Accordingly, candidates lower on the list are less familiar on the national level and their pages are less active on Facebook in most cases.

The findings reflect the current situation in a technologically advanced country, where Facebook is well-established as an online platform both for the general public and political players. Our research corresponds with previous work showing that with time the expectations of the Equalization hypothesis are replaced with the predictions of the Normalization hypothesis, as dominant and established parties and candidates take over the social media scene, and this holds true across all facets that we examined: between parties, between candidates in realistic electability places, and within parties as well.

For future studies, it may be interesting to examine whether there is a shift in other countries to normalization as Facebook assumes more dominance. Due to this study’s limitations we did not examine the content on party and candidate Facebook political pages, and therefore could not analyze the mechanics of content popularization. Such studies will provide a deeper understanding of the significance of Facebook activity during political campaigns.

Footnotes

Acknowledgements

Many thanks to the institute for the study of new media, politics and society in Ariel University, as well as for Yonatan Amir, Michael Barmatz, and Reut Avrahami for their assistance in collecting the data and preparing the manuscript.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biographies

Azi Lev-On is a faculty member in the School of Communication in Ariel University. His research focuses on the uses and perceived effects of social media, public participation and deliberation online, online communities, collective action and campaigns, and behaviors in computer-mediated environments, employing a variety of methods such as content and link analysis, surveys and laboratory experiments. For more information, see

Sharon Haleva-Amir teaches in Tel Aviv University and Beit Berl Academic College, and is a research fellow at the Haifa Center of Law and Technology (HCLT). Her research focuses on Israeli e-Politics as well as law and technology, employing both qualitative and quantitative methods such as computer automated online tools, websites and social media content analysis, interviews and questionnaires. For more information, see .

References

Atmor

Siani

(2011) Party websites in the 2009 elections: a comparative view. In: Cohen

Lev-On

(eds) Connected: Politics, Technology and Society in Israel. Tel Aviv: Israeli Political Science Association, pp. 177–210.

Baxter

Marcella

(2013) Do online election campaigns sway the vote? A study of the 2011 Scottish parliamentary election. Libri 63(3): 190–205.

Bonchek

(1997) From broadcast to netcast: the internet and the flow of political information. Doctoral Dissertation, Harvard University, Cambridge, MA.

Caspi

Lev

(2009) Premature Americanization: new media in the elections for the 18th Knesset. Kesher 39: 6–16.

Chen

Smith

(2010) Adoption and use of digital media in election campaigns: Australia, Canada and New Zealand compared. Public Communication Review 1(1): 3–26.

Chen

Smith

(2011) Digital media in the 2008 Canadian election. Journal of Information Technology & Politics 8(4): 399–417.

ComScore (2011) It’s a social world: top 10 need-to-knows about social networking and where it’s headed. Available at: http://www.comscore.com/it_is_a_social_world

Corrado

Firestone

(1996) Elections in Cyberspace: Towards a New Era in American Politics. Washington, DC: Aspen Institute.

De Rosa

(2013) The Five Stars Movement in the Italian political scenario: a case for cybercratic centralism? JeDEM: eJournal of eDemocracy and Open Government 5(2): 128–140.

10.

Farrell

(2012) The consequences of the internet for politics. Annual Review of Political Science 15: 35–52.

11.

Foot

Schneider

(2006) Web Campaigning. Cambridge, MA: MIT Press.

12.

Gibson

(2012) From brochureware to “MyBo”: an overview of online elections and campaigning. Politics 32(2): 77–84.

13.

Gibson

McAllister

(2011) Do online election campaigns win votes? The 2007 Australian “YouTube” election. Political Communication 28(2): 227–244.

14.

Gibson

McAllister

(2015) Normalizing or equalizing party competition? Assessing the impact of the web on election campaigning. Political Studies 63: 529–547.

15.

Gibson

Ward

(1998) UK political parties and the Internet “politics as usual” in the new media? The International Journal of Press/Politics 3(3): 14–38.

16.

Gibson

Ward

(2000) A proposed methodology for studying the function and effectiveness of party and candidate web sites. Social Science Computer Review 18(3): 301–319.

17.

Gueorguieva

(2008) Voters, MySpace, and YouTube: the impact of alternative communication channels on the 2006 election cycle and beyond. Social Science Computer Review 26(3): 288–300.

18.

Hansen

Kosiara-Pedersen

(2014) Cyber-campaigning in Denmark: application and effects of candidate campaigning. Journal of Information Technology & Politics 11(2): 206–219.

19.

Hindman

(2005) The real lessons of Howard Dean: reflections on the first digital campaign. Perspectives on Politics 3(1): 121–128.

20.

Jackson

(2011) Perception or reality: how MPs believe the Internet helps them win votes. Journal of Political Marketing 10(3): 230–250.

21.

Kalnes

(2009) Norwegian parties and Web 2.0. Journal of Information Technology & Politics 6(3–4): 251–266.

22.

Klinger

(2013) Mastering the art of social media: Swiss parties, the 2011 national election and digital challenges. Information, Communication & Society 16(5): 717–736.

23.

Koc-Michalska

Gibson

Vedel

(2014) Online campaigning in France, 2007–2012: political actors and citizens in the aftermath of the web.2.0 evolution. Journal of Information Technology & Politics 11(2): 220–244.

24.

Larsson

Moe

(2014) Triumph of the underdogs? Comparing Twitter use by political actors during two Norwegian election campaigns. SAGE Open 4(4): 2158–2440.

25.

Lazer

Pentland

Adamic

et al . (2009) Life in the network: the coming age of computational social science. Science 323: 721–723.

26.

Lev-On

(2011) Campaigning online: Use of the Internet by parties, candidates and voters in national and local election campaigns in Israel. Policy & Internet 3(1): 1–28.

27.

Lev-On

(Under Review) Online campaigning moves to Facebook: Contenders’ perspectives.

28.

Lev-On

Neriya-Ben Shahar

. 2011. A forum of their own: Views about the Internet among ultra-Orthodox Jewish women who browse designated closed forums. First Monday 16(4).

29.

Lilleker

Vedel

(2013) The Internet in campaigns and elections. In: Dutton

(ed.) Oxford Handbook of Internet Studies. Oxford: Oxford University Press, pp. 401–420.

30.

Lilleker

Koc-Michalska

Schweitzer

et al . (2011) Informing, engaging, mobilizing or interacting: searching for a European model of web campaigning. European Journal of Communication 26(3): 195–213.

31.

Livak

Lev-On

Doron

(2011) MK websites and the personalization of Israeli politics. Israel Affairs 17(3): 445–466.

32.

Mann

Lev-On

(2016) Annual Report: The Media in Israel 2015: Agendas, Uses and Trends. Ariel, Israel: Institute for the Study of New Media, Politics and Society.

33.

Margolis

Resnick

(2000) Politics as Usual: The Cyberspace “Revolution.” London: SAGE.

34.

Noam

(2005) Why the Internet is bad for democracy. Communications of the ACM 48(10): 57–58.

35.

Rahat

Sheafer

(2007) The personalization(s) of politics: Israel, 1949–2003. Political Communication 24(1): 65–80.

36.

Schweitzer

(2008) Innovation or normalization in E-campaigning? A longitudinal content and structural analysis of German party websites in the 2002 and 2005 national elections. European Journal of Communication 23(4): 449–470.

37.

Schweitzer

(2011) Normalization 2.0: a longitudinal analysis of German online campaigns in the national elections 2002–9. European Journal of Communication 26(4): 310–327.

38.

Seymour-Ure

(1977) Parliament and mass communication in the twentieth century. In: Walkland

(ed.) The House of Commons in the Twentieth Century. Oxford: Clarendon Press, pp. 527–592.

39.

Small

(2008) Equal access, unequal success: major and minor Canadian parties on the net. Party Politics 14(1): 51–70.

40.

Southern

(2015) Is web 2.0 providing a voice for outsiders? A comparison of personal web site and social media use by candidates at the 2010 UK general election. Journal of Information Technology & Politics 12(1): 1–17.

41.

Strandberg

(2009) Online campaigning: an opening for the outsiders? An analysis of Finnish parliamentary candidates’ websites in the 2003 election campaign. New Media & Society 11(5): 835–854.

42.

Strandberg

(2013) A social media revolution or just a case of history repeating itself? The use of social media in the 2011 Finnish parliamentary elections. New Media & Society 15: 1329–1347.

43.

Steinfeld

Lev-On

(Under Review). MPs on FB: An analysis of politicians’ Facebook pages.

44.

Tremlett

(2015). The Podemos revolution: how a peripheral group of radical academics changed European politics. The Guardian, 31 March. Available at: http://www.theguardian.com/world/2015/mar/31/podemos-revolution-radical-academics-changed-european-politics

45.

Vergeer

Hermans

Sams

(2013) Online social networks and micro-blogging in political campaigning: the exploration of a new campaign tool and a new campaign style. Party Politics 19(3): 477–501.