THE DYNAMICS OF THE DEMOBILIZATION OF THE PROTEST CAMPAIGN IN 

ASSAM 

Tijen Demirel-Pegg 

Indiana University-Purdue University Indianapolis 

Abstract: This study highlights the role that critical events play in the demobilization of protest 

campaigns. Social movement scholars suggest that protest campaigns demobilize as a 

consequence of polarization within the campaign or the cooptation of the campaign leaders. I 

offer critical events as an alternative causal mechanism and argue that protest campaigns in 

ethnically divided societies are particularly combustible as they have the potential to trigger 

unintended or unorchestrated communal violence. When such violence occurs, elite strategies 

change, mass support declines and the campaign demobilizes. An empirical investigation of the 

dynamics of the demobilization phase of the anti-foreigner protest campaign in Assam, India 

between 1979 and 1985 confirms this argument. A single group analysis is conducted to compare 

the dynamics of the campaign before and after the communal violence by using time series event 

data collected from The Indian Express, a national newspaper. The study has wider implications 

for the literature on collective action as it illuminates the dynamic and complex nature of protest 

campaigns. 

International Interactions 

This is the author's manuscript of the article to be published in final edited form at:  

Demirel-Pegg, Tijen (2016), “The Dynamics of the Demobilization of the Protest Campaign in 

Assam,” International Interactions. 

http://www.tandfonline.com/doi/full/10.1080/03050629.2016.1128430 

 



1 
 

Introduction 

The year 2014 was marked by intense protests against the police after the killings of 

Michael Brown in Ferguson and Eric Garner in New York and the subsequent failure to indict 

the police officers that killed them. While large scale protests were held across the United States, 

an unexpected event happened: two police officers were shot to death in New York on December 

20, 2014 by a man who was allegedly outraged by the Brown and Garner cases. The momentum 

for the Black Lives Matter movement in New York City was derailed as the police killings 

generated a backlash effect against demonstrators by critics and local leaders who maintained 

that it was not appropriate to keep protesting given the horrific nature of the police killings. 

Meanwhile, the New York police blamed the protesters for those killings. These events show that 

in the midst of a mobilization process, protesters can withdraw from the streets when unexpected 

events occur, and external pressure encourages them to stop demonstrating.  

This study highlights the role that critical events play in the demobilization of protest 

campaigns. In the context of an ongoing protest campaign, critical events may alter protesters’ 

tactics and their willingness to continue demonstrating, leading to the demobilization of the 

campaign. Typically, social movement scholars argue that protest campaigns demobilize as a 

consequence of a polarization between radical and moderate protesters or the cooptation of the 

campaign leaders. I offer critical events as an alternative causal mechanism and argue that 

protest campaigns in ethnically divided societies are particularly combustible as they have the 



2 
 

potential to trigger unintended communal violence.1 When such violence occurs, elite strategies 

change, mass support declines and campaigns demobilize. 

The connection between critical events and the demobilization of a campaign has 

remained largely under-theorized mainly because scholars have typically associated critical 

events with increased protest activity (Meyer and Staggenborg 1996, Rasler 1996).  For instance, 

Rasler (1996) shows that the burning down of the Abadan cinema resulting in 400 deaths was 

one of the important critical events that escalated the protests during the Iranian revolution. 

Similarly, Lohmann (1994) argues that President Gorbachev’s visit to East Germany in October 

1989 where he publicly urged East German leaders to implement reforms was seen as a signal 

that the Soviet military forces would not intervene if people started protesting.  In both cases, 

critical events are seen as having an escalatory effect on mass protests against the regime. 

Nonetheless, critical events can also have a de-escalatory effect. According to Staggenborg 

(1993:320), critical events are significant turning points that, whether intentional or not, alter 

expectations and perceptions of threats and opportunities, affecting the outcomes of protest 

campaigns. This definition allows for the possibility that critical events can lead to the 

demobilization of a protest campaign. For instance, the 1990/91 Gulf War was responsible for 

the demobilization of the First Intifada protest campaign. Until the Gulf War, Palestinian leaders 

                                                           
1 Communal violence  refers to an intense, sudden, though not necessarily wholly unplanned, 

lethal attack by civilian members of one ethnic group on civilian members of another ethnic 

group, with the victims chosen because of their group membership (Horowitz 2001:1).  The 

intention to kill and the civilian nature of the perpetrators and the victims are the defining 

characteristics of communal violence. 



3 
 

used nonviolent tactics to convey their message to the international community, knowing that 

Israel’s military response would evoke international criticism. The onset of the Gulf War, 

however, shifted international attention away from the campaign, and created an opportunity for 

the Israelis to clamp down on the Palestinian activists by imposing a total curfew, limiting the 

Intifada’s activities significantly (Alimi 2009:228-29). 

To investigate how critical events can demobilize a protest campaign, I conduct a 

quantitative analysis of a single case study, the anti-foreigner protest campaign in the northeast 

Indian state of Assam between 1979 and 1985. The critical event in the Assam protest campaign 

proved to be the unexpected rise of communal violence in February 1983.  By using original and 

fine-grained data collected from The Indian Express, a daily newspaper, I demonstrate that the 

1983 communal violence played a major role in the demobilization of the campaign in this 

ethnically divided state.  Though the communal violence was neither planned nor directed by the 

protesters, it had a significant impact on the campaign by compelling the campaign leaders to 

suspend protests temporarily, and leading to the withdrawal of one of the main ethnic groups that 

originally supported the campaign.  Consequently, protests declined dramatically, never reaching 

the levels they did during the first four years of the campaign.  

This study makes several important contributions to the theoretical literature.  First, it 

examines the link between critical events and the demobilization process of a protest campaign.  

Second, this research expands our understanding of how protest campaigns demobilize.  So far, 

scholars have mostly identified the use of violence or the creation of formal organizations as the 

main determinants of demobilization (Chenoweth and Stephan 2011, Nepstad 2011, Piven and 

Cloward 1977, Tarrow 1989).  This study identifies communal violence in ethnically divided 

societies as another causal explanation. In societies where multiple ethnic groups compete for 



4 
 

power, communal violence that occurs during a protest campaign, yet is not directed by the 

campaign leaders, is likely to lead to demobilization.  Communal violence pushes protest leaders 

to change strategies and results in the withdrawal of and rank and file support.  Third, this 

research brings a more nuanced understanding to the effects of violence on the demobilization of 

protest campaigns.  Traditional studies of nonviolent resistance campaigns assume that 

opposition groups choose to pursue nonviolent strategies and strictly adhere to such tactics even 

when faced with severe repression.  If they fail to maintain nonviolent discipline, then the 

campaign is more likely to fail to achieve its political goal as protesters leave the streets in the 

face of violence.  In short, the use of nonviolent tactics is an important factor that avoids the 

premature demobilization of the campaign before groups gain important concessions from the 

government (Chenoweth and Stephan 2011, Nepstad 2011, Schock 2013, Schock 2005, Sharp 

1973).  On the other hand, social movement scholars who have studied the dynamics of protest 

campaigns note that radical groups frequently use violence, and this development allows the state 

to distinguish the radicals from the moderates.  By making concessions to the moderates, state 

leaders are able to demobilize the protest campaign by coopting the moderates while isolating the 

radicals (della Porta and Tarrow 1986, Tilly 1978).  Both of these strands of research assume that 

violence is a tactical choice designed to achieve political goals.2 Yet, I argue that violence is not 

                                                           
2 My understanding of protest campaign is similar to that of social movement scholars in that it 

involves both types of actions: nonviolent ones as well as low level forms of violence, short of 

organized military operations.  While demonstrations, boycotts, or strikes are nonviolent tactics 

typically used during protest campaigns, the potential for a violent outbreak exists as people 

might throw stones, break windows, or set vehicles on fire.   



5 
 

necessarily strategic, and when it is unorchestrated in the context of a protest campaign, it can 

produce a shock that may demobilize a protest campaign, as happened in Assam.  Thus, the aim 

of this study is to examine protest campaigns and improve our understanding of their dynamics 

in terms of how and when the scale and scope of protest actions escalates or de-escalates.  

 

Theoretical Foundations: The Demobilization of Protest Campaigns and Critical Events 

 Scholars who have studied protest campaigns frequently suffer from two limitations.  

First,  scholars have mostly focused on the factors that predict success and have only included 

failed nonviolent campaigns as negative cases to identify the conditions for the achievement of 

the campaign’s stated goals (Chenoweth and Stephan 2011, Nepstad 2011, Schock 2005).  This 

approach tends to overlook dynamics that are unique to unsuccessful outcomes.  Second, 

scholars of protest campaigns usually pay more attention to the origins and spread of 

mobilization rather than the demobilization process (Koopmans 1997, McAdam 1983, Tarrow 

1998, Tilly 1978).  Although demobilization is a crucial process that occurs within the trajectory 

of protest campaigns, it has attracted surprisingly little theoretical attention.3  Of the few scholars 

                                                           
3 Demobilization refers to the process by which protesting decreases in scale and scope and 

eventually ends (Tilly and Tarrow 2007).  While demobilization typically evokes an 

unsuccessful outcome, these two concepts are distinct.  Although the demobilization of a 

nonviolent campaign often leads to its failure, demobilization might also precede a successful 

outcome. For example, the demobilization of the Baltic and Crimean Tatar protests in 1987 and 

1988 under coercive measures eventually led to a successful outcome as they triggered other 

protests, ending with the fall of the Soviet Union (Beissinger 2002). 

 



6 
 

that do study demobilization, the focus is on the cooptation or polarization processes.  For 

instance,  Piven and Cloward (1977) argue that the development of formally structured 

organizations greatly increases the prospects for elite influence and the cooptation of  campaign 

leaders, thus, diminishing the campaign’s disruptive potential.  Others such as Tarrow (1989) and 

della Porta and Tarrow (1986) maintain that demobilization is the inevitable outcome of the 

expansion of a protest campaign and is largely driven by the polarization between moderates and 

radicals.  

I advance a causal mechanism for demobilization that is different from either cooptation 

or polarization.  I suggest that protest campaigns can demobilize because of critical events that 

constrain the leadership’s protest strategies.  Critical events are important turning points that 

affect public and elite attention to campaigns (Staggenborg 1993:320).  

Scholars have examined critical events in different contexts in political science. For 

instance, several International Relations scholars have studied the impact of critical events, or 

shocks in the de-escalation or termination of strategic rivalries (Goertz and Diehl 1995, 

Rasler,Thompson and Ganguly 2013).  While critical events are not sufficient to terminate or de-

escalate rivalries, they can contribute to a process of de-escalation of rivalries when political 

leaders take advantage of the uncertainty caused by the disruption of routine expectations 

(Rasler, et al. 2013:13-19).  Scholars of Historical Institutionalism also have sought to 

understand institutional change via critical junctures (Collier and Collier 2002, Mahoney 2001, 

Polanyi 1957).  Critical junctures relax structural influences (economic, cultural, organizational, 

and ideological) on political action significantly for a short period of time, widening the options 

available to political actors.  The consequences of powerful actors’ decisions during critical 

junctures typically have long term consequences (Capoccia and Kelemen 2007).  Both strands of 



7 
 

research, however, mostly emphasize the relaxation of structural constraints and the widening of 

political options, rather than structural changes that narrow the options available to political 

actors.  

 While studies of protest campaigns have examined the escalatory impact of critical events 

(Rasler 1996, Shawki 2012), with the notable exception of Staggenborg (1993), few scholars 

have provided theoretical guidelines as to how to identify critical events. Staggenborg suggests 

six types of critical events based on the extent of control the actors exert: large-scale 

socioeconomic and political events, natural disasters and epidemics, accidents, critical 

encounters, strategic initiatives, and policy outcomes (Staggenborg 1993:322).  Large-scale 

socioeconomic and political events such as wars and depressions; natural disasters and 

epidemics; and industrial or nuclear accidents are critical events which neither campaign actors 

nor their opponents exert much control over.  In contrast, strategic initiatives and policy 

outcomes are orchestrated by the campaign actors or their opponents.  In strategic initiatives, 

campaign actors or the counter-campaign create events designed to mobilize support.  Policy 

outcomes are official state responses to mobilization.  States might adopt favorable or 

unfavorable policies that affect levels of support for campaigns.    

 Critical encounters, which the communal violence in Assam falls into, are events that 

involve face-to-face interactions or physical contact between authorities or other actors who are 

somewhat associated with the campaign.  Such events occur when other social actors attempt to 

assert control using methods ranging from ridicule to violence, causing a sudden change in the 

trajectory of the campaign (Staggenborg 1993:327-28).  Staggenborg argues that the level of 

control in critical encounters ranges from full control by the state or campaign leaders to little or 

no control.  Thus, critical encounters might include events that occur during the campaign with 



8 
 

the full knowledge and sanctioning of the elites or campaign leaders as well as events that are 

unauthorized or not approved by them.  The shootings of the two police officers in New York 

during the Black Lives Matter campaign in the United States, for instance, happened within the 

context of the campaign but were not sanctioned by the elites or the campaign leaders.  Similarly, 

in Assam, the killings of the mostly Muslim peasants by Assamese villagers during the protest 

campaign were critical encounters that were beyond the campaign leadership’s control.  While 

the campaign leaders had called for a boycott of the 1983 State Assembly elections that 

increased violence associated with protests in major towns, the violent attacks in rural areas were 

not controlled by the leaders of the Assam protest campaign (Kimura 2013:7).  Therefore, the 

communal violence in Assam is distinct from the protest campaign in the sense that neither the 

protesters nor the campaign or the state leadership planned it.  Nonetheless, the communal 

violence occurred in rural areas when ethnic tension was significantly elevated in the region.  

 

The Causal Mechanism 

In societies where communal divisions are salient, opposition leaders broaden the base of 

resistance to include members of different ethnic groups.  Nonetheless, the inclusion of different 

ethnic groups makes the unity of the campaign precarious particularly if campaign activities 

highlight ethnic cleavages (Horowitz 2001, Svensson and Lindgren 2011).  Mass actions can 

exacerbate tensions by provoking fear or anger between different communal groups.  When an 

ethnic group demonstrates strength by protesting in large numbers, others feel threatened 

especially if the protesters’ claims have ethnic overtones.  In such instances, protest campaigns 

become highly combustible, increasing the chances of communal violence.  For instance, Uzbek 

demonstrations over ethnic claims to land and housing led to communal violence in Kyrgyzstan 



9 
 

in 1990.  Similarly, the Azeri protests against Armenian demands for the unification of Karabakh 

with Armenia triggered communal violence, resulting in mass killings of Armenians (Horowitz 

2001, Varshney 2002).  

Although communal violence occurs outside the control of the protest leadership, it can 

dramatically affect the trajectory of the campaign.  Opposition elites face pressure to shut down 

the campaign to prevent more ethnic violence.  Moreover, protesters that are members of the 

ethnic groups targeted during the violence withdraw their support from the campaign as a result 

of the increased prominence of ethnic allegiances.  The leaders also face the challenge of finding 

ways to reinvigorate enthusiasm and maintain solidarity particularly when political elites or the 

media call for a slowdown to give the communities time to heal.  Even if the campaign resumes, 

the loss of momentum leads to a decline in protest activity and lower levels of support.  A 

striking example is the collapse of the civil disobedience campaign in India as a result of Hindu-

Muslim violence in the early 1920s.  As the British were fighting against the Ottoman sultan 

during World War I, the Muslims in India turned against the British, providing Gandhi the 

opportunity to unite Hindus and Muslims in a civil disobedience campaign for the freedom of 

India.  Even though Muslim-Hindu unity was created for some time, in mid-1921, Muslim 

clerics in Malabar in the state of Kerala started to protest against the British treatment of the 

caliph in Turkey.  These protests first turned into a violent rebellion against the British, and then 

degenerated into Hindu-Muslim violence (Varshney 2002:142), creating a Hindu-Muslim divide 

within the Indian national movement across the country.  Gandhi and the Congress Party were 

unable to maintain the campaign as communal divisions became too obvious.  Consequently, 

Gandhi ended the campaign in 1922 and the Congress Party withdrew from mass mobilization 

until 1929  (Varshney 2002:134-35). 



10 
 

Communal violence leading to the demobilization of a protest campaign differs from 

previously identified causal pathways in the literature.  For Piven and Cloward (1977) and 

Tarrow (1989), the broadening of the support base leaves campaigns vulnerable to internal 

divisions, resulting in a withdrawal of support, and finally, demobilization.  However, the factors 

that trigger the withdrawal of public support are different.  According to Piven and Cloward 

(1977), the expansion of the protest campaign prompts the government to coopt the opposition to 

channel their disruptive behavior into organized forms of contention.  Yet, campaign leaders 

usually overestimate their ability to keep the campaign strong and effective via organizations 

while simultaneously underestimating the amount of resources needed to maintain an 

organization.  As leaders get isolated from the protesters, disruptive behavior declines, leaving 

the remaining agitated protesters discouraged and vulnerable to repression.   

Alternatively, Tarrow (1989) highlights the competitive dynamics among groups and its 

potential to lead to internal divisions before governments can coopt campaign leaders.  When 

protest campaigns first emerge, disruptive protests diffuse to different locations and segments of 

the society.  Competition for mass support intensifies as both established groups, such as trade 

unions or political parties, and new groups join the campaign.  Nonetheless, the early enthusiasm 

for protesting wanes over time and the personal costs of participation wear people down.  The 

decline in participation further encourages competition for mass support.  While moderates lead 

the shift toward conventional forms of collective action such as strikes and demonstrations, 

smaller and newer groups turn radical as they employ violent tactics to distinguish themselves 

from the moderates.  Similar to Piven and Cloward’s (1977) cooptation model, the state represses 

the radicals selectively while accommodating the moderates, reinforcing the polarization 

between these groups.  Faced with repression, the radicals resort to more violent tactics, resulting 



11 
 

in a further withdrawal of public support.  The split between the moderates and the radicals 

eventually leads to the demise of the protest campaign (Tarrow 1989).   

The causal mechanism I offer, however, allows for a disruption of the trajectory of 

protests by an unexpected and unorchestrated event.  As Figure 1 illustrates, if a critical event 

such as communal violence occurs in an ethnically divided society, the campaign leadership 

refrains from employing strategies that might exacerbate ethnic tensions. Governing elites 

usually assert their authority and criticize campaign leaders for mobilizing ethnic political 

demands and bringing ethnic issues to the forefront.  In addition, political elites often pressure 

campaign leaders to suspend their activities in a highly tense ethnic and political environment to 

prevent the outbreak of more violence.  As Horowitz (2001:286) suggests, ethnic groups can be 

provoked more easily to engage in large scale violence during mass demonstrations or 

processions, particularly if the core issues closely relate to ethnic demands.  Given the elevated 

tension as well as the political elite’s pressure in the aftermath of communal violence, campaign 

leaders will typically avoid organizing protests.   

Furthermore, as communal violence highlights the divisions within the society, ethnic 

groups, particularly those that have been victimized by the violence, withdraw their support from 

the protest campaign as it becomes more difficult to act collectively across ethnic lines. When 

the protest campaign secures the support of different ethnic, religious, and/or socioeconomic 

groups, protest leaders face challenges in keeping these different groups together after a major 

shock like communal violence.  If one of the groups included in this broad coalition has 

affiliations to the group that has been targeted, campaign solidarity declines. The group affiliated 

to the victims questions the purpose of the campaign and pressures the leaders to modify 

campaign policies that undermine the interests of the targeted groups. Other groups also make 



12 
 

demands to strengthen their profile, further undermining the campaign’s broad coalition. Thus, 

the ability of the leaders to organize mass events decreases. Furthermore, the leaders spend more 

time and energy into finding ways to re-energize the campaign and rebuild solidarity.  

 

 

 

FIGURE 1 Causal Pathway for the Demobilization of a Protest Campaign 

 

Assam and the Anti-Foreigner Protest Campaign, 1979-1985 

Assam is an ethnically heterogeneous society which has also witnessed large scale 

immigration from many parts of the subcontinent (particularly East Bengal) over the last hundred 

years (Baruah 1986).  According to the 1971 census records (before the anti-foreigner protest 

campaign started in 1979) 72.5% of the population in Assam was Hindu, 24.6% was Muslim, 

and 2.61% was Christian.  The Muslim population in Assam includes ethnic Assamese and 

DEMOBILIZATION 

Protest Campaign 

in Ethnically 

Divided Societies 

Critical Event: 

Communal 

Violence 

Changes in elite 

strategies 

Withdrawal of 

mass support 



13 
 

Muslims of Bengali descent while Hindus also include ethnic Assamese and Hindu Bengalis that 

immigrated to Assam.  People are divided along linguistic lines, which, simultaneously also 

stands for the division between “indigenous” and “immigrant” groups.  As of 1971, 

approximately 68% of the population in Assam spoke languages or dialects that are considered to 

be of “indigenous” people while 29 % of the population spoke “immigrant” languages or dialects 

(Baruah 1986:1188). 

 The protest campaign in Assam emerged to drive out both Hindu and Muslim 

immigrants (Kimura 2013).  Threatened by rapid population growth following a large influx of 

Muslim immigrants from Bangladesh and concerned by the growing political power of the 

middle-class mostly Hindu Bengali immigrants, ethnic Assamese started a state-wide protest 

campaign in 1979 to pressure the Indian Government to identify illegal immigrants, delete their 

names from voting lists, and deport them to Bangladesh.  The campaign was led by the All 

Assam Students Union (AASU) and the All Assam Gana Sangram Parishad (AAGSP), an 

umbrella organization formed in 1979 to direct the protest activities against illegal immigrants 

(Baruah 1999).  While the campaign leaders demanded the deportation of all those who entered 

after 1951, the Indian state felt a humanitarian commitment toward the Hindu refugees who 

crossed into India from East Pakistan to escape the atrocities of the Pakistani state before it 

became Bangladesh in 1971.  Therefore, the Indian Government insisted that 1971 should be the 

baseline year, allowing foreigners who entered before that year to stay (Kimura 2013).  

The campaign gained mass support among ethnic Assamese, including indigenous 

Muslims who had settled in Assam before colonization.  The leaders of the campaign made a 

deliberate attempt to frame the immigration issue along secular lines by highlighting the 

illegality of it rather than identifying certain ethnic immigrants, especially the Muslims, as the 



14 
 

unwanted population (Baruah 1986, Kimura 2013).  Thousands of students, government 

employees, peasants and workers participated in demonstrations, strikes, and sit-ins throughout 

the first four years.  The government responded with a mix of repression and accommodation as 

it used force against the protesters at times while also periodically negotiating with the campaign 

leaders.  In early 1983, the campaign intensified in the wake of the state legislative assembly 

elections scheduled for February.  The central government insisted on using the electoral rolls of 

1979 without deleting the names of foreigners while the campaign leaders strongly opposed and 

called for an electoral boycott.  Both nonviolent and violent protests intensified before the 

elections as many Assamese citizens supported the boycott.  In addition to the strikes aimed at 

preventing election preparations and the filing of nominations, the AASU and the AAGSP 

organized a general strike that was supported by state employees.  As clashes between the police 

and the protesters increased, the Assam government brought in additional police from other 

states.  In the midst of this tense political environment, communal violence occurred when 

Assamese and Tiwa (an indigenous minority group) peasants in the rural areas attacked mostly 

Muslims of Bengali descent (Kimura 2013).  The first large-scale ethnic violence took place in 

the Darrang District on the northern bank of the Brahmaputra River on February 11.  One week 

later, about 1,600 to 2,000 Muslims of East Bengali origin were killed in several villages around 

Nellie in the Nagaon District (Kimura 2013:68).  According to Kimura (2013), the Nellie 



15 
 

incident was one of the largest incidents of ethnic violence in India’s rural areas since India 

gained its independence in 1947.4  

After the communal violence, the AASU and AAGSP suspended the campaign for a few 

months, resuming protest activities in August 1983.  Yet, protest activity remained at 

dramatically lower levels after the ethnic violence of 1983.  As a result of extensive negotiations, 

the campaign leaders and the Indian Government signed the Assam Accord in August 1985, both 

agreeing to disenfranchise all the foreigners that entered Assam between 1966 and 1971 for ten 

years and deporting those that came after 1971 (Baruah 1999).  Figure 2 illustrates the trajectory 

of the protest campaign and the occurrence of communal violence in Assam in 1983.5 

 

                                                           
4 Smaller scale communal violence in 1980 in Assam is not considered as a critical event as it did 

not change the overall trajectory of the protest campaign. See Capoccia and Kelemen (2007) for 

a discussion of identification of critical events.  

5 This figure is generated by using the count of events.  



16 
 

 

FIGURE 2  Weekly Protest Activity Before and After Communal Violence in Assam 

 

Assam as a “Least Likely” Case 

Assam is an ideal case to examine the effects of communal violence on a protest 

campaign because it is a “least likely” case, meaning that it is harder to prove the argument as 

opposed to a case that easily fits the argument (George and Bennett 2005).  First and foremost, 

communal violence took place in the spring of 1983, but the campaign ended in August 1985, 

more than two years later.  In most other instances, communal violence ends protest campaigns 

rather abruptly, as in the case of the Tamil protests against the declaration of Sinhalese as the 

sole official language in Sri Lanka in 1958 (DeVotta 2004).  In Assam, though, protests actually 

resumed a few months after communal violence and continued intermittently for two more years.  

Therefore, demonstrating the demobilizing effects of communal violence is more challenging in 

0

20
0

40
0

60
0

80
0

10
00

12
00

14
00

16
00

18
00

20
00

N
u
m

b
e
r 

o
f 
C

o
m

m
u

m
a
l 
V

io
le

n
c
e
 E

v
e
n

ts

0

10

20

30

40

50

60

70

80

90
10

0
11

0
12

0

N
u
m

b
e
r 

o
f 
P

ro
te

s
t 
C

a
m

p
a

ig
n
 E

v
e
n

ts

Ja
nu

ar
y 
19

79

Ja
nu

ar
y 
19

80

Ja
nu

ar
y 
19

81

Ja
nu

ar
y 
19

82

Ja
nu

ar
y 
19

83

Ja
nu

ar
y 
19

84

Ja
nu

ar
y 
19

85

Week of Event

Protest Campaign Communal Violence



17 
 

the Assam case.  Moreover, the campaign in Assam was well-established since 1979, long before 

the ethnic violence took place.  Protests were organized regularly and were usually well attended.  

A strong campaign like the one in Assam should be less vulnerable to the negative effects of 

communal violence than other more ad hoc campaigns like the one in Nigeria after the re-

election of Goodluck Jonathan in 2011.  Therefore, a demonstration of the decline in protest 

activity as a result of communal violence in Assam should provide strong support for the 

demobilizing effect of ethnic violence in the context of ongoing protest campaigns.  

 

Research Design 

Data6 

The analysis is based on data collected from The Indian Express, a national daily 

newspaper, from January 1, 1979 to August 16, 1985.  The data set covers both nonviolent and 

violent forms of protest action such as demonstrations, strikes and clashes with the police, as 

well as incidents of communal violence.7  Events were coded for the date, actor, location, target, 

number of participants, number of injured and number of deaths.8  The state’s repressive and 

accommodative actions were also coded for the same time period.  Events that lacked 

information on the exact date, actor, and type of action were omitted from the data set.  In total, 

                                                           
6 The codebook used for this data set is posted in the Appendix. 

7 Ethnic violence is distinguishable from protest violence as communal violence occurred in rural 

areas whereas protest activity typically took place in major cities. Ambiguous cases were 

excluded.  

8 Ambiguous events were excluded. 



18 
 

8,488 events9 were hand coded by the author10 based on a list of events compiled mostly from 

Krain (2000) and Moore and Lindstrom’s (1996) Violent International Conflict Data Project 

(VICD) and weighted by a scale that is based on Krain’s (2000) scale for opposition and state 

actions ranging from 1 to 25.11  A detailed explanation of the weighting scheme and its 

modification of Krain’s scale can be found in the Appendix.  The data were then aggregated in 

weekly intervals. 

The underrepresentation of the frequency of the actual events, the overreporting of 

violent events and the selective reporting of events are three obvious issues that could potentially 

challenge the validity of the data as a consequence of coding from a single newspaper.  In terms 

of underrepresentation, The Indian Express had correspondents based in Assam throughout the 

protest campaign directly reporting from the area and providing detailed information regarding a 

wide range of collective actions, thus reducing the chance of underrepresentation.  Moreover, the 

data illustrate similar trends to those found by scholars who have studied the anti-foreigner 

protest campaign and communal violence in Assam (Baruah 1999, Kimura 2013), somewhat 

minimizing concerns over validity.  Regarding the issue of overreporting violence and 

underreporting nonviolence (Day,Pinckney and Chenoweth 2015), I used secondary sources 

                                                           
9 Every action of every actor was coded separately.  For instance, if students shouted slogans and 

threw stones, two separate actions were coded. Clashes are coded for both protesters and the 

police. 

10 Since only one coder was involved, potential issues of inter-coder reliability do not arise. 

11 The correlation coefficients for all variables between the number of events and the weighted 

data range from .87 to .98.  



19 
 

(Baruah 1986, Baruah 1999, Kimura 2013) to check the accuracy of the reporting whenever 

possible.  Nonetheless, given the media attention on Assam and the protesters after the mass 

killings, underreporting of nonviolence seems unlikely.  Finally, selective reporting of events 

might introduce bias to the data because reporters do not (and cannot) report every single 

collective action that occurs in the public domain.  Therefore, they are left to make decisions 

regarding what to report and how much detail to present.  The concern is that newspapers report 

events that are likely to attract more attention and underreport minor events, thus, disregarding 

the vast majority of contentious activity (McCarthy,McPhail and Smith 1996).  Nonetheless, 

others have argued that selective reporting is an issue when and where protests occur frequently.  

In locations like Assam in the late 1970s, when protests were less frequent, they are more 

newsworthy and receive greater coverage (Earl,Martin,McCarthy and Soule 2004:71).  

Moreover, the longitudinal design of the study reduces potential challenges to the internal 

validity of the analyses since selective reporting would be an issue both before and after the 

communal violence.  

 

Model and Estimation 

To understand the effect of communal violence on protest activity, I conduct a single 

group time series analysis that is commonly used in epidemiology, which compares a pre-

intervention time series with a post-intervention time series (Linden and Adams 2011).  

Accordingly, I model communal violence as an “intervention” on the protest activity series, the 

dependent variable, to examine two types of changes in protest activity: the immediate change in 

protest activity right after the first occurrences of communal violence and the long term change 

in the trajectory of protest activity until the end of the campaign.  



20 
 

Protest activity, the dependent variable, is the total weekly score of nonviolent and 

violent forms of collective action including demonstrations, strikes, boycotts, and arson.  The 

minimum values of protest typically reflect the weeks when protesters use few nonviolent 

tactics; while the maximum values reflect the weeks when protesters resort to nonviolent tactics 

more frequently and/or resort to violent tactics such as clashing with the police.  Hence, the 

demobilization phase represents the period before the end of the campaign when the values of 

the dependent variable are declining, indicating that protesters used both nonviolent and violent 

forms of tactics less frequently.  

To assess the changes of the communal violence on protest activity, I construct a time 

variable denoting the number of weeks from the start of the campaign until its end.  The 

coefficient of this variable represents the trajectory of protest activity until the communal 

violence begins.  The intervention variable is a dummy variable coded 1 for the weeks the 

communal violence are on (weeks 7 to 12 in 1983) and 0 otherwise.  The coefficient of 

intervention is the intercept at the time when communal violence occurs and indicates any 

immediate changes in protest activity right after the communal violence.  Finally, a post-

intervention time variable is constructed to denote the number of weeks after the communal 

violence.  The coefficient of this variable represents the change in the slope of protest activity.  If 

the communal violence indeed has a demobilizing effect, then the coefficients of either 

intervention or post-intervention time, or both should be negative and have significant p values.  

A significant negative coefficient for intervention would signify a sudden drop in protest activity 

and a significant, negative slope after intervention would suggest a decay, which could be steep 

or gradual, depending on the slope.  



21 
 

I also control for state actions as both repression and accommodation have the potential 

to affect protest activity.  While several studies have shown that repression increases protest 

activity by creating a backfire effect (Chenoweth and Stephan 2011, Francisco 1995, Rasler 

1996), others have found that carefully targeted and consistent repression is likely to demobilize 

protesters (Beissinger 2002, Tarrow 1989).  Repression is the total weekly score of repressive 

actions carried out by the lower level authorities, typically the police, directly confronting the 

protesters, such as blocking the roads; as well as those sanctioned by higher level authorities 

including the imposition of curfews or restrictions on civil liberties.  

Similarly, the state’s accommodative tactics can have a positive or negative effect on 

protests.  Accommodation can increase protest activity as it signals weakness (Rasler 1996); or it 

can lead to a decline in protests as it creates splits between the moderates and the radicals 

(Tarrow 1989) or coopts the protesters (Piven and Cloward 1977).  Accommodation is measured 

as the total weekly score of state concessions that include actions like negotiations, removing 

restrictive laws and regulations, or withdrawing the army from the region.12 

I estimate both a Zero-Inflated Negative Binomial Regression (ZINB) and a Poisson 

autoregressive (PAR) model of order (p) to examine the effects of communal violence on the 

protest campaign.  Both ZINB and PAR are count models that are appropriate for the nature of 

the dependent variable in this analysis.  A ZINB analysis is appropriate for count models that are 

overdispersed13 and have a high number of zero observations of the dependent variable (Long 

                                                           
12 Descriptive statistics for the variables can be found in the Appendix. 

13 The variance of protest activity is greater than its mean, indicating overdispersion. 



22 
 

1997).14  The data are also ordered in time, meaning that past values influence future values.15 As 

Brandt et al (2000) show, estimating event counts that are time dependent with a standard count 

model might lead to inefficient estimates.  The systematic dynamics of the data have to be taken 

into account while working with time series event count data such as the series on the Assam 

protest campaign, which a PAR analysis is able to do. The PAR model is suitable for cyclical 

and short-memoried processes that are mean reverting (Brandt and Williams 2001).  As Figure 2 

demonstrates, the protest activity series has a cyclical pattern as the peaks in the series are 

followed by declines.  The plot for the Autocorrelation Function (ACF) of the series (Figure 3) 

shows autocorrelation only in the first four lags indicating that the process has a short-term 

memory.  Therefore, I use the PAR model developed by Brandt and Williams (2001) to account 

for the time-series nature of the data and a ZINB to deal with the high frequency of zero 

observations.  

                                                           
14 Vuong test statistics indicate that ZINB is preferable to a regular negative binomial.  

15 A transfer function model assessing the impact of intervention on a time series is not suitable 

as the data are based on event counts (Brandt and Williams 2001). 



23 
 

 

FIGURE 3  Autocorrelation Function Plot for Protest Activity 

 

Findings  

Table 1 shows the results of the ZINB and PAR (2)16 analyses for the protest activity in 

Assam between January 1979 and August 1985.17  The Critical Event Model with the single 

                                                           
16 The AR parameters in larger lags were insignificant.  

17 The same analyses were conducted on the unweighted data and the results are similar.  Also, 

the Critical Event Model was run on nonviolent and violent protest activity separately. Overall, 

the results still confirm the findings of the Critical Event Model.  The short-term positive impact 

of communal violence on violent protests can perhaps be explained due to the radicals’ initial 

-0
.2

0

0.
00

0.
20

0.
40

0.
60

A
u
to

c
o

rr
e
la

ti
o

n
s
 o

f 
P

ro
te

s
t 
A

c
ti
v
it
y

0 10 20 30 40
Lag

Bartlett's formula for MA(q) 95% confidence bands



24 
 

group analysis controlling for repression and accommodation suggests that protest activity 

decreased significantly when communal violence began in the middle of February 1983.  Both 

intervention and post intervention are statistically significant (p<.001 and p<.05 respectively) in 

the ZINB analysis.  The PAR results are also similar as both variables significantly decrease 

protest activity, indicating that protest activity declined immediately after communal violence 

began and continued to decline in its aftermath. 

 

TABLE 1 ZINB and PAR Estimates of Protest Activity in Assam 
 

Variable 

 

Critical Event Model 

 

Protest Violence 

Model 

 

Pre-Communal 

Violence Model 

 

Post-Communal 

Violence Model 

 ZINB PAR ZINB PAR ZINB PAR ZINB PAR 

Time   0.001 

 (0.002) 

  0.006*** 

 (0.002) 

 -0.006** 

 (0.002) 

-0.001 

(0.002) 

 

- 

 

- 

 

- 

 

- 

Intervention 

(Communal 

Violence) 

-2.774*** 

 (0.657) 

-2.090*** 

 (0.529) 

 -1.598** 

 (0.491) 

-1.054† 

(0.621) 

 

- 

 

- 

 

- 

 

- 

Post-Intervention  -0.015* 

 (0.006) 

-0.036*** 

(0.006) 

  0.006 

 (0.006) 

-0.018** 

(0.006) 

 

- 

 

- 

 

- 

 

- 

Repression  

 

  0.005*** 

 (0.001) 

 0.004*** 

(0.001) 

  0.003*** 

 (0.001) 

 

0.003***

(0.001) 

0.004** 

(0.001) 

 0.003*** 

(0.001) 

 0.004 

(0.003) 

 0.012*** 

(0.002) 

Accommodation  -0.012 

 (0.009) 

-0.033*** 

(0.008) 

 -0.009  

 (0.008) 

-0.019** 

(0.007) 

 0.019 

(0.012) 

-0.014†
 

(0.007)
 

-0.017 

(0.013) 

-0.005 

(0.042) 

High violence 

 

 

     - 

 

     - 

  0.003** 

 (0.001) 

 0.001  
(0.001) 

 

- 

 

- 

 

    - 

 

    - 

Constant  4.112*** 

(0.311) 

 3.003*** 

(0.246) 

  4.896*** 

 (2.502) 

 3.506** 

(0.205) 

4.379***

(0.205) 

3.925*** 

(0.129) 

 3.727*** 

(0.195) 

 2.325*** 

(0.242) 

AR parameter 1 

(lag one) 

 

     - 

 0.573*** 

(0.055) 

     

      - 

    

0.529***

(0.056) 

 

-  

  0.670*** 

 (0.071) 

   

- 

 0.440** 

(0.090) 

AR parameter 2 

(lag two) 

 

     - 

-0.123** 

(0.041) 

  

      -    

-0.084† 

(0.051) 

      

- 

 -0.081 

 (0.051) 

 

- 

 0.008 

(0.087) 

N    345     343 345 343 215 213     125   123 

Probability>chi-

square 

  

  0.000 

 

      - 

 

0.000 

 

- 

 

0.000 

 

- 

 

  0.106 

 

     - 

Adjusted R2      -    0.56 -   0.54 -    0.57  -   0.29 
† p<.10, *p<0.05, **p<.01, ***p<.001; All significant tests are two-tailed, standard errors in parentheses 

                                                           

desire to show their continued determination to carry on.  The results of robustness checks can be 

found in the Appendix. 



25 
 

These findings confirm that the communal violence led to a sudden drop in protest 

activity and continued to stifle the campaign until the Assam Accord was signed in August 1985.  

Table 2 illustrates the factor changes in the rate of protest activity for a clear interpretation of the 

coefficients.  Accordingly, protest activity decreases by a factor of 0.06 after the communal 

violence begins, and continues to decline by a factor of 0.99 after the communal violence is over.  

 

TABLE 2 Factor Change in Expected Rate of Protest Activity18 

 

  

Critical Event 

Model 

 

Protest 

Violence Model 

 

Pre-Communal 

Violence Model 

 

Post-Communal 

Violence Model 

 

Time 

 

- 

 

-0.99 

  

Intervention (Communal 

Violence) 

 

-0.06 

 

-0.20 

 

- 

 

- 

 

Post-Intervention 

 

-0.99 

 

- 

 

- 

 

- 

 

Repression 

 

1.01 

 

1.00 

 

1.00 

 

- 

 

Accommodation 

 

- 

 

- 

 

 0.98  

 

- 

 

High Violence 

 

- 

 

1.00 

  

 

 

 

The Critical Event Model also shows that repression leads to an increase in protest 

activity in both the ZINB and PAR analyses suggesting the existence of a backfire effect 

(Chenoweth and Stephan 2011, Francisco 1995, Rasler 1996).  This effect is relatively strong, as 

repression increases protest activity by a factor of 1.01, holding other variables constant (Table 

2).  This finding is particularly interesting because it suggests that the campaign continued to 

demobilize despite the backfire effect of repression.  Accommodation has a significant and 

                                                           
18 Only the expected rates of significant coefficients are listed 



26 
 

negative effect only in the PAR analysis.  The negative impact of accommodation indicates that, 

independent of communal violence, accommodative actions might have contributed to the 

demobilization of the protest campaign, confirming Piven and Cloward’s cooptation model or 

Tarrow’s polarization model.  To understand the effects of accommodation and repression more 

precisely, I examine the impact of these variables on protest activity during the pre- and post-

intervention phases in Models 3 and 4 respectively.     

In the Protest Violence Model, I control for high levels of protest violence as previous 

studies have shown that heavy protest violence leads to the demobilization of a campaign as 

participation declines when violence erupts and polarization between the radicals and the 

moderates occurs (Chenoweth and Stephan 2011, della Porta and Tarrow 1986, Tarrow 1989). 19 

For this model only, I construct a new dependent variable for protests that does not include 

heavy violence and create a new control variable, high violence representing acts such as 

abductions, armed attacks, and clashes with security forces that occur during the protest 

                                                           
19 Although studies have demonstrated that elections are associated with high levels of violence 

(Dunning 2011), elections did not have a significant impact on protest activity. Parliamentary 

elections in December 1979 and the State Assembly elections in February 1983 were the only 

two elections held in Assam during the time period under study. The 1984 parliamentary 

elections were postponed and held in December 1985. A dummy variable coded 1 for three 

months before and after each election was statistically insignificant in the ZINB estimates of the 

Critical Event Model.  



27 
 

campaign.20 The results of the ZINB and PAR analyses suggest that high violence is significant, 

but in the opposite direction. High violence does not seem to deter protests in Assam, and 

instead, increases it by a factor of 1.00 (Table 2).  Thus, the Protest Violence Model shows that 

high levels of violence during the elections was not the cause of the demobilization of the 

campaign.  The heavy violence in Assam was primarily associated with the protesters’ election 

boycott as anti-election groups, mostly ethnic Assamese, tried to intimidate pro-election groups 

(mostly Bengalis) from participating (Baruah 1999, Kimura 2013).   

  The Pre- and Post-Communal Violence models examine the effects of state actions 

before and after communal violence respectively to understand if state policies had differential 

effects across the two time periods. For instance, Indian leaders could have been more inclined to 

make concessions to the protesters after ethnic violence began to alleviate the tension in the 

region.  Alternatively, the leaders might have resorted to more repressive measures in order to 

crush the protesters and bring order to Assam.  Both ZINB and PAR analyses show that 

repression has a positive and significant effect on protest activity during the pre- and post-

communal violence periods, suggesting that the effect of repression remained relatively 

consistent across the protest campaign.  Accommodation has no significant effect in the pre and 

post-intervention period, indicating that the negotiations between the Indian government and the 

campaign leaders, and the attempts to coopt the opposition did not demobilize the campaign.  As 

Figure 1 also shows, protest activity had already declined dramatically when the parties agreed to 

participate in the last round of negotiations, which started in April 1985, almost two years after 

                                                           
20 The full range of events are listed under the High Intensity Collective Action category in the 

Codebook posted in the Appendix. 



28 
 

the ethnic killings took place.  Thus, neither model suggests that sate actions led to the decline in 

protest activity. 

Predictive margins generated based on the ZINB analysis of the Critical Event Model 

provide a more nuanced understanding of the effect of the communal violence on protest activity 

(Figure 4).  The predicted number of weekly protest events declines steadily over the course of 

time. As Figure 4 illustrates, the predicted number of protest events is 60 immediately after the 

communal violence. About six months later, however, the number of protest events goes down to 

40, and in about a year, the number is below 30, indicating a more than 50% decrease in protest 

activity. In approximately two years after the communal violence, the predicted number of 

protest events further declines to slightly above ten a week.  When these numbers are put into 

context, the first year in particular seems quite significant. One year after the communal 

violence, none of the issues the protesters were campaigning on had been resolved and the 

leaders of the campaign were still trying to mobilize support. So, a more than 50% decline in the 

first year suggests that the communal violence had a significant effect on the demobilization of 

the campaign.  



29 
 

 

 

FIGURE 4 Predicted number of Events after the Communal Violence 

 

To evaluate the predictive power of the Critical Event Model, I generated out-of-sample 

forecasting, which is an ideal tool for cross-validation.  The main idea is to compare the 

predicted values generated by the model to the observed values.21  Moreover, these forecasts can 

be easily compared to those of other models.  Since polarization did not occur in the Assam case, 

forecasts on a polarization model  (Tarrow 1989) cannot be tested. 22  However, because the 

                                                           
21 Generating predicted values of observations that are not included in any of the analyses is 

another option for cross validation.  

22 The United Liberation Front of Assam (ULFA), a separatist insurgent group, was established 

in 1979 but did not actively challenge the state until the late 1980s. 

0

10

20

30

40

50

60

70

80

P
re

d
ic

te
d
 N

u
m

b
e

r 
O

f 
E

v
e
n

ts

0 10 20 30 40 50 60 70 80 90 10
0

11
0

12
0

Number of Weeks After Communal Violence

95% Confidence Interval Band



30 
 

Indian government used accommodative tactics to coopt the protesters, I am able to generate out-

of-sample forecasts based on a Cooptation Model that has the accommodation variable only 

(Piven and Cloward 1977) and compare them to the out-of-sample forecasts of the Critical Event 

Model.  

To obtain the out-of-sample forecasts, I first run ZINB regressions to estimate the 

parameters for both models by holding back the weeks after the communal violence ends.23  

Based on these estimates, I then generate out-of-sample probabilities of protest activity for the 

period after the critical event and calculate Theil’s U to evaluate the predictive power of each 

model.  Theil’s U is a ratio where the numerator is the sum of squared prediction errors and the 

denominator is the sum of squared prediction error from a naïve model, which expects no change 

from one observation to the other (Farnum and Stanton 1989:26).  Values below one indicate a 

better performance of the statistical model compared to the naïve model.   

The plots in Figure 4 illustrate the number of observed protest events and the out-of-

sample predictions based on the Critical Event and the Cooptation Models after the intervention. 

Even though neither one of the models predict protest activity very accurately, the Critical Event 

Model performs much better than the Cooptation Model. 24 The Theil’s U statistics for the 

Critical Event and the Cooptation Models are 0.99 and 3.36 respectively, indicating that although 

neither greatly outperforms the naïve model, the Critical Event Model is still a better predictor of 

                                                           
23 I hold back the weeks after June 1983, approximately 33% of the observations, after the 

communal violence. These estimates can be found in the Appendix. 

24 The AIC scores for the Critical Event and Cooptation models are 1427.15 and 1535.25 

respectively. 



31 
 

protest activity during the demobilization period.  As Figure 4 shows, the Cooptation Model 

predicts higher values of protest activity while the Critical Event Model predicts no activity 

during the demobilization period.25 

 

 

FIGURE 5 Out-of-sample Forecasts for the Critical Event and Cooptation Models 

 

Overall, the empirical analyses show that the communal violence in February 1983 led to 

the demobilization of the protest campaign.  Until the communal violence, the campaign was 

                                                           
25 The Critical Event plot suggests that the variables present in the model are not sensitive 

enough to protest activity occurrence to be able to detect those spikes. This might be due to the 

inability of the model to account for the autoregressive component of the series.  



32 
 

strong and had most of the characteristics identified by scholars of nonviolent resistance as likely 

to lead to a successful outcome.  First, the campaign started off by employing nonviolent tactics 

and secured substantive mass support from ethnic Assamese as well as from non-immigrant 

Muslims, the most important factor for pressuring the government to change its unfavorable 

policies according to Chenoweth and Stephan (2011).  The campaign leaders mobilized 

thousands of people, including students, teachers, peasants, state government officials and oil 

workers and brought them out into the streets.  For instance, in November 1979, nearly two 

million people Assam participated in the civil disobedience campaign (Baruah 1999).  Second, 

the high level of participation allowed for tactical innovation throughout the first four years, 

increasing the campaign’s resilience in the face of state repression (Chenoweth and Stephan 

2011, McAdam 1983, Schock 2005).  For instance, in late December 1979, shortly after 

launching a civil disobedience campaign, the protesters imposed an economic blockade, stopping 

the transportation of crude oil, plywood, and jute from Assam to the rest of the country (Indian 

Express, December 30, 1979).  Moreover, many state police and government officials supported 

the campaign against illegal immigrants, dramatically reducing the level of repression used 

against the protesters (Kimura 2013).  According to Nepstad (2011) and Chenoweth and Stephan 

(2011), such loyalty shifts are crucial for the success of a campaign.  Finally, potentially more 

radical supporters of the campaign did not have a violent insurgent group to turn to (Chenoweth 

and Stephan 2011), keeping the mass support for the campaign relatively intact.  In short, the 

protest campaign in Assam featured several mechanisms that should have enabled the leaders to 

continue to mobilize people and put pressure on the government before communal violence 

began. 



33 
 

So, how exactly did the communal violence set the demobilization process of the 

campaign in motion?  Communal violence turned into a critical event that significantly limited 

the campaign leader’s available strategies.  After the extent of the violence became known, 

members of the Indian central government repeatedly called for peace and calm in the region 

(The Indian Express, February 22 & March 15, 1983) while political actors and civil society 

groups offered competing explanations for the occurrence of such brutal violence.  The campaign 

leaders blamed the Indian Government for imposing the elections without changing the electoral 

rolls (Kimura 2013) whereas several members of the government accused the protesters of 

inciting violence in the region.  For instance, right after the most publicized mass killings in 

Nellie on February 18, Prime Minister Indira Gandhi publicly blamed the protest leaders for the 

violence in Assam (The Indian Express, February 22, 1983).  The campaign leadership was also 

criticized for their failure to condemn violence immediately after the communal violence started 

(Gandhi 1983). When the temporary suspension of the campaign was finally announced in late 

March, the leaders emphasized the need to focus on relief, rehabilitation, and integration (The 

Indian Express, March 28, 1983).  Even though protests resumed in August 1983, the intensity of 

the campaign decreased dramatically (Figure 2). 

The communal violence further entrenched ethnic divisions within the society alienating 

ethnic Assamese and immigrant Muslims (Gupta 1983). The tension among these two groups 

had important consequences for the campaign.  After the extensive killings of Muslim peasants, 

many Muslim protesters withdrew their support, including those who had settled before 

colonization and, hence, were seen as indigenous.  For instance, one of the top Muslim student 

leaders resigned from his position after the occurrence of communal violence (Kimura 2013).  

Consequently, the campaign leaders had to devote considerable time and resources into 



34 
 

addressing internal divisions.  Several Muslim members of the AASU issued an ultimatum to the 

AASU leadership demanding a correction in what they perceived as a pro-Hindu communal tilt 

(Baruah 1999). Similarly, several other indigenous groups, such as the Bodos, began to make 

demands that emphasized their distinctiveness from ethnic Assamese.  In January and December 

1984, the AASU organized two separate national conventions to address the rifts among such 

groups by emphasizing the urgent need to provide a unified political platform (Baruah 1999). 

While the lack of primary sources makes it impossible to pinpoint which of these two causal 

mechanisms (the leaders backing off versus withdrawal of support) is more important, the 

secondary sources clearly demonstrate the role of communal violence in the demobilization 

process.  

 Two significant concessions by the central government in New Delhi led to the end of the 

protest campaign.  First, in 1984, the Election Commission of India decided that the electoral 

rolls in Assam were to be revised based on the 1971 electoral rolls, which meant that the 

immigrants who entered after 1971 would be disenfranchised (The Indian Express, July 13, 

1984).  The campaign leaders saw the Election Commission’s decision as a confidence building 

measure as many illegal immigrants who settled in Assam after the creation of Bangladesh in 

1971 would be disenfranchised. Second, the Indian government left the state government out of 

the negotiations, recognizing the power of the campaign leaders, who insisted that the current 

government was not elected by the people of Assam in 1983 as many ethnic Assamese boycotted 

the elections (Baruah 1999).  Therefore, they argued, the state government was illegitimate and 

should not participate in any discussions (Narayan 1985).  

 In early 1985, the lack of enthusiasm for continued protests was already evident.  Many 

Assamese defied the campaign leaders’ call to boycott the Republic Day celebrations in early 



35 
 

1985 by joining them in the main stadium in Guwahati (The Indian Express, May 8, 1985).  The 

final round of negotiations between the campaign leaders and the central government began in 

April 1985, and the Assam Accord was eventually signed on August 15, 1985.  Accordingly, all 

foreigners who had entered between 1966 and 1971 would be disenfranchised for ten years and 

those who had entered after 1971 would be deported.  The current state government would be 

dissolved and new elections would be held in December 1985 based on the new electoral rolls 

(The Indian Express, August 16, 1985).    

  

Conclusion 

 This empirical analysis of the dynamics of the demobilization of the protest campaign in 

Assam has several implications for the study of nonviolent campaigns.  First, it offers an 

alternative causal mechanism for the demobilization of a protest campaign by demonstrating that 

critical events can trigger demobilization.  Communal violence in ethnically divided societies 

can curtail mass support for the protest campaign and force campaign leaders to suspend protest 

activities.  While scholars have often highlighted the escalatory effects of critical events, this 

analysis shows that they can also be important turning points in the demobilization of a protest 

campaign.  

Second, this study suggests that the participation of large numbers of people and the 

inclusion of various sectors and groups do not always set off mechanisms that lead to success. 

Even if repression triggers a backfire effect and motivates larger number of people to participate, 

campaigns can still demobilize. Indeed, similar to Piven and Cloward’s (1977) and Tarrow’s 

(1989) models, the Assam case shows that broadening the support base makes coalition 

maintenance challenging in ethnically divided societies.  When communal violence against 



36 
 

Muslim immigrants occurred in Assam, Muslim protesters withdrew their support, forcing the 

campaign leadership temporarily to shift their strategy from confronting political authorities to 

trying to unify the different groups within the campaign. 

 Third, the Assam case shows that violent protest activity does not necessarily undermine 

the strength of a protest campaign as scholars of nonviolent resistance maintain.  Although 

protesters resorted to violent tactics, mobilization levels remained high throughout the first four 

years of the campaign.  Instead, a critical event, communal violence, eventually led to the 

demobilization of the protest campaign.  In line with Davenport and Moore’s (2012) call for 

better mapping and understanding of diverse forms of political conflict and violence, this study 

indicates that various types of violence can have differential effects on protest dynamics. 

 From a policy standpoint, the implications of the Assam campaign for states and activists 

are quite alarming.  In ethnically divided societies, protests that highlight ethnic differences can 

have unintended violent consequences and lead to deeper societal cleavages in the long term.  

States should be aware of this potential and calculate the costs of repressing and fueling protests 

by taking into account the possibility of large-scale communal violence.  Likewise, protesters 

should develop strategies to prevent the escalation of ethnic tension as it can have detrimental 

effects on the campaign. 

At a more general level, the Assam case illustrates the unpredictable nature of protest 

dynamics.  Scholars of political contention have long recognized the challenges associated with 

identifying causal pathways due to the unpredictability of contentious interactions between 

multiple actors.  This study explains how an unpredictable event can demobilize a protest 

campaign.  Although this study has limitations in its generalizability across different campaigns 

due to the unique communal patterns and protest dynamics in Assam, it highlights an alternative 



37 
 

causal path to demobilization and, thus, brings scholars one step closer to understanding the 

complex dynamics of contentious politics.  

  



38 
 

References 

 

 

Alimi, Eitan Y. (2009) Mobilizing under the Gun: Theorizing Political Opportunity Structure in 

a Highly Repressive Setting. Mobilization 14(2):219-37. 

Baruah, Sanjib. (1986) Immigration, Ethnic Conflict, and Political Turmoil--Assam, 1979-1985. 

Asian Survey 26(11):1184-206. 

———. (1999) India against Itself : Assam and the Politics of Nationality. Critical Histories. 

Philadelphia: University of Pennsylvania Press. 

Beissinger, Mark R. (2002) Nationalist Mobilization and the Collapse of the Soviet State. 

Cambridge: Cambridge University Press. 

Brandt, Patrick T., and John T. Williams. (2001) A Linear Poisson Autoregressive Model: The 

Poisson Ar(P) Model. Political Analysis 9(2):164-84. 

Brandt, Patrick T., John T. Williams, Benjamin A. Fordham, and Brian Pollins. (2000) Dynamic 

Modeling for Persistent Event Count Time Series. American Journal of Political Science 

44(4):823-43. 

Capoccia, Giovanni, and R. Daniel Kelemen. (2007) The Study of Critical Junctures: Theory, 

Narrative, and Counterfactuals in Historical Institutionalism. World Politics 59(3):341-

69. 

Chenoweth, Erica, and Maria J. Stephan. (2011) Why Civil Resistance Works : The Strategic 

Logic of Nonviolent Conflict. New York: Columbia University Press. 

Collier, Ruth Berins, and David Collier. (2002) Shaping the Political Arena: Critical Junctures, 

the Labor Movement, and Regime Dynamics in Latin America. Notre Dame, Ind. :: 

University of Notre Dame Press. 

Davenport, Christian, and Will H. Moore. (2012) The Arab Spring, Winter, and Back Again? 

(Re)Introducing the Dissent-Repression Nexus with a Twist. International Interactions 

38(5):704-13. 

Day, Joel, Jonathan Pinckney, and Erica Chenoweth. (2015) Collecting Data on Nonviolent 

Action: Lessons Learnd and Ways Forward. Journal of Peace Research 52(1):129-33. 

della Porta, Donatella, and Sidney Tarrow. (1986) Unwanted Children: Political Violence and 

the Cycle of Protest in Italy: 1966-1973. European Journal of Political Research 14. 

DeVotta, Neil. (2004) Blowback: Linguistic Nationalism, Institutional Decay, and Ethnic 

Conflict in Sri Lanka. Stanford: Stanford University Press. 

Dunning, Thad. (2011) Fighting and Voting: Violent Conflict and Electoral Politics. The Journal 

of Conflict Resolution 55(3):327-39. 

Earl, Jennifer, Andrew Martin, John D. McCarthy, and Sarah A. Soule. (2004) The Use of 

Newspaper Data in the Study of Collective Action. Annual Review of Sociology 3065-80. 

Farnum, Nicholas.R., and LaVerne W. Stanton. (1989) Quantitative Forecasting Methods. PWS-

Kent Pub. 

Francisco, Ronald A. (1995) The Relationship between Coercion and Protest: An Empirical 

Evaluation in Three Coercive States. The Journal of Conflict Resolution 39(2):263-82. 

Gandhi, Rajmohan. April 11. 1983, 1983 Assam: Some Questions and Answers. The Indian 

Express, 5. 

George, Alexander L., and Andrew Bennett. (2005) Case Studies and Theory Development in the 

Social Sciences. Bcsia Studies in International Security. Cambridge, Mass.: MIT Press. 



39 
 

Goertz, Gary, and Paul F. Diehl. (1995) The Initiation and Termination of Enduring Rivalries: 

The Impact of Political Shocks. American Journal of Political Science 39(1):30-52. 

Gupta, Shekhar February 24, 1983 Assam Violence Has No Set Pattern. The Indian Express, 7. 

Horowitz, Donald L. (2001) The Deadly Ethnic Riot. Berkeley: University of California Press. 

Kimura, Makiko. (2013) The Nellie Massacre of 1983: Agency of Rioters. New Delhi: Sage 

Publications India. 

Koopmans, Ruud. (1997) Dynamics of Repression and Mobilization: The German Extreme 

Right in the 1990s. Mobilization 2(2):149-65. 

Krain, Matthew. (2000) Repression and Accommodation in Post-Revolutionary States. 1st ed. 

New York: St. Martin's Press. 

Linden, Ariel, and John L. Adams. (2011) Applying a Propensity Score-Based Weighting Model 

to Interrupted Time Series Data: Improving Causal Inference in Programme Evaluation. 

Journal of Evaluation in Clinical Practice 17(6):1231-38. 

Lohmann, Susanne. (1994) The Dynamics of Informational Cascades: The Monday 

Demonstrations in Leipzig, East German, 1989-91. World Politics 47(October):42-101. 

Long, Scott. (1997) Regression Models for Categorical and Limited Dependent Variables. 

Thousand Oaks: Sage Publications. 

Mahoney, James. (2001) The Legacies of Liberalism : Path Dependence and Political Regimes 

in Central America. Baltimore :: Johns Hopkins University Press. 

McAdam, Doug. (1983) Tactical Innovation and the Pace of Insurgency. American Sociological 

Review 48(6):735-54. 

McCarthy, John D., Clark McPhail, and Jackie Smith. (1996) Images of Protest: Dimensions of 

Selection Bias in Media Coverage of Washington Demonstrations, 1982 and 1991. 

American Sociological Review 61(3):478-99. 

Meyer, David S., and Suzanne Staggenborg. (1996) Movements, Countermovements, and the 

Structure of Political Opportunity. American Journal of Sociology 101(6):1628-60. 

Moore, Will H., and Ronny Lindstrom. (1996) The Violent Intranational Conflict Data Project 

(Vicdp) Codebook. Department of Political Science, University of California - Riverside. 

Narayan, Hemendra. February 19, 1985 Show of Strength by Assam Agitation Leaders. Indian 

Express. 

Nepstad, Sharon Erickson. (2011) Nonviolent Revolutions : Civil Resistance in the Late 20th 

Century. New York: Oxford University Press. 

Piven, Frances Fox, and Richard A. Cloward. (1977) Poor People's Movements: Why They 

Suceed, How They Fail. New York: Pantheon Books. 

Polanyi, Karl. (1957) The Great Tansformation. New York :: Rinehart. 

Rasler, Karen. (1996) Concessions, Repression, and Political Protest in the Iranian Revolution. 

American Sociological Review 61(1):132-52. 

Rasler, Karen A., William R. Thompson, and Sumit Ganguly. (2013) How Rivalries End. 1st ed. 

ed. Philadelphia :: University of Pennsylvania Press. 

Schock, Kurt. (2013) The Practice and Study of Civil Resistance. Journal of Peace Research 

50(3):277-90. 

———. (2005) Unarmed Insurrections : People Power Movements in Nondemocracies. 

Minneapolis, MN: University of Minnesota Press. 

Sharp, Gene. (1973) The Politics of Nonviolent Action. Boston, MA.: Porter Sargent. 



40 
 

Shawki, Noha. (2012) The 2008 Food Crisis as a Critical Event for the Food Sovereignty and 

Food Justice Movements. International Journal of Sociology of Agriculture and Food 

19(3):423-44. 

Staggenborg, Suzanne. (1993) Critical Events and the Mobilization of the Pro-Choice 

Movement. Research in Political Sociology 6319-45. 

Svensson, Isak, and Mathilda Lindgren. (2011) From Bombs to Banners? The Decline of Wars 

and the Rise of Unarmed Uprisings in East Asia. Security Dialogue 43(2):219-37. 

Tarrow, Sidney G. (1989) Democracy and Disorder : Protest and Politics in Italy, 1965-1975. 

Oxford: Oxford University Press. 

———. (1998) Power in Movement : Social Movements and Contentious Politics. Cambridge 

Studies in Comparative Politics. 2nd ed. Cambridge, U.K. ; New York: Cambridge 

University Press. 

Tilly, Charles. (1978) From Mobilization to Revolution. Reading, Mass.: Addison-Wesley Pub. 

Co. 

Tilly, Charles, and Sidney G. Tarrow. (2007) Contentious Politics. Boulder, Colorado: Paradigm 

Publishers. 

Varshney, Ashutosh. (2002) Ethnic Conflict and Civic Life : Hindus and Muslims in India. New 

Haven [Conn.]: Yale University Press. 

 

  



41 
 

APPENDIX 

for  

The Dynamics of the Demobilization of the Protest Campaign in Assam 

By Tijen Demirel-Pegg 

 

This appendix includes the following sections:  

Section A: Explanation of the Coding Procedures and the Codebook used in the data collection 

Section B: Descriptive Statistics 

Section C: Robustness Checks 

 

  



42 
 

 

A. EXPLANATION OF CODING PROCEDURES AND THE CODEBOOK 
 

1. Coding Procedures 
 

I followed multiple steps to construct the codebook and weight the data. First, I identified 

the four broad event categories that I needed data for: 1) state accommodation; 2) state 

repression; 3) group collective action; and 4) external support to the group. These categories 

were followed by the lists of actors involved in the Assamese protest campaign. 

For state actions, I relied on a majority of Krain’s (2000) list of repressive and 

accommodative events.26 Krain includes a variety of repressive state actions, ranging from 

seizing assets to use of torture, as well as a comprehensive list of accommodative actions ranging 

from increasing security of target group to incorporation into government.  Even though these 

lists cover a wide range of events, I discovered some additional categories that were essential to 

my four cases throughout the coding process.  Therefore, I added violent arrests of opposition 

leaders; halt negotiations; curfews; declaration of President’s Rule or Governor’s Rule;27 

teargassing; violent disruption of group organization; beating up; clashes with group; damaging 

property; bombing; burning houses and bridges; and burning entire villages to the repressive 

events lists. These events came up with some frequency and provided detailed information about 

the exact form of events.  

I then labeled the repressive events based on their objectives and decision making 

agencies. Preventive repression corresponds to repressive acts imposed by higher level 

authorities, such as the government or the judiciary, to demobilize the contenders and to prevent 

future acts of contention. The declaration of martial law, and the restriction of the right of 

assembly, for instance, fall into this category of repression. Reactionary repression, on the other 

hand, stands for repressive acts of lower level agents, such as the police, applied during the 

course of a contentious episode in an attempt to respond to collective action and maintain or re-

establish order. Teargassing the protesters or damaging property are examples of reactionary 

repression. I consider arrests as a separate category because it is not always clear if arrests result 

from centralized policy directives, or if they are consequences of ad-hoc decisions made by the 

local police. Finally, I grouped all repressive events under six broader categories: Restrict; Seize; 

Warn; Judicial Actions; Non-Judicial Actions; and Use of Force. 

Regarding the state’s accommodative policies, I expanded Krain’s list by adding 

organizing discussion meetings or committees, and withdrawal of army. The Indian state used 

these two types of policies during the protest waves and insurgencies as a form of 

accommodation.  Again, I grouped similar types of events under broader categories: 

Negotiations; Judicial Accommodation; Non-Judicial Accommodation; and Group Recognition. 

For the purpose of systematic coding, I arbitrarily assigned the broad categories a 3-digit code 

and the specific event types a 4-digit code which started with the first two digits of the broader 

category.  All repressive events that involve violence were assigned a 9 at the very end (thus, 

violent actions are all 5-digit codes ending with 9).  All the codes for accommodative events start 

                                                           
26 I did not include mass killings because some form of military armed attack, burning villages or bombing already 
covered this category.   
27 Article 356 of the Constitution of India grants the President the right to dissolve or suspend a state legislature and 
place the state under direct federal rule if he is satisfied that there has been failure of the constitutional machinery in 
the state. 



43 
 

with a 1 and the repressive ones start with a 2. This code assignment was particularly helpful 

later throughout the data management process because it made it easier to identify similar types 

of events and their respective broader categories for aggregation purposes. It also helped me to 

detect miscoding, if any, more easily. The following are examples of categories of 

accommodation and repression with their respective codes.  

For group actions, I used a majority of the Violent International Conflict Data Project 

(VICDP) events list (Moore and Lindstrom 1996).  The VICDP is a data project, which has 

generated conflict and cooperation data at the sub-national level. Events from the Integrated Data 

for Event Analysis Framework (IDEA) (Bond,Bond,Oh,Jenkins and Taylor 2003) was also 

helpful in identifying specific events that were not included in the VICDP project, such as 

hijacking, and taking hostages. For the purpose of distinguishing between situations in which 

violence is or is not used, I added the “violent” versions of demonstrations, strikes and sit-ins to 

the list. I also created two separate categories for demonstrations based on size. Demonstrations 

with less than 200 participants are coded as small scale and the ones with more than 200 are 

coded as large scale. Even though the cut-off point of 200 participants is arbitrary, there is a 

pragmatic reason behind it. The Indian Express usually reports massive demonstrations as 

featuring “hundreds” or “thousands” of demonstrators. I am conservatively interpreting 

“hundreds” as at least 200, and therefore, using it as my cut-off point. 

After completing the list for group actions, I grouped those under five broader categories: 

Accommodative Group Actions; Low Intensity Collective Action; Medium Intensity Collective 

Action; and High Intensity Collective Action.  Again, the broader category was assigned a 3-

digit code and the specific events were assigned a 4-digit code.  All the events that involved 

violence were assigned a 9 at the end. All collective action events start with a 5.  

In addition to collective action I included a “Splits” category for groups to code 

information about group cohesion. The list of events in this category does not reflect a wide 

range of group characteristics because the scope of the study is not on organizational aspects of 

the groups. Instead, it includes the events that are the results of splits within a group. These 

specific events are disagreement between factions; forming an organization as a result of split; 

expelling members; armed attack against other factions; and militants surrendering. These 

events also have codes starting with a 5 because they are related to groups.  

For the list of acts of external support to the group, I built on Heraclides’ (1990:368-69) 

list of events which are specifically designed to fit cases with external involvement in 

secessionist conflicts.  For example, some of the events in his list including providing access to 

communication, providing sanctuary and training, or giving direct military assistance such as 

armed intervention constituted the basis of the list for external support to the group.  I also added 

firing across borders and granting asylum to leadership of group to the list to make it more 

comprehensive.  All the events for external support to the group are arranged under two larger 

categories: Material and Diplomatic/Political Support. Material Support has three sub-categories 

based on the level of involvement: Low, Moderate and High. The 3 digit codes for these main 

categories start with an 8 and the specific events have 5 digit codes, just like the rest of the 

codebook.  

While compiling the list of events for state, group, and external actor actions, I 

deliberately attempted to include similar policy options across the board. Obviously, the same 

exact policy options that are available to the state are not available for the group and vice versa. 

However, if the state has the option of putting down a demonstration, then demonstration is 



44 
 

included among the list of group actions; and so is allowing demonstrations in state’s own 

territory for external support actions. 

Finally, for the list of actors and targets, I used the actor/target list in Moore and 

Lindstrom’s (1996) VICDP as a starting point for my initial deliberations on what kind of actors 

to include. The specific party names and factions, names of governors, chief ministers, names of 

religious or ethnic groups along with broader categories of students, peasants, workers, and 

businessmen, for instance, are listed in the codebook. These specific actors were organized based 

on 7 types of actors and targets: State, Social Actors, Political Parties, Religious Groups / 

Populations, Insurgent Groups, State Enforcement, and External Actors. All the broader 

categories have 5 digit codes and the specific actors and targets have 6 digit codes. If there is 

more detailed information available, then those more specific ones have 7 digit codes. For 

example, in Assam, the code for the State is 33100, for the Governor 331011, and for Governor 

Shri T.S. Misra, it is 3310113.  

Once the entire list of events was compiled, I assigned scores to each event to reflect the 

intensity of these events more realistically. It would simply not be accurate to assume that the 

state’s restriction of assembly has the same impact as launching an armed attack.  If both of these 

events would be assigned a score of 1 (or simply counted by frequency), it might very well be 

possible that restriction of assembly ends up having more weight in the statistical analysis than 

launching armed attacks against the group if its frequency is higher. In order to make the analysis 

more accurate, the weights of those events needs to be taken into consideration.  

For assigning weights, my starting point was, once again, Krain’s (2000) scale of 

repression and accommodation.  Krain’s scale is initially based on the rankings of 30 experts in 

the field. The experts were asked to rank the repressive and accommodative events on a 1 to 5 

scale. The average scores for each event were calculated and then rounded to the nearest whole 

number. To be able to make finer distinctions, Krain took the square of the raw average score 

and assigned new scores to the events ranging from 1 to 25 (Krain 2000:42-43). So, for example, 

while arresting opposition leader and armed attack received a score of 3 by the expert rankings, 

Krain scored them as 7 and 12 respectively.  By expanding the range of the weights, Krain 

gained enough flexibility to rank events which originally fell into the same score category.   

As a wider range of scores allows me to rank the events with more precision and thus to 

see the increases or decreases in the intensity of the actions, I decided to use a range between 1 

and 25 as well.  I first assigned similar scores to the repressive and accommodative events that 

Krain had on his list.  Then I ranked the events which I had added in the previous step that were 

not in Krain’s list. For example, while arresting opposition leaders is on Krain’s original list, I 

included violent arrests of opposition leaders into my list of repressive events. These additions 

inevitably led to the modification of several scores. The addition of a violent event to the list 

meant that it had to have a score more than the non-violent version of it.  So, the score was 

assigned based on other violent events with similar intensity. As a general principle, the violent 

events were assigned a 4 point higher score than the non-violent versions.28 Once this rule was 

established, similar events were modified based on the same principle. Finally, I made some 

adjustments to the scale based on case-specific characteristics. Throughout the data collection 

process, I read thousands of newspaper articles on the same subject, which gave me a good 

understanding of the actions of the state and the groups. I was, therefore, able to understand how 

the state used certain types of policies in what kind of context and used this insight to modify the 

                                                           
28 This 4 point difference is enough to account for the difference in intensity, yet does not place the violent and non-
violent versions of the same event at the extreme ends of the scale.  



45 
 

scale. For example, while declaration of martial law has a score of 10 in Krain’s scale, it is 21 in 

my scale because the Indian state declared martial law in situations where other repressive 

measures had already been taken and martial law was seen as the only option left to halt the 

further deterioration of law and order. Therefore, declaration of martial law was assigned a 

dramatically higher score than violent or non-violent putdown of demonstrations. Even with 

these various modifications, the correlation coefficient for accommodative and repressive events 

between the new scale and Krain’s scale is relatively strong (.92 and .78 respectively). 

I then applied these similar guidelines to assign scores to group actions and external 

support events. The most severe forms of collective action and external support events were 

assigned scores closer to 25 and the least severe were assigned scores closer to 1. Actions 

included as options under both the state actions or group actions received similar scores.  For 

example, armed attack receives a score of 20 both in the list of state’s repressive actions and the 

group’s most threatening actions list. Similarly, negotiating receives the same score of 6 in both 

the state’s and group’s conciliatory actions lists.  

This systematic and consistent procedure of assigning weights was particularly important 

in ensuring some degree of balance and objectivity in this otherwise subjective process. 

Subjectivity is a legitimate concern both with Krain’s scale and also with my scale as, indeed, it 

is with any other weighting scheme.  As there are no objective criteria used to rank such events, 

any ranking inevitably involves some kind of a judgment call on the researcher’s part. However, 

the application of the same principles across all types of actions will reduce the bias to a 

minimum and prevent it from affecting the analysis in a significant way.  
  



46 
 

2.  CODEBOOK 

EVENT LIST FOR ACCOMMODATION 

     

Main Category Specific Category Code  

Adjusted 

Weight  

(W)  

    

NEGOTIATIONS    110   

 Broad statement of guarantee of rights 1111 4 

 Low-level concession  1112 1 

 Compromise reached  1113 15 

 Negotiations begun 1114 6 

    

    

JUDICIAL ACCOMMODATION 120   

 

Reduction in fines, taxes, fees, 

 punishment for compellance 1211 4 

 Removing restrictive law/regulation  1212 5 

 Allowing emigration 1213 7 

 Allowing return from exile  1214 8 

 

Outlawing repressive apparatus institution, 

practice  1215 8 

 Prosecution of enemies of target group  1216 1 

 Stays of execution  1217 5 

 Freeing prisoners 1218 10 

 General Amnesty  1220 14 

 Legalizing Group Membership  1221 12 

    

    

NON-JUDICIAL ACCOMMODATION 130   

 Organizing discussion meetings, committees  1311 3 

 Increasing access to information 1312 4 

 Providing selective incentives 1313 8 

 Giving positive publicity 1314 4 

 

Introducing enabling 

mechanisms/entitlement 1315 6 

 Increasing opportunities for participation 1316 9 

 

Small-scale diversion of 

resources to group 1317 5 

 Abandoning project hurtful to group 1318 4 

 Increasing security of target group 1320 1 

 Increasing access to markets 1321 9 

 

Large-scale redistribution of 

assets/resources 1322 10 

 Direct aid (money, forces) to challenger 1323 12 

 Withdrawal of Army  1324 16 

    



47 
 

GROUP RECOGNITION 140   

 Encouraging/allowing separatism 1411 18 

 Incorporation into government  1412 25 

    

OTHER    910   

        

 

EVENT LIST FOR REPRESSION     

Main Category Specific Category Code  

Adjusted 

Weight  

(W)  

RESTRICT   210   

Preventive Repression 

Restricting access to resources, 

necessities, jobs 2111 6 

 Restricting assembly 2112 5 

 

Restricting freedom of  

communication, speech 

and distribution of information 2113 5 

 Restricting emigration, mobility  2114 6 

 Prohibitive fines, taxes, fees, regulation  2115 4 

    

    

SEIZE   220   

Preventive Repression Seizing assets 2211 7 

Arrest Arrests/ Detentions 2212 9 

 Mass Arrests (>25) 2213 14 

 Non-violent Arrest of Opposition Leaders  2418 10 

 Violent Arrest of Opposition Leaders 24199 14 

    

    

WARN   230   

Preventive Repression Public threat of violence/punishment 2311 5 

 Show of force 2312 6 

 Exemplary punishment/deterrent 2313 9 

 Show trials/political trials 2314 8 

 Compellance: Punishing non-performance 2315 4 

 Expulsion/purge from party or ruling elite 2316 3 

 Making groups more visible, easily spotted 2317 4 

 Halt Negotiations 2318 6 

JUDICIAL ACTIONS    240   

Preventive Repression Curfew  2411 10 

 Martial law declared 2412 21 

 “Special” extra-legal courts set-up 2413 8 

 

Suspension of parts of constitution or  

the regular workings of government / issuing 

ordinance 2414 16 

 Outlawing organizations, groups, industries  2415 15 



48 
 

 

Suspending or censoring news 

media/speech  2416 14 

 Ousting groups from government 2417 15 

 Exiles/expulsions 2420 10 

 Trials in absentia  2421 14 

 Forced conscription 2422 8 

 President's Rule/Governor's Rule  2423 18 

 Religious suppression/persecution 2424 9 

    

NON-JUDICIAL ACTIONS  250   

Reactionary Repression 

Non-violent putdown of 

demonstration/strikes/hartals 2511 7 

 

Violent putdown of 

demonstration/strikes/hartals,  

lathi-charging and teargassing 25119 11 

 Non-violent disruption of group organization 2512 7 

 Violent disruption of group organization 25129 11 

 

Calling in  

additional police posts/ secret police/special 

forces 2514 6 

    

USE OF FORCE    260   

Reactionary Repression Beating up  26119 11 

 Clashes with group  26129 14 

 Use of torture 26139 16 

 Forced resettlement 26149 18 

 Armed attack 26159 20 

 Forced labor 26169 12 

 

Forced subjugation/integration 

of communities  26179 9 

 Disappearance  26189 21 

 Damaging property  26199 14 

 Bombing  26209 21 

 Burning houses, bridges, etc.  26219 21 

 Burning entire villages 26229 25 

 Manufactured famine or draught 26239 20 

 Mass rape 26249 21 

 Concentration Camps 26259 18 

OTHER   920   

Note: All actions that involve violence end with 9.   

EVENT LIST FOR COLLECTIVE ACTION 

      

Main Category Specific Category Code  

 Weight  

(W)  

     

ACCOMMODATIVE ACTIONS 510   

 

Broad statement of guarantee 

of rights and security of government officials 5111 4 



49 
 

 Low-level concession  5112 1 

 Compromise reached  5113 15 

 Negotiations begun 5114 6 

 Freeing hostages/prisoners 5115 10 

  

Abandoning plans/actions/strikes/boycotts / 

demonstrations etc. hurtful to state 5116 4 

 

Allowing government access to 

resources, mines, oil facilities, etc.  5117 9 

 Renouncing secessionist goals 5118 16 

 Acceptance of an Agreement 5120 18 

 Termination of Separatist Movement 5121 20 

 Announcing ceasefire 5122 15 

    

 CONTENTIOUS ACTIONS     

LOW INTENSITY COLLECTIVE ACTION 520   

 

Mild negative statements 

about federal government  5211 1 

 

Mild negative statements 

about local government 5212 1 

 

Strong negative statements 

about federal government 5213 3 

 

Strong negative statements 

about local government 5214 3 

 Organizing meetings  5311 6 

 Distributing information  5312 5 

 Forming an organization 5313 8 

    

MEDIUM INTENSITY COLLECTIVE ACTION 540   

 Non-violent Strike (hartal)  5411 7 

  Violent Strike (hartal)  54119 11 

 Non-Violent Bandh (general strike) 5412 9 

 Violent Bandh (general strike)  54129 11 

 

Non-violent Small-scale demonstrations 

 (less than 200 participants)  5413 7 

 

Violent Small-scale demonstrations 

(less than 200participants) 54139 11 

 

Non-violent Large-scale demonstrations 

 (more than 200 participants)  5414 10 

 

Violent Large-scale demonstrations 

 (more than 200 participants) 54149 11 

 Non-violent Sit-ins  5415 7 

 Violent Sit-ins  54159 11 

 Commercial Boycotts  5416 7 

 Electoral boycotts  5417 10 

 Halt negotiations 5418 6 

 Forming a political wing / party  5420 12 



50 
 

 Running candidate in the election  5421 12 

 Threat of violence/agitation 5422 5 

 Violation of curfew, martial law, etc.  5423 5 

    

HIGH INTENSITY COLLECTIVE ACTION 550   

 

Surveillance, links with spies,  

acquiring information through clandestine means 5424 7 

 Assaults 55109 11 

 Robbery/stealing-damaging property 55119 14 

 Clashes  55129 15 

 Direct confrontation with army or paramilitaries 55299 20 

 Sabotage 55139 17 

 Abduct tourists/journalists/citizens 55149 15 

 Abduct and execute tourists/journalists/citizens  55249 20 

 Abduct politicians 55159 16 

 Abduct and execute politicians 55259 20 

 Seizing  buildings  55169 17 

 Using training camps  55179 16 

 Securing access to arms  55189 16 

 Armed attack 55199 20 

 Bombing  55209 21 

 Riots 55219 17 

 Burning houses, bridges, etc.  55229 21 

 Burning entire villages 55239 25 

    

OTHER   950   

    

All actions that involve violence en with 9.   

    

    

SPLITS       

Main Category Specific Category Code  

 Weight  

(W)  

 Disagreements between factions 5611 5 

 Forming an organization as a result of splits  5612 10 

 Expelling members 5614 7 

 Armed attack against other factions 56139 20 

 Militants surrendering 5615 10 

    

EVENT LIST FOR EXTERNAL SUPPORT TO THE STATE 

      

Main Category Specific Category Code  

 Weight  

(W)  

MATERIAL SUPPORT 810   

LOW Humanitarian Aid 8111 2 



51 
 

MEDIUM   820   

 Providing Financial Assistance 8211 10 

 

Providing Vital Access to 

communication 8212 10 

 

Threatening to use force against group 

on state's behalf 8213 5 

HIGH    830   

 Non-injury destructive action against group 8311 13 

 Providing Arms 8312 15 

 Providing Military Personnel  8313 17 

 Border seal 8314 15 

 Border skirmishes (support state) 8315 17 

 Full scale military intervention support 8316 21 

DIPLOMATIC / POLITICAL SUPPORT 840   

 Mild Verbal Statements 8411 1 

 Strong Verbal Statements 8412 3 

 

Support in Intergovernmental  

Organizations 8413 10 

 Campaigns in Support of State 8414 10 

 

Banning groups to exist/function 

in the state's own territory / expulsion of group 8415 8 

 

Banning demonstrations/protests, etc. in the 

state's own territory 8416 8 

 Detaining or arresting group members 8417 10 

OTHER   981   

 

EVENT LIST FOR EXTERNAL SUPPORT TO THE GROUP 

      

Main Category Specific Category Code  

Weight  

(W)  

TANGIBLE SUPPORT 850   

LOW Humanitarian Aid 8511 2 

    

    

MODERATE  860   

 Providing Sanctuary/training 8611 14 

 Providing Base of Operation 8612 14 

 Providing Financial Assistance 8613 10 

 

Providing Vital Access to 

communication 8614 10 

 

Threatening the state to intervene 

on group's behalf 8615 5 

       

HIGH - PHYSICAL 

INVOLVEMENT   870   

 Non-injury destructive action against state 8711 13 

 Providing Arms 8712 15 



52 
 

 

Providing Military Personnel under secessionist 

command 8713 17 

 Border closures 8714 15 

 Border skirmishes (support group) 8715 17 

 Firing Across Borders 8716 16 

 Full scale military intervention 8717 21 

    

    

DIPLOMATIC / POLITICAL SUPPORT 880   

 Mild Verbal Statements 8811 1 

 Strong Verbal Statements 8812 3 

 

Support in Intergovernmental  

Organizations 8813 10 

 Campaigns in Support of Groups 8814 10 

 

Allowing groups to exist/function 

in the state's own territory 8815 8 

 

Allowing demonstrations/protests, etc. in the 

state's own territory 8816 8 

 

Freeing detained/arrested group 

members out of prison 8817 10 

 Granting Asylum to leadership of group 8818 14 

    

OTHER   985   

 

 

   

ACTORS /TARGETS FOR ASSAM    

    

STATE   33100  

 Federal Government 331001  

 Local Government 331002  

 Governor 331011  

 Shri L.P. Singh, ICS 3310111  

 Shri P. Mehrotra 3310112  

 Shri T.S. Misra 3310113  

 Shri B. N. Sing 3310114  

 Shri Harideo Joshi 3310115  

 

Shri Justice Anisetti Roghuvir, Chief 

Justice, Assam 3310116  

 Shri D. D. Thakur 3310117  

 Shri Lok Nath Misra 3310118  

 Chief Minister 331003  

 CM Shri Golap Borbora 3310031  

 CM Shri Jogenda Nath Hazarika 3310032  

 CM Shrimati Anowara Taimur 3310033  

 CM Shri Kesab Chandra Gogoi 3310034  

 CM Shri Hiteswar Saikia 3310035  

 CM Shri Prafulla Kumar Mahanta 3310036  

    



53 
 

SOCIAL ACTORS   33110  

 Media 331101  

 Students 331102  

 Workers 331103  

 Peasants 331104  

 Educators 331105  

 Politicians 331106  

 Political Activists 331107  

 Elites 331108  

 Local Tourists 331109  

 Foreign Tourists 331110  

 Foreign Journalists 331111  

 Foreign Politicians 331112  

 Bengali nationals trying to immigrate 331113  

 Civil servants 331114  

 Nepali nationals trying to immigrate 331115  

 Businessmen/traders 331116  

 Foreign businessmen/traders 331117  

 Other 33910  

    

    

POLITICAL PARTIES 33120  

 Congress Party  331201  

 Janata Party 331202  

 BJP  331203  

 Asom Gana Parishad (AGP) 331204  

 Asom Gana Parishad (Splinter) 3312045  

 United Minorities Front  331205  

 Other 33920  

    

RELIGIOUS GROUPS 33130  

 Muslims  331301  

 Hindus 331302  

 Sikhs  333103  

 Buddhists 331304  

 Christian 331305  

 Other 33930  

    

ETHNIC GROUPS   33140  

 Population 331401  

 Assamese  331402  

 Bangladeshi immigrants 331403  

 Tribals 331404  

 Nepalese 331405  

 Bodos 331406  

 Hindi-speaking people 331407  

 Other  33940  

     



54 
 

AGITATING/INSURGENT GROUPS 33150  

 United Liberation Front of Assam 331501  

 All Assamese Student Union (AASU) 331502  

 

All Assam Gana Sangram Parishad 

(AAGSP) 331503  

 Asom Sahitya Sabha  331504  

 

All Assam Minority Students Union (not 

insurgent) 331505  

 Lok Parishad 331506  

 

Assam Jatiyatabadi Yuva Chhatra Parishad 

(AJYCP) - Assam Nationalistic Youth and 

Students Forum) 331507  

 Other 33950  

    

    

STATE ENFORCEMENT 33160  

 Police  331601  

 Paramilitary troops  331602  

 Army 331603  

 Border Security Force  331604  

 Other 33960  

    

EXTERNAL ACTORS 33170  

 US 331701  

 United Kingdom 331702  

 United Nations 331703  

 Regional Organizations 331704  

 China 331705  

 Bhutan 331706  

 Bangladesh 331707  

 Nepal 331708  

 Burma (Myanmar) 331709  

 Pakistan 331710  

 Other 33970  

    

       

 

  



55 
 

B. DESCRIPTIVE STATISTICS 

TABLE B-1 Descriptive Statistics for the Main Variables (Weighted) 

Variable Observation Mean 

Standard 

Deviation Minimum Maximum 

Protest Activity 345 45.81 131.57 0 1278 

Repression 345 38.95 110.18 0 866 

Accommodation 345 5.60 13.19 0 96 

Heavy Violence 345 32.56 80.22 0 780 

 

  



56 
 

C. ROBUSTNESS CHECKS 

 

TABLE C-1 ZINB and PAR Estimates for Number of Protest Activity in Assam (Unweighted 

Data) 

 

Variable 

 

Critical Event Model 

 

Protest Violence 

Model 

 

Pre-Communal 

Violence Model 

 

Post-Communal 

Violence Model 

 ZINB PAR ZINB PAR ZINB PAR ZINB PAR 

Time   -0.003 

  (0.002) 

  0.003* 

  (0.001) 

  -0.007** 

 (0.002) 

 -0.001 

 (0.002) 

 

     - 

 

       - 

 

      - 

 

      - 

Intervention 

(Riot) 

-1.953** 

  (0.704) 

-1.266* 

 (0.532) 

 -1.304* 

 (0.634) 

 -0.761 

 (0.633) 

 

     - 

 

       - 

 

      - 

 

      - 

Post-Intervention   -0.006 

  (0.006) 

  -  

0.024*** 

(0.005) 

  0.009 

 (0.008) 

 -0.015* 

 (0.006) 

 

     - 

 

       - 

 

      - 

 

      - 

Repression  

 

   0.051*** 

  (0.010) 

       

0.044*** 

(0.004) 

  0.032*** 

  (0.008) 

  0.039*** 

 (0.004) 

0.041***

(0.009) 

0 .040*** 

(0.004) 

  0.075† 

 (0.043) 

 0.226*** 

(0.036) 

Accommodation   -0.077 

  (0.050) 

-0.085* 

(0.037) 

  -0.049 

  (0.053) 

 -0.068* 

 (0.039) 

-0.040 

(0.050) 

 0.072* 

(0.036)
 

 -0.192 

 (0.138) 

-0.064 

(0.210) 

High violence 

 

 

- 

 

- 

   0.053* 

  (0.023) 

  0.005 

(0.009) 

 

     - 

 

      - 

 

      - 

 

     - 

Constant   2.296*** 

  (0.310) 

  

1.172*** 

(0.208) 

  2.736*** 

 (0.309) 

1.414*** 

(0.203) 

2.059*** 

(0.171) 

1.577*** 

(0.113) 

1.190*** 

(0.320) 

 -0.018 

 (0.224) 

AR parameter 1 

(lag one) 

 

      - 

    

0.567*** 

(0.056) 

     

- 

 0.520*** 

(0.054) 

 

     -  

 0.658*** 

(0.069) 

   

      - 

  0.437*** 

 (0.089) 

AR parameter 2 

(lag two) 

 

      - 

-0.102* 

(0.047) 

  

-     
-0.066 

(0.048) 

      

     - 

-0.138* 

(0.057) 

 

      - 

  0.004 

 (0.087) 

N   345     343 345   343   215 213    125 123 

Probability>chi-

square 

  

0.000 

 

- 

 

0.000 

 

- 

 

  0.000 

 

       - 

 

   0.029 

 

- 

Adjusted R2 -    0.59 -  0.54      -     0.62       - 0.33 
† p<.10, *p< .05, **p<.01, ***p<.001; All significant tests are two-tailed, standard errors in parentheses 

 

 

  



57 
 

 

TABLE C-2 ZINB and PAR Estimates for Violent and Nonviolent Protest Activity in Assam 

 

Variable 

 

Nonviolent Protest Activity 

 

Violent Protest Activity 

 ZINB PAR ZINB PAR 

Time -0.006*** 

        (0.002) 

-0.001 

(0.001) 

  0.012** 

         (0.004) 

    0.022** 

(0.007) 

Intervention 

(Riot) 

        -1.259* 

        (0.532) 

         -1.788* 

         (0.726) 

2.743* 

(1.357) 

           3.215*** 

          (0.529) 

Post-Intervention         -0.006 

        (0.006) 

         -0.014* 

         (0.006) 

          -0.037* 

(0.018) 

-0.090* 

          (0.036) 

Repression  

 

 0.003*** 

        (0.001) 

   0.004*** 

         (0.000) 

   0.008** 

(0.003) 

          -0.000 

          (0.000) 

Accommodation         -0.017* 

        (0.008) 

-0.025** 

         (0.007) 

          -0.047 

(0.030) 

 0.030* 

          (0.012) 

Constant   4.887*** 

        (0.258) 

  3.580*** 

        (0.195) 

    2.611** 

          (0.797) 

           0.923 

         (1.389) 

AR parameter 1 (lag 

one) 

 

- 

 0.492*** 

        (0.055) 

 

- 
0.088† 

          (0.046) 

AR parameter 2 (lag 

two) 

 

- 

        -0.048 

        (0.049) 

 

-  
    0.087 

          (0.059) 

N 

 

  345             343 345 343 

Probability>chi-square  0.000               -             0.000                 - 

 

Adjusted R2 
 

- 

     

           0.59 

 

- 

 

0.32 
† p<.10, *p< .05, **p<.01, ***p<.001; All significant tests are two-tailed, standard errors in parentheses 

 

 

 

 

 

 

 

 

 

 

 



58 
 

TABLE C-3 Critical Event Model and Cooptation Model Estimates of Protest Activity (January 

1979 – June 1983) 

 
 

Variable 

 

Critical Event 

Model 

 

Cooptation 

Model 

 

Time 

       

    0.002 

   (0.002) 

     

    0.004† 

   (0.002) 

 

Intervention 

(Communal Violence) 

  

   -2.777*** 

   (0.733) 

 

- 

 

Post-Intervention 

 

   -0.163* 

   (0.073) 

 

- 

 

Repression  

 

  

    0.004*** 

   (0.001) 

 

- 

 

Accommodation 

 

   -0.016 

   (0.012) 

  

    0.004 

   (0.013) 

 

Constant 

  

    4.089 

   (0.350)*** 

  

    4.265*** 

   (0.381) 

 

N 

  

       232 

 

      232 

 

Probability>chi-square 

  

    0.000 

 

 

    0.2083 

† p<.10, *p<0.05, **p<.01, ***p<.001; All significant tests are two-tailed, standard errors in parentheses 

 

 

  



59 
 

 

References: 

 

Alimi, Eitan Y. (2009) Mobilizing under the Gun: Theorizing Political Opportunity Structure in 

a Highly Repressive Setting. Mobilization 14(2):219-37. 

Baruah, Sanjib. (1986) Immigration, Ethnic Conflict, and Political Turmoil--Assam, 1979-1985. 

Asian Survey 26(11):1184-206. 

———. (1999) India against Itself : Assam and the Politics of Nationality. Critical Histories. 

Philadelphia: University of Pennsylvania Press. 

Beissinger, Mark R. (2002) Nationalist Mobilization and the Collapse of the Soviet State. 

Cambridge: Cambridge University Press. 

Bond, Doug, Joe Bond, Churl Oh, J. Craig Jenkins, and Charles Lewis Taylor. (2003) Integrated 

Data for Events Analysis (Idea): An Event Typology for Automated Events Data 

Development. Journal of Peace Research 40(6):733-45. 

Brandt, Patrick T., and John T. Williams. (2001) A Linear Poisson Autoregressive Model: The 

Poisson Ar(P) Model. Political Analysis 9(2):164-84. 

Brandt, Patrick T., John T. Williams, Benjamin A. Fordham, and Brian Pollins. (2000) Dynamic 

Modeling for Persistent Event Count Time Series. American Journal of Political Science 

44(4):823-43. 

Capoccia, Giovanni, and R. Daniel Kelemen. (2007) The Study of Critical Junctures: Theory, 

Narrative, and Counterfactuals in Historical Institutionalism. World Politics 59(3):341-

69. 

Chenoweth, Erica, and Maria J. Stephan. (2011) Why Civil Resistance Works : The Strategic 

Logic of Nonviolent Conflict. New York: Columbia University Press. 

Collier, Ruth Berins, and David Collier. (2002) Shaping the Political Arena: Critical Junctures, 

the Labor Movement, and Regime Dynamics in Latin America. Notre Dame, Ind. :: 

University of Notre Dame Press. 

Davenport, Christian, and Will H. Moore. (2012) The Arab Spring, Winter, and Back Again? 

(Re)Introducing the Dissent-Repression Nexus with a Twist. International Interactions 

38(5):704-13. 

Day, Joel, Jonathan Pinckney, and Erica Chenoweth. (2015) Collecting Data on Nonviolent 

Action: Lessons Learnd and Ways Forward. Journal of Peace Research 52(1):129-33. 

della Porta, Donatella, and Sidney Tarrow. (1986) Unwanted Children: Political Violence and 

the Cycle of Protest in Italy: 1966-1973. European Journal of Political Research 14. 

DeVotta, Neil. (2004) Blowback: Linguistic Nationalism, Institutional Decay, and Ethnic 

Conflict in Sri Lanka. Stanford: Stanford University Press. 

Dunning, Thad. (2011) Fighting and Voting: Violent Conflict and Electoral Politics. The Journal 

of Conflict Resolution 55(3):327-39. 

Earl, Jennifer, Andrew Martin, John D. McCarthy, and Sarah A. Soule. (2004) The Use of 

Newspaper Data in the Study of Collective Action. Annual Review of Sociology 3065-80. 

Farnum, Nicholas.R., and LaVerne W. Stanton. (1989) Quantitative Forecasting Methods. PWS-

Kent Pub. 

Francisco, Ronald A. (1995) The Relationship between Coercion and Protest: An Empirical 

Evaluation in Three Coercive States. The Journal of Conflict Resolution 39(2):263-82. 



60 
 

Gandhi, Rajmohan. April 11. 1983, 1983 Assam: Some Questions and Answers. The Indian 

Express, 5. 

George, Alexander L., and Andrew Bennett. (2005) Case Studies and Theory Development in the 

Social Sciences. Bcsia Studies in International Security. Cambridge, Mass.: MIT Press. 

Goertz, Gary, and Paul F. Diehl. (1995) The Initiation and Termination of Enduring Rivalries: 

The Impact of Political Shocks. American Journal of Political Science 39(1):30-52. 

Gupta, Shekhar February 24, 1983 Assam Violence Has No Set Pattern. The Indian Express, 7. 

Heraclides, Alexis. (1990) Secessionist Minorities and External Involvement. International 

Organization 44(3):341-78. 

Horowitz, Donald L. (2001) The Deadly Ethnic Riot. Berkeley: University of California Press. 

Kimura, Makiko. (2013) The Nellie Massacre of 1983: Agency of Rioters. New Delhi: Sage 

Publications India. 

Koopmans, Ruud. (1997) Dynamics of Repression and Mobilization: The German Extreme 

Right in the 1990s. Mobilization 2(2):149-65. 

Krain, Matthew. (2000) Repression and Accommodation in Post-Revolutionary States. 1st ed. 

New York: St. Martin's Press. 

Linden, Ariel, and John L. Adams. (2011) Applying a Propensity Score-Based Weighting Model 

to Interrupted Time Series Data: Improving Causal Inference in Programme Evaluation. 

Journal of Evaluation in Clinical Practice 17(6):1231-38. 

Lohmann, Susanne. (1994) The Dynamics of Informational Cascades: The Monday 

Demonstrations in Leipzig, East German, 1989-91. World Politics 47(October):42-101. 

Long, Scott. (1997) Regression Models for Categorical and Limited Dependent Variables. 

Thousand Oaks: Sage Publications. 

Mahoney, James. (2001) The Legacies of Liberalism : Path Dependence and Political Regimes 

in Central America. Baltimore :: Johns Hopkins University Press. 

McAdam, Doug. (1983) Tactical Innovation and the Pace of Insurgency. American Sociological 

Review 48(6):735-54. 

McCarthy, John D., Clark McPhail, and Jackie Smith. (1996) Images of Protest: Dimensions of 

Selection Bias in Media Coverage of Washington Demonstrations, 1982 and 1991. 

American Sociological Review 61(3):478-99. 

Meyer, David S., and Suzanne Staggenborg. (1996) Movements, Countermovements, and the 

Structure of Political Opportunity. American Journal of Sociology 101(6):1628-60. 

Moore, Will H., and Ronny Lindstrom. (1996) The Violent Intranational Conflict Data Project 

(Vicdp) Codebook. Department of Political Science, University of California - Riverside. 

Narayan, Hemendra. February 19, 1985 Show of Strength by Assam Agitation Leaders. Indian 

Express. 

Nepstad, Sharon Erickson. (2011) Nonviolent Revolutions : Civil Resistance in the Late 20th 

Century. New York: Oxford University Press. 

Piven, Frances Fox, and Richard A. Cloward. (1977) Poor People's Movements: Why They 

Suceed, How They Fail. New York: Pantheon Books. 

Polanyi, Karl. (1957) The Great Tansformation. New York :: Rinehart. 

Rasler, Karen. (1996) Concessions, Repression, and Political Protest in the Iranian Revolution. 

American Sociological Review 61(1):132-52. 

Rasler, Karen A., William R. Thompson, and Sumit Ganguly. (2013) How Rivalries End. 1st ed. 

ed. Philadelphia :: University of Pennsylvania Press. 



61 
 

Schock, Kurt. (2013) The Practice and Study of Civil Resistance. Journal of Peace Research 

50(3):277-90. 

———. (2005) Unarmed Insurrections : People Power Movements in Nondemocracies. 

Minneapolis, MN: University of Minnesota Press. 

Sharp, Gene. (1973) The Politics of Nonviolent Action. Boston, MA.: Porter Sargent. 

Shawki, Noha. (2012) The 2008 Food Crisis as a Critical Event for the Food Sovereignty and 

Food Justice Movements. International Journal of Sociology of Agriculture and Food 

19(3):423-44. 

Staggenborg, Suzanne. (1993) Critical Events and the Mobilization of the Pro-Choice 

Movement. Research in Political Sociology 6319-45. 

Svensson, Isak, and Mathilda Lindgren. (2011) From Bombs to Banners? The Decline of Wars 

and the Rise of Unarmed Uprisings in East Asia. Security Dialogue 43(2):219-37. 

Tarrow, Sidney G. (1989) Democracy and Disorder : Protest and Politics in Italy, 1965-1975. 

Oxford: Oxford University Press. 

———. (1998) Power in Movement : Social Movements and Contentious Politics. Cambridge 

Studies in Comparative Politics. 2nd ed. Cambridge, U.K. ; New York: Cambridge 

University Press. 

Tilly, Charles. (1978) From Mobilization to Revolution. Reading, Mass.: Addison-Wesley Pub. 

Co. 

Tilly, Charles, and Sidney G. Tarrow. (2007) Contentious Politics. Boulder, Colorado: Paradigm 

Publishers. 

Varshney, Ashutosh. (2002) Ethnic Conflict and Civic Life : Hindus and Muslims in India. New 

Haven [Conn.]: Yale University Press.