Note: To view the code click on the ‘Show’/‘Hide’ buttons on the right side of the page.

# Load required libraries
library(lmerTest)
library(dplyr)
library(tidyr)
library(gt)
library(ggplot2)
library(gridExtra)
library(stargazer)
library(patchwork)
library(sjPlot)
library(mediation)

source("af_utils.R")
source("af_graph.R")
source("af_mediation.R")

df <- readRDS("Israel Survey/data/il_pe.RDS")

# # Filter out rows with NA, NULL, or empty strings in any of the must have columns
# must_have_list <- c("gender", "age", "marital_status", "education", "religiosity", "religion")
# df <- df %>%
#   filter(if_all(all_of(must_have_list), ~ !is.na(.) & . != "" & . != "NULL"))

1 General

We measured how distinct destabilizing events (security and political) affect the different dimensions of extremism. The analysis utilized our Extremism Gauge.The results provide compelling empirical evidence supporting our innovative theoretical framework for understanding socio-political extremism. Using our novel conceptualization and operationalization, we demonstrate how different destabilizing events distinctly affect the three dimensions of extremism across diverse Israeli population groups. These preliminary analyses not only validate our multidimensional extremism measure but also reveal its unique capacity to capture nuanced variations across cognitive, behavioral, and social dimensions of extremism — a capability absent in existing unidimensional methods.

Research Question:

  • How do destabilizing events affect the socio-political extremism within the Israeli population?

Hypotheses:

  • Political events affect the level of the population’s socio-political extremism
  • Different political events affect socio-political extremism differently
  • Different extremism dimensions are affected differently
  • Different population groups are affected differently by a specific political event

2 Dataset

Our analysis draws from a comprehensive six-wave study conducted among Jewish Israelis between 2021 and 2024. Waves 1, 2, 5, and 6 each comprised original nationally representative samples of Israeli Jews, while waves 3-4 constituted a panel study. Study 1 was conducted between December 6, 2021, and January 13, 2022 (N = 2016; Israeli Jews = 1609, Arab citizens of Israel = 407). Study 2 took place between May 23, 2022, and June 24, 2022 (N = 2010; Israeli Jews = 1610, Arab citizens of Israel = 400). Study 3 was a panel survey with data collected between October 24, 2022, and October 31, 2022 (pre-elections; N = 1381; Israeli Jews = 963, Arab citizens of Israel = 418) and between January 12, 2023, and January 25, 2023 (post-elections; N = 987; Israeli Jews = 764, Arab citizens of Israel = 223). Study 5 was conducted between June 20, 2023, and June 26, 2023 (N = 1962; Israeli Jews = 1651, Arab citizens of Israel = 311). Study 6 was conducted between April 4, 2024, and April 30, 2024 (N = 1500; Israeli Jews = 1200, Arab citizens of Israel = 300).

3 Destabilizing Events

A series of significant events unfolded in Israel, marked by escalating tensions and political upheaval. This robust research design allowed us to examine the impact of these events on socio-political extremism of the Jewish population.

Terror attacks in major Israeli cities (Waves 1-2, 12/21-05/22)

A wave of terror attacks targeting major Israeli cities, including Jerusalem, Beer-Sheba, Benei-Brak, and Hadera.

Dissolution of the “unity” government (Waves 2-3, 05/22-10/22)

The fall of the “unity” (Bennet-Lapid) government, created a political shift, favorable for Netanyahu’s Likud party.

Introduction of the judicial reform by the new Netanyahu’s government (Waves 3-4, 10/22-01/23)

Shortly after winning the election, Netanyahu’s new government introduced a controversial judicial reform proposed by Yariv Levin. The reform was widely criticized for potentially undermining democratic institutions.

Dismissal of the minister of defense (Waves 4-5, 01/23-06/23)

Netanyahu dismissal of the defense minister Yoav Galant, triggered an immediate massive public response. Within hours, hundreds of thousands of citizens protested the decision, compelling Netanyahu to retract his action.

October 7th War (Waves 5-6, 06/23-04/24)

The most devastating event occurred on October 7th, when a large-scale terrorist attack from Gaza resulted in terrorists breaching border areas. The attack was characterized by extreme violence, with terrorists killing, sexually assaulting, and kidnapping numerous innocent civilians, marking a tragic and traumatic moment in Israel’s history.

# Create the data frame
event_data <- data.frame(
  Waves = c("1-2", "2-3", "3-4", "4-5", "5-6"),
  Threat_Type = c("Security", "Political", "Political", "Political", "Security"),
  Intensity = c("High", "High", "High", "High", "High"),
  Time_Difference_Months = c(2, 4, 0.5, 3, 6),
  Comments = c("Operation Break the Wave ('Shover Galim')",  
               "Approval of the early dissolution of the 24th Knesset", 
               "Justice Minister Yariv Levin presents the judicial reform", 
               "Gallant's dismissal",  
               "The October 7th War")
)

# Fix column names for `cols_label()`
event_data %>%
  gt() %>%
  cols_label(
    Waves = "Waves",
    Threat_Type = "Threat Type",
    Intensity = "Intensity",
    Time_Difference_Months = "Time Distance (Months)",
    Comments = "Comments"
  ) %>%
  tab_header(
    title = "Overview of Events and Threat Types"
  )
Overview of Events and Threat Types
Waves Threat Type Intensity Time Distance (Months) Comments
1-2 Security High 2.0 Operation Break the Wave ('Shover Galim')
2-3 Political High 4.0 Approval of the early dissolution of the 24th Knesset
3-4 Political High 0.5 Justice Minister Yariv Levin presents the judicial reform
4-5 Political High 3.0 Gallant's dismissal
5-6 Security High 6.0 The October 7th War

4 Method

The conceptualization of socio-political extremism centers around three main dimensions: Cognitive, behavioral, and social. The cognitive dimension is constructed from people’s ideological attitudes and political positions. The behavioral dimension is built from people’s endorsement and support of violent protest against the government. The social dimension is constructed from people’s support of personal political exclusion towards various outgroups based on a least-liked paradigm. We measured these three dimensions in all waves.

We evaluated the changes in extremism levels between each pair of consecutive waves. We assumed that political events are primarily responsible for these changes. Considering the short time between successive waves, usually less than 6 months, we expected that the events’ effect will be more substantial than other state or population-related variables.

We measured differences in extremism through three main groups of indicators: Extremism Dimensions, Extremism Point (threshold), and Extremism Ranking.

The indicators of Extremism Dimensions represent the mean extremism level of the population in each dimension. We use a combined indicator calculated based on the squared root of the mean of each extremism dimension indicator. The Extremism Point represents the level in each dimension used to identify respondents’ measures as extreme. We set this threshold based on the median + MAD of the levels in each dimension. Extremism Ranking represents the percentage of respondents with 1/2/3 dimensions beyond the extremism point.

We used paired t-test to analyze the difference in extremism levels of panel survey pairs. We used ANOVA to analyze the difference in extremism levels of cross-sectional survey pairs.

# Respondent ID var name
id_var_name <- "respondent_id"

# List of survey wave pairs
wave_pairs <- list(c(1,2), c(2,3), "panel" = c(3,4), c(4,5), c(5,6))

# Respondents Demographics
demographic_vars <- c("gender", "age", "marital_status", "education")

# Variable labels
var_labels <- c("Overall" = "pe_extremism", 
                "Ideology" = "pe_ideology",
                "Violence" = "pe_violence", 
                "Intolerance" = "pe_intolerance",
                "Ideology" = "ideology_threshold_change",
                "Violence" = "violence_threshold_change", 
                "Intolerance" = "intolerance_threshold_change",
                "ER1" = "er1_diff", 
                "ER2" = "er2_diff", 
                "ER3" = "er3_diff" )

event_types <- c("Type", event_data[["Threat_Type"]])
event_descriptions <- c("Description", event_data[["Comments"]])
events_df <- as.data.frame(rbind(event_types, event_descriptions))

5 Results Summary

The results reveal significant and statistically meaningful shifts in extremism levels across various population segments following the destabilizing political and security events.

The first terror event triggered a notable rise in intolerance, increasing by 4.7% overall. This surge was particularly pronounced among the right-wing population, with intolerance and violence rising by 7.7% and 6.8%, respectively. Interestingly, ideological extremism also rose within left-wing (10%) and center-wing (9.7%) groups.

The political transition with the fall of the “unity” (Lapid-Bennet) government demonstrated substantial extremism changes. The combined extremism indicator dropped by 4.1%, with violence decreasing by 7.5% and intolerance by 4.2%. This reduction was most significant within right-wing populations, national-religious, and Orthodox groups. A noteworthy 14% decrease in support for violence was observed among left-wing populations, a phenomenon that merits further scholarly investigation.

The Judicial Reform announcement produced distinct shifts: right-wing ideological extremism decreased by 4.2%, left-wing ideological extremism increased by 6.7%, and center-wing violence extremism rose by 9.7%, suggesting the followed aggressive political protests.

The public demonstrations that prevented Netanyahu’s dismissal of Defense Minister Galant produced nuanced effects. The secular population saw a 7.4% decrease in combined extremism, while the National Religious population experienced a substantial 15.8% increase. The Orthodox Jewish population witnessed a 19.8% rise in ideological extremism.

The October 7th war had the most profound impact on socio-political extremism. The overall Jewish population experienced increases across multiple extremism dimensions: 7.2% in combined extremism, 4.3% in ideology, 9% in violence, and 8.1% in intolerance. These changes were consistent across political and religious groups.

Extremism Rank 3 is the percentage of population that exceeds all extremism thresholds, thus representing the size of the most extreme group within the population. The research found that the right-wing population and religious groups (Religious, National Religious, and Orthodox) demonstrated the highest sensitivity to political events, having more substantial extremism fluctuations than secular populations. This finding is also revealed through these fgroups’ overall extremism changes (-18% to 16% ), showing substantially higher fluctuations than secular and traditional populations (-7% to 12%, -7% to 4%, respectively).

6 Conclusions

These preliminary findings powerfully demonstrate that different events have substantially different impacts on (i) overall extremism, (ii) different extremism dimensions, and (iii) different population groups. This empirical evidence strongly validates our novel conceptualization and operationalization of socio-political extremism, while underscoring the critical importance of conducting systematic research into how political events shape socio-political extremism. Our findings reveal complex patterns of extremism that cannot be captured by existing unidimensional measures, highlighting the transformative potential of our approach for understanding and addressing socio-political extremism. Building on these preliminary results, we hypothesize that perceived threat serves as a critical mediating mechanism explaining the differential effects of various events on extremism dimensions across population groups. We propose that threat perception varies systematically across different populations as a function of event type, thereby producing distinct patterns of extremist responses across cognitive, behavioral, and social dimensions - a theoretical framework we aim to rigorously test in the proposed research.

7 Detailed Results

# Dependent Variables
dimension_vars <- c("pe_extremism", 
                    "pe_ideology", 
                    "pe_violence", 
                    "pe_intolerance")

threshold_vars <- c("ideology" = "pe_ideology_point", 
                    "violence" = "pe_violence_point", 
                    "intolerance" = "pe_intolerance_point")

extremism_rank_var_name <- "pe_extremism_rank"

7.1 Overall Jewish Population

# Table title
title <- "**Destablizing Events and Socio-Political Extremism**"
subtitle <- "***Overall Jewish Population***"

result <- af_pair_compare (df, 
                           pair_var = "nwave", pairs_list = wave_pairs, 
                           dep_vars = dimension_vars, 
                           rand_vars = demographic_vars, id_var = id_var_name)

tbl_summary <- af_summary_pair_analysis_table(df, result$summary_results,
                                              dimension_vars, threshold_vars,
                                              extremism_rank_var_name,
                                              events_df, var_labels,
                                              tbl_title = title, tbl_subtitle = subtitle)

tbl_summary
Destablizing Events and Socio-Political Extremism
Overall Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 2.6% -4.1% ** 1.1% -1.6% 7.2% ***
Ideology 2.3% -1.5% -0.7% 1.2% 4.3% *
Violence 2.3% -7.5% *** 1.1% -0.1% 9% ***
Intolerance 4.7% ** -4.2% * 1.3% -2.9% 8.1% ***
% Change in Extremsim Point
Ideology -14.19% 0% 2.69% -2.62% 0%
Violence -7.74% -8.47% -2.06% 8.46% 5%
Intolerance 3.96% -4.61% 4.06% -1.7% 5.69%
Change in Extremism Ranking
ER1 9.39% 2.42% -9.97% 6.44% 3.52%
ER2 4.91% -0.6% -1.81% 0.56% 2.11%
ER3 3.67% -0.98% -2.69% 2.43% 0.27%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism point represent the threshold (Median+MAD) for identifying a respondent measure as extreme. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.2 By Political Orientation

af_plot_histogram(df, "pe_left_center_right")

group_var <- "pe_left_center_right"
group_list <- c("right", "center", "left")

# Table title
title <- "**Political Events, Political Orientation and Socio-Political Extremism**"
subtitle <- "***<group>-wing Jewish Population***"

outputs <- 
  af_pair_compare_by_group(df, 
                           group_var, group_list, 
                           threshold_vars, dimension_vars,
                           extremism_rank_var_name,
                           events_df, var_labels, 
                           title, subtitle)

7.2.1 Right-wing

outputs[["right"]]
Political Events, Political Orientation and Socio-Political Extremism
right-wing Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 4.9% ** -7.5% *** 0.5% -2.1% 4.3% *
Ideology -2.5% -2.3% -4.2% * 2.2% 3.3%
Violence 6.8% * -10.7% *** -0.4% 0% 5.3% **
Intolerance 7.7% *** -8% *** 1.8% -4.1% 4.9% *
Change in Extremism Ranking
ER1 10.33% -0.83% -7.39% 3.96% -0.82%
ER2 8.54% -3.65% -2.4% 1.06% 1.93%
ER3 5.98% -1.75% -4.23% 3.71% -0.24%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.2.2 Center-wing

outputs[["center"]]
Political Events, Political Orientation and Socio-Political Extremism
center-wing Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall -1.3% -1.5% 3.6% -1.4% 11.5% ***
Ideology 9.7% * -3.3% 5.9% 6.6% 5.2%
Violence -3.8% -4.7% 9.7% * -3% 12.5% ***
Intolerance -0.5% -0.9% 1.1% -2.5% 13.4% ***
Change in Extremism Ranking
ER1 2.54% 1.69% -2.11% 3.18% 8.12%
ER2 0.13% 1.38% 1.29% -1% 1.76%
ER3 0.42% -0.42% 0% 0.66% 0.87%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.2.3 Left-wing

outputs[["left"]]
Political Events, Political Orientation and Socio-Political Extremism
left-wing Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 1.5% -3.8% 0.8% 3.1% 12.1% *
Ideology 10% * 5.5% 6.7% * -8.8% 7%
Violence -5.6% -14% * 4.4% 10.5% 22.7% **
Intolerance 2.6% -6.2% -3.4% 6.6% 12.5%
Change in Extremism Ranking
ER1 20.19% 12.5% -38.9% 28.09% 13.38%
ER2 1.61% 1.87% -4.72% 2.81% 2.84%
ER3 1.63% -0.68% -0.94% 0.56% 1.05%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.3 By Religiosity

af_plot_histogram(df, "religiosity")

group_var <- "religiosity"
group_list <- c("Secular", "Traditional", "Religious", "National Religious", "Orthodox")

# Table title
title <- "**Political Events, Religiosity and Socio-Political Extremism**"
subtitle <- "***<group> Jewish Population***"

outputs <- 
  af_pair_compare_by_group(df, 
                           group_var, group_list, 
                           threshold_vars, dimension_vars, 
                           extremism_rank_var_name,
                           events_df, var_labels, 
                           title, subtitle)

7.3.1 Secular

outputs[["Secular"]]
Political Events, Religiosity and Socio-Political Extremism
Secular Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 2.1% 1.7% 1.1% -7.4% ** 11.7% ***
Ideology 4.2% 4.8% 1.1% -5.1% 8%
Violence -0.9% -1.3% 1.1% -7% 17.3% ***
Intolerance 4.7% 1.5% 1.3% -9.1% ** 12% ***
Change in Extremism Ranking
ER1 11.29% 10.16% -17.32% 5.82% 10.63%
ER2 2.77% 2.95% -0.81% -4.22% 3.62%
ER3 1.78% -0.17% -1.61% 0.66% 1.12%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.3.2 Traditional

outputs[["Traditional"]]
Political Events, Religiosity and Socio-Political Extremism
Traditional Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 1.8% -7.2% * 2.5% 3.8% 2.6%
Ideology -1.2% -5.2% 1% 6.6% 2.8%
Violence 1.9% -8.4% 0.7% 6.4% 0.7%
Intolerance 3.9% -8% * 3.3% 3.3% 3.5%
Change in Extremism Ranking
ER1 5.9% -3.64% -1.53% 5.89% -2.9%
ER2 4.68% -0.81% -2.5% 5.22% -1.51%
ER3 3.71% -0.6% -3.11% 3.3% -0.61%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.3.3 Religious

outputs[["Religious"]]
Political Events, Religiosity and Socio-Political Extremism
Religious Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 6% -1.2% 3.6% -7.6% 3.2%
Ideology 12.7% * -6.2% -1.4% -6.9% -2.2%
Violence 7.8% -3% 10.5% -12.1% -0.8%
Intolerance 7% -0.4% 2.1% -5.8% 5.8%
Change in Extremism Ranking
ER1 14% 6% 4.17% -12.94% -7.11%
ER2 11.13% 3.27% 3.33% -9.21% -1.67%
ER3 6.4% 3.6% -10% 6.14% -1.43%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.3.4 National Religious

outputs[["National Religious"]]
Political Events, Religiosity and Socio-Political Extremism
National Religious Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 10.2% -18.3% ** 1.8% 15.8% * -0.1%
Ideology -7.1% -15.5% -1.4% 5.7% 2.9%
Violence 18.4% -26% * 1.5% 26.4% * 0.2%
Intolerance 15.5% * -17.8% ** 2.6% 15.9% -0.9%
Change in Extremism Ranking
ER1 16.45% -9.7% -9.53% 17.83% 3.32%
ER2 12.73% -14.38% -3.57% 14.57% -2.39%
ER3 9.62% -7.44% -2.17% 1.52% 0.41%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001

7.3.5 Orthodox

outputs[["Orthodox"]]
Political Events, Religiosity and Socio-Political Extremism
Orthodox Jewish Population
Between Survey Waves 1-2 2-3 3-4p 4-5 5-6
Events
Type Security Political Political Political Security
Description Operation Break the Wave ('Shover Galim') Approval of the early dissolution of the 24th Knesset Justice Minister Yariv Levin presents the judicial reform Gallant's dismissal The October 7th War
% Change in Extremsim Medians
Overall 6.1% -13.2% *** -6.4% -4.1% 16.1% ***
Ideology 0.9% -3.5% -11.7% 19.8% * -1.3%
Violence 10.1% -23.9% *** -4.2% -2.6% 27.2% ***
Intolerance 7% -11.8% ** -6.3% -10.7% 19.8% ***
Change in Extremism Ranking
ER1 10.31% -7.91% -20.97% 18.78% 5.49%
ER2 9.8% -14.3% -6.47% 2.52% 14.06%
ER3 7.38% -6.23% -1.15% 3.8% 1.33%
Change in extremism medians represents difference in the median of respondents' extremism dimensions. Extremism ranking represent the percentage of respondents with at least 1/2/3 extreme measures. Pairs of panel surveys are marked with 'p'.
* p < 0.05, ** p < 0.01, *** p < 0.001