R Markdown

This is an R Markdown document. Markdown is a simple formatting syntax for authoring HTML, PDF, and MS Word documents. For more details on using R Markdown see http://rmarkdown.rstudio.com.

When you click the Knit button a document will be generated that includes both content as well as the output of any embedded R code chunks within the document. You can embed an R code chunk like this:

# Read the DDI metadata XML file provided with the Pew dataset
PEW_Metadata <- xmlParse(
  "D:/CPR_internship/Datawork/India-Full-Public-Dataset-PEW/Pew India DDI metadata.xml"
)



# List all files available in the dataset folder
list.files("D:/CPR_internship/Datawork/India-Full-Public-Dataset-PEW")
## [1] "CODEBOOK_India.pdf"                                             
## [2] "India recode syntax for public release.txt"                     
## [3] "India Religion Public Data - Pew Research Center (All Vars).csv"
## [4] "India Religion Public Data - Pew Research Center (All Vars).sav"
## [5] "pew-2020_datawork.R"                                            
## [6] "pew-regional-analysis.R"                                        
## [7] "Pew India DDI metadata.xml"                                     
## [8] "READ ME.txt"
# Read the Pew Research Center survey data stored in SPSS format
pew_data <- read_sav(
  "D:/CPR_internship/Datawork/India-Full-Public-Dataset-PEW/India Religion Public Data - Pew Research Center (All Vars).sav"
)

Basic Demographics - Demographic Profile of Respondents

Urban-Rural Distribution

Table 1. Distribution of Respondents by Place of Residence

Table 1. Distribution of Respondents by Place of Residence
Category Count Percentage
Urban 9201 30.67
Rural 20798 69.33

Table 2. Region-wise Distribution of Urban and Rural Respondents

Table 2. Region-wise Distribution of Urban and Rural Respondents
Region Residence Count Percentage
Northeast Urban 236 14.51
Northeast Rural 1390 85.49
North Urban 2381 32.90
North Rural 4857 67.10
Central Urban 396 19.47
Central Rural 1638 80.53
East Urban 1081 15.78
East Rural 5771 84.22
West Urban 2722 40.75
West Rural 3957 59.25
South Urban 2385 42.82
South Rural 3185 57.18

Gender Composition

Table 3. Distribution of Respondents by Gender

Table 3. Distribution of Respondents by Gender
Gender Count Percentage
Male 14594 48.65
Female 15405 51.35

Religious Composition

Table 4. Distribution of Respondents by Religion

Table 4. Distribution of Respondents by Religion
Religion Count Percentage
Hindu 22975 76.59
Muslim 3336 11.12
Christian 1011 3.37
Sikh 1782 5.94
Buddhist 719 2.40
Jain 109 0.36
Some other religion 54 0.18
No religion 13 0.04

Figure 1. Distribution of Respondents by Religion

Figure 1. Distribution of Respondents by Religion

Figure 1. Distribution of Respondents by Religion

Caste Composition

Table 5. Distribution of Respondents by Caste Category

Table 5. Distribution of Respondents by Caste Category
Caste_Category n Percentage
General Category 10346 34.49
Other Backward Class 8080 26.93
Scheduled Caste 7248 24.16
Scheduled Tribe 3127 10.42
Most Backward Class 759 2.53
Other / DK / Refused 439 1.46

Figure 2. Distribution of Respondents by Caste Category

Figure 2. Distribution of Respondents by Caste Category

Figure 2. Distribution of Respondents by Caste Category

General Category Sub-Groups

Table 6. Distribution of General Category Sub-Groups

Table 6. Distribution of General Category Sub-Groups
General_Caste n Percentage
Neither / Just General Category 5927 57.29
Other Forward Caste 2614 25.27
Brahmin 1132 10.94
Don’t Know / Refused 673 6.50

Educational Attainment

Table 7. Distribution of Respondents by Educational Attainment

Table 7. Distribution of Respondents by Educational Attainment
Education n Percentage
No Formal Schooling 13649 45.50
Secondary Education 8168 27.23
Primary Education 5102 17.01
Post-Secondary Education 2723 9.08
Refused 253 0.84
Don’t Know 104 0.35

How People Get Information

This section examines how frequently respondents follow news and the major sources from which they obtain information.

Frequency of Following News

Frequency of Following News

Table 8. Frequency of Following News (Weighted Estimates)
News_Frequency Weighted_Count Percentage
Every Day 12142.76 40.48
Never 4534.78 15.12
More Than Once a Week 4402.04 14.67
Seldom 4279.93 14.27
Once a Week 2473.28 8.24
Once or Twice a Month 1653.90 5.51
A Few Times a Year 338.66 1.13
Don’t Know 126.29 0.42
Refused 47.34 0.16

Frequency of Following News by Place of Residence

Table 9. Frequency of Following News by Place of Residence
News Frequency Rural (%) Urban (%)
A Few Times a Year 1.3 0.9
Every Day 35.7 51.3
More Than Once a Week 14.5 15.3
Never 18.3 8.6
Once a Week 8.5 7.9
Once or Twice a Month 6.5 3.4
Seldom 15.2 12.6
Note:
Percentages are weighted using Pew survey weights and calculated within each residence category.

Frequency of Following News by Place of Residence

Figure 4. Frequency of Following News by Place of Residence

Figure 4. Frequency of Following News by Place of Residence

Sources of News

Sources of News Among News Consumers

Table 10. Sources of News Among News Consumers
News Source Sample N Weighted Percentage
Television 21562 81.7
Family / Friends 6166 30.1
Newspapers / Magazines 7654 29.2
Internet 5747 20.9
Social Media 3108 12.3
Radio 1217 5.2
Other 233 0.8
Note:
Respondents could select up to three sources. Percentages do not sum to 100.

Figure 5. Sources of News Among News Consumers

Figure 5. Sources of News Among News Consumers

Figure 5. Sources of News Among News Consumers

Economic Perceptions

This section examines how respondents evaluate their current economic situation and whether perceptions differ between urban and rural residents.

Current Economic Situation

Table 11. Current Economic Situation

Table 11. Current Economic Situation
Economic_Situation Weighted_Count Percentage
Somewhat good 15445.6 52.0
Very good 7496.3 25.2
Somewhat bad 4651.0 15.7
Very bad 2105.2 7.1

Figure 6. Current Economic Situation

Figure 6. Current Economic Situation

Figure 6. Current Economic Situation

Economic Situation by Place of Residence

Current Economic Situation by Residence

Table 12. Current Economic Situation by Residence
Economic_Situation Rural Urban
Somewhat bad 16.8 13.3
Somewhat good 51.5 53.1
Very bad 8.4 4.4
Very good 23.3 29.3

Current Economic Situation by Residence

Figure 7. Current Economic Situation by Residence

Figure 7. Current Economic Situation by Residence

Future Economic Expectations

Expectations About Future Economic Situation

Table 13. Expectations About Future Economic Situation
Future_Economic_Situation Weighted_Count Percentage
Improve 19751.5 70.0
Remain the same 7679.7 27.2
Worsen 800.5 2.8

Expectations About Future Economic Situation

Figure 8. Expectations About Future Economic Situation

Figure 8. Expectations About Future Economic Situation

Perceptions of Major Problems in India

Respondents were asked whether a set of major issues facing India constituted a very big problem, a moderately big problem, a small problem, or not a problem at all. This section compares perceptions between urban and rural respondents.

Perceptions of Major Problems in India by Place of Residence

Perceptions of Major Problems in India by Place of Residence

Perceptions of Major Problems in India by Place of Residence

Women’s Safety Perceptions

This section examines public perceptions regarding the most important steps to improve women’s safety in India.

Overall Views on Women’s Safety

Table: Views on Improving Women’s Safety in India
Women_Safety Weighted_Count Percentage
Teach boys to respect women 15251.2 50.8
Teach girls appropriate behaviour 7655.8 25.5
Both / Depends / Other 4023.5 13.4
Improve law and order / policing 2044.4 6.8
Don’t Know / Refused 531.8 1.8
Women are already safe 492.3 1.6

Figure: Overall Women’s Safety Views

Gender and Residence Differences

Equal Rights for Women (Q9)

This section examines perceptions regarding the importance of equal rights for women in India.

Overall Importance of Equal Rights for Women

Importance of Equal Rights for Women in India
Women_Rights Weighted_Count Percentage
Very important 24108.4 80.4
Somewhat important 4668.1 15.6
Not too important 626.1 2.1
Don’t Know / Refused 398.7 1.3
Not at all important 197.6 0.7

Figure: Overall Views on Equal Rights

Gender and Residence Differences

Preferences for Governance (Q10)

This section examines respondents’ preferences between democratic governance and strong leadership.

Overall Preferences for Governance

Preferences for Governance in India
Government_Preference Weighted_Count Percentage
Strong Leader 14512.8 48.4
Democratic Government 13713.7 45.7
Don’t Know / Refused 1772.5 5.9

Figure: Overall Governance Preferences

Gender and Residence Differences

Political Influence in Religious Matters (Q12)

This section examines perceptions of how much influence political actors should have in religious matters in India.

Overall Views on Political Influence in Religion

Views on Political Influence in Religious Matters
Political_Influence Weighted_Count Percentage
Some influence 9860.5 32.9
A large influence 8711.2 29.0
No influence at all 5164.1 17.2
Not too much influence 4233.9 14.1
Don’t Know / Refused 2029.3 6.8

Figure: Overall Views

Gender and Residence Differences

Perceived Discrimination in India (Q11)

This section examines perceptions of discrimination faced by different social and religious groups in India, including differences by residence and gender.

Overall Perceptions of Discrimination

Perceived Discrimination Against Social and Religious Groups
Group Weighted_Yes Total_Weighted Percentage
Women 6926.1 28490.8 24.3
Scheduled Castes 6146.9 26839.7 22.9
Scheduled Tribes 5805.0 26291.5 22.1
Hindus 6148.3 28766.5 21.4
Muslims 5411.0 26695.4 20.3
OBCs 4880.8 26765.4 18.2
Christians 3016.0 23311.9 12.9

Figure: Overall Perceived Discrimination

Discrimination by Residence

## Religious Freedom in India (Q17a, Q17b)

religious_freedom_compare <- pew_data %>%
  mutate(
    Residence = ifelse(Urban == 1, "Urban", "Rural"),
    Gender = ifelse(QGEN == 1, "Male", "Female"),
    Group = paste(Residence, Gender, sep = " - ")
  ) %>%
  
  dplyr::select(weight, Group, Q17a, Q17b) %>%
  
  tidyr::pivot_longer(
    cols = c(Q17a, Q17b),
    names_to = "Question",
    values_to = "Response"
  ) %>%
  
  mutate(
    Question = case_when(
      Question == "Q17a" ~ "People of Other Religions",
      Question == "Q17b" ~ "Your Own Religion"
    ),
    
    Freedom = case_when(
      Response == 1 ~ "Very free",
      Response == 2 ~ "Somewhat free",
      Response == 3 ~ "Not too free",
      Response == 4 ~ "Not at all free",
      TRUE ~ NA_character_
    )
  ) %>%
  
  filter(!is.na(Freedom)) %>%
  
  group_by(Question, Group, Freedom) %>%
  
  summarise(
    Weighted_N = sum(weight, na.rm = TRUE),
    .groups = "drop"
  ) %>%
  
  group_by(Question, Group) %>%
  
  mutate(
    Percentage = 100 * Weighted_N / sum(Weighted_N)
  ) %>%
  
  ungroup()

religious_freedom_compare$Group <- factor(
  religious_freedom_compare$Group,
  levels = c(
    "Urban - Male",
    "Urban - Female",
    "Rural - Male",
    "Rural - Female"
  )
)

ggplot(
  religious_freedom_compare,
  aes(
    x = Percentage,
    y = Group,
    fill = Freedom
  )
) +
  
  geom_col(width = 0.75) +
  
  facet_wrap(~Question, ncol = 2) +
  
  scale_fill_manual(
    values = c(
      "Very free" = "#1b7837",
      "Somewhat free" = "#a6dba0",
      "Not too free" = "#f4a582",
      "Not at all free" = "#b2182b"
    )
  ) +
  
  scale_x_continuous(
    labels = function(x) paste0(x, "%"),
    expand = expansion(mult = c(0, 0.02))
  ) +
  
  labs(
    title = "Perceptions of Religious Freedom in India",
    subtitle = "Comparison across urban-rural residence and gender",
    x = "Percentage of Respondents",
    y = NULL,
    fill = NULL,
    caption = "Source: Pew Research Center Survey Data (weighted estimates)"
  ) +
  
  theme_minimal(base_size = 14) +
  
  theme(
    plot.title = element_text(face = "bold", size = 18),
    plot.subtitle = element_text(size = 12, colour = "grey30"),
    strip.text = element_text(face = "bold", size = 13),
    axis.text.y = element_text(face = "bold"),
    legend.position = "bottom",
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank()
  )

## Views on Inter-Caste and Inter-Religious Marriage

This section examines attitudes toward restricting inter-caste and inter-religious marriages. The analysis is conducted separately for women and men, and further compared across urban and rural respondents. Responses are grouped into two categories: Important (very/somewhat important) and Not Important (not too/not at all important).

Figure: Women’s Views on Marriage Restrictions

## Views on Inter-Caste and Inter-Religious Marriage

This section examines attitudes toward inter-caste and inter-religious marriage restrictions. Responses are collapsed into two categories: Important (very/somewhat important) and Not Important (not too/not at all important). The analysis is conducted separately for women and men.

Women’s Views on Marriage Restrictions

## Views on Gender Roles Within the Family

This section examines public perceptions of gender roles within the household. Respondents were asked who should primarily be responsible for (i) earning money, (ii) taking care of children, and (iii) making household financial decisions. Responses are categorized as: Men, Women, Both, and Depends/Other.

Overall Distribution of Views on Family Roles

family_roles <- pew_data %>%
  mutate(
    Group = case_when(
      Urban == 1 & QGEN == 1 ~ "Urban Male",
      Urban == 1 & QGEN == 2 ~ "Urban Female",
      Urban == 2 & QGEN == 1 ~ "Rural Male",
      Urban == 2 & QGEN == 2 ~ "Rural Female"
    )
  ) %>%
  
  dplyr::select(
    weight,
    Group,
    Q22a,
    Q22b,
    Q22c
  ) %>%
  
  pivot_longer(
    cols = c(Q22a, Q22b, Q22c),
    names_to = "Responsibility",
    values_to = "Response"
  ) %>%
  
  mutate(
    Responsibility = case_when(
      Responsibility == "Q22a" ~ "Earning Money",
      Responsibility == "Q22b" ~ "Taking Care of Children",
      Responsibility == "Q22c" ~ "Making Decisions About Expenses"
    ),
    
    Response = case_when(
      Response == 1 ~ "Men",
      Response == 2 ~ "Women",
      Response == 3 ~ "Both",
      Response == 4 ~ "Depends / Other",
      TRUE ~ NA_character_
    )
  ) %>%
  
  filter(!is.na(Response))

family_roles_summary <- family_roles %>%
  group_by(
    Group,
    Responsibility,
    Response
  ) %>%
  summarise(
    Weighted_N = sum(weight),
    .groups = "drop"
  ) %>%
  
  group_by(
    Group,
    Responsibility
  ) %>%
  
  mutate(
    Percentage = 100 * Weighted_N / sum(Weighted_N)
  )


ggplot(
  family_roles_summary,
  aes(
    x = Percentage,
    y = Responsibility,
    fill = Response
  )
) +
  
  geom_col(width = 0.75) +
  
  geom_text(
    aes(
      label = ifelse(
        Percentage >= 10,
        paste0(round(Percentage,1), "%"),
        ""
      )
    ),
    position = position_stack(vjust = 0.5),
    colour = "white",
    fontface = "bold",
    size = 3.8
  ) +
  
  facet_wrap(~ Group, ncol = 2) +
  
  scale_fill_manual(
    values = c(
      "Men" = "#2166ac",
      "Women" = "#b2182b",
      "Both" = "#4d9221",
      "Depends / Other" = "#969696"
    )
  ) +
  
  scale_x_continuous(labels = function(x) paste0(x, "%")) +
  
  labs(
    title = "Views on Family Responsibilities in India",
    subtitle = "Comparison across urban-rural residence and gender",
    x = "Percentage of Respondents",
    y = NULL,
    fill = NULL,
    caption = "Source: Pew Research Survey (2020), weighted estimates"
  ) +
  
  theme_minimal(base_size = 14) +
  
  theme(
    plot.title = element_text(face = "bold", size = 18),
    plot.subtitle = element_text(size = 12, colour = "grey30"),
    strip.text = element_text(face = "bold", size = 13),
    axis.text.y = element_text(face = "bold"),
    legend.position = "bottom",
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank()
  )

## Gender Attitudes in India

This section examines gender attitudes across urban-rural residence and gender. It focuses on two statements: (1) A wife must always obey her husband, and
(2) Men should have more rights to a job when jobs are scarce.

Responses are grouped into Agree and Disagree categories, and weighted percentages are calculated using survey weights.

gender_attitudes <- pew_data %>%
  mutate(
    Group = case_when(
      Urban == 1 & QGEN == 1 ~ "Urban Male",
      Urban == 1 & QGEN == 2 ~ "Urban Female",
      Urban == 2 & QGEN == 1 ~ "Rural Male",
      Urban == 2 & QGEN == 2 ~ "Rural Female"
    )
  ) %>%
  
  dplyr::select(
    weight,
    Group,
    Q23a,
    Q23b
  ) %>%
  
  pivot_longer(
    cols = c(Q23a, Q23b),
    names_to = "Statement",
    values_to = "Response"
  ) %>%
  
  mutate(
    Statement = case_when(
      Statement == "Q23a" ~
        "A wife must always obey her husband",
      
      Statement == "Q23b" ~
        "Men should have more rights to a job\nwhen jobs are scarce"
    ),
    
    Response = case_when(
      Response %in% c(1, 2) ~ "Agree",
      Response %in% c(3, 4) ~ "Disagree",
      TRUE ~ NA_character_
    )
  ) %>%
  
  filter(!is.na(Response))


gender_attitudes_summary <- gender_attitudes %>%
  group_by(
    Group,
    Statement,
    Response
  ) %>%
  summarise(
    Weighted_N = sum(weight),
    .groups = "drop"
  ) %>%
  
  group_by(
    Group,
    Statement
  ) %>%
  
  mutate(
    Percentage = 100 * Weighted_N / sum(Weighted_N)
  )


ggplot(
  gender_attitudes_summary,
  aes(
    x = Percentage,
    y = Group,
    fill = Response
  )
) +
  
  geom_col(width = 0.75) +
  
  geom_text(
    aes(
      label = paste0(round(Percentage,1), "%")
    ),
    position = position_stack(vjust = 0.5),
    colour = "white",
    fontface = "bold",
    size = 4
  ) +
  
  facet_wrap(
    ~ Statement,
    ncol = 2
  ) +
  
  scale_fill_manual(
    values = c(
      "Agree" = "#b2182b",
      "Disagree" = "#2166ac"
    )
  ) +
  
  scale_x_continuous(
    labels = function(x) paste0(x, "%")
  ) +
  
  labs(
    title = "Gender Attitudes in India",
    subtitle = "Comparison by residence and gender",
    x = "Percentage of Respondents",
    y = NULL,
    fill = NULL,
    caption = paste(
      "Questions: (1) A wife must always obey her husband;",
      "(2) When jobs are scarce, men should have more rights",
      "to a job than women.\n",
      "Responses collapsed into Agree (completely/mostly agree)",
      "and Disagree (mostly/completely disagree).\n",
      "Source: Pew Research Center, Religion in India Survey (2020).",
      "Weighted estimates."
    )
  ) +
  
  theme_minimal(base_size = 14) +
  
  theme(
    plot.title = element_text(
      face = "bold",
      size = 18
    ),
    
    plot.subtitle = element_text(
      size = 12,
      colour = "grey30"
    ),
    
    strip.text = element_text(
      face = "bold",
      size = 13
    ),
    
    axis.text.y = element_text(
      face = "bold"
    ),
    
    legend.position = "bottom",
    
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank()
  )

## Views on Inheritance and Political Leadership

This section examines two important dimensions of social attitudes in India: (1) who should have greater rights to parental inheritance, and
(2) perceptions of whether men, women, or both make better political leaders.

The analysis compares responses across gender and rural–urban residence using weighted survey estimates.

inheritance_compare <- pew_data %>%
  mutate(
    Group = case_when(
      Urban == 1 & QGEN == 1 ~ "Urban Male",
      Urban == 1 & QGEN == 2 ~ "Urban Female",
      Urban == 2 & QGEN == 1 ~ "Rural Male",
      Urban == 2 & QGEN == 2 ~ "Rural Female"
    ),
    
    Inheritance_View = case_when(
      Q24 == 1 ~ "Sons",
      Q24 == 2 ~ "Daughters",
      Q24 == 3 ~ "Equal rights",
      Q24 == 4 ~ "Other / Depends"
    )
  ) %>%
  filter(!is.na(Inheritance_View)) %>%
  group_by(Group, Inheritance_View) %>%
  summarise(
    Weighted_N = sum(weight),
    .groups = "drop"
  ) %>%
  group_by(Group) %>%
  mutate(
    Percentage = 100 * Weighted_N / sum(Weighted_N)
  ) %>%
  ungroup()


ggplot(
  inheritance_compare,
  aes(
    x = Group,
    y = Percentage,
    fill = Inheritance_View
  )
) +
  
  geom_col(
    width = 0.75
  ) +
  
  geom_text(
    aes(
      label = ifelse(
        Percentage >= 5,
        paste0(round(Percentage,1), "%"),
        ""
      )
    ),
    position = position_stack(vjust = 0.5),
    colour = "white",
    fontface = "bold",
    size = 4
  ) +
  
  scale_fill_manual(
    values = c(
      "Equal rights" = "#238b45",
      "Sons" = "#cb181d",
      "Daughters" = "#2171b5",
      "Other / Depends" = "#969696"
    )
  ) +
  
  labs(
    title = "Who Should Have Greater Rights to Parents' Inheritance?",
    subtitle = "Comparison by gender and place of residence",
    x = NULL,
    y = "Percentage (%)",
    fill = NULL,
    caption = paste(
      "Question: In your opinion, who should have a greater right",
      "to parents' inheritance?\n",
      "Weighted estimates from Pew Research Center India Survey 2019–2020."
    )
  ) +
  
  theme_minimal(base_size = 14) +
  
  theme(
    plot.title = element_text(face = "bold", size = 18),
    plot.subtitle = element_text(size = 12, colour = "grey30"),
    axis.text.x = element_text(face = "bold"),
    legend.position = "bottom",
    panel.grid.major.x = element_blank(),
    panel.grid.minor = element_blank()
  )

leadership_compare <- pew_data %>%
  mutate(
    Residence = ifelse(Urban == 1, "Urban", "Rural"),
    Gender = ifelse(QGEN == 1, "Male", "Female"),
    
    Group = paste(Residence, Gender),
    
    Leadership_View = case_when(
      Q26 == 1 ~ "Men better leaders",
      Q26 == 2 ~ "Women better leaders",
      Q26 == 3 ~ "Equally good leaders",
      TRUE ~ NA_character_
    )
  ) %>%
  filter(!is.na(Leadership_View)) %>%
  group_by(Group, Leadership_View) %>%
  summarise(
    Weighted_N = sum(weight),
    .groups = "drop"
  ) %>%
  group_by(Group) %>%
  mutate(
    Percentage = 100 * Weighted_N / sum(Weighted_N)
  )


leadership_compare$Group <- factor(
  leadership_compare$Group,
  levels = c(
    "Urban Male",
    "Urban Female",
    "Rural Male",
    "Rural Female"
  )
)


ggplot(
  leadership_compare,
  aes(
    x = Group,
    y = Percentage,
    fill = Leadership_View
  )
) +
  
  geom_col(
    width = 0.75
  ) +
  
  geom_text(
    aes(
      label = ifelse(
        Percentage > 5,
        paste0(round(Percentage,1), "%"),
        ""
      )
    ),
    position = position_stack(vjust = 0.5),
    color = "white",
    fontface = "bold",
    size = 4
  ) +
  
  scale_fill_manual(
    values = c(
      "Men better leaders" = "#b2182b",
      "Women better leaders" = "#2166ac",
      "Equally good leaders" = "#1b7837"
    )
  ) +
  
  scale_y_continuous(
    labels = function(x) paste0(x, "%"),
    expand = expansion(mult = c(0, 0.02))
  ) +
  
  labs(
    title = "Views on Political Leadership by Gender and Residence",
    subtitle = paste(
      "Comparison of opinions on whether men, women,",
      "or both are equally good political leaders"
    ),
    x = NULL,
    y = "Percentage of Respondents",
    fill = NULL,
    caption = paste(
      "Question: Which of these statements comes closest to your opinion?\n",
      "Men generally make better political leaders than women; ",
      "Women generally make better political leaders than men; ",
      "Women and men make equally good political leaders.\n",
      "Source: Pew Research Center, Religion in India Survey (2019–2020). ",
      "Weighted estimates."
    )
  ) +
  
  theme_minimal(base_size = 14) +
  
  theme(
    plot.title = element_text(face = "bold", size = 18),
    plot.subtitle = element_text(size = 12, colour = "grey30"),
    axis.text.x = element_text(face = "bold", size = 12),
    legend.position = "bottom",
    panel.grid.major.x = element_blank(),
    panel.grid.minor = element_blank(),
    plot.caption = element_text(hjust = 0, size = 10)
  )