2. LGBTQ+ Candidate Profile

How Do LGBTQ+ Candidates Differ from the Average?

source(here::here("code", "00_setup.R"))
df <- readRDS(paths$analysis_full_rds)
df <- df %>%
  mutate(ideology_category = factor(ideology_category, levels = ideology_levels))

1 Overview

The previous chapter established who runs for municipal office in Brazil. This chapter asks the central comparative question: How do LGBTQ+ candidates differ from their non-LGBTQ+ counterparts?

We compare the two groups across demographics (age, gender, race, education), political positioning (ideology, party, position), and electoral outcomes. Every comparison shown here treats the non-LGBTQ+ population as the baseline and highlights where LGBTQ+ candidates diverge — and by how much.

The dataset contains 463,601 candidates, of whom 3,134 (0.7%) are identified as LGBTQ+.

Results by Position Type

Analyses in this chapter are presented separately for city councilors (proportional representation) and mayors/vice-mayors (plurality/majority) using tabbed panels. City councilors constitute 93% of all candidates and are shown in the default tab. The “All Candidates” tab pools across positions for reference. This disaggregation matters because the two electoral systems have fundamentally different competitive dynamics.

2 Identification Overview

2.1 How Were LGBTQ+ Candidates Identified?

LGBTQ+ candidates in this dataset were identified through two complementary sources:

• TSE (Tribunal Superior Eleitoral): For the first time in 2024, the TSE allowed candidates to voluntarily declare their sexual orientation and gender identity on official registration forms.
• VOTE LGBT: A civil society project that independently identified LGBTQ+ candidates through public campaigning, social media presence, and community referrals.

Some candidates appear in both sources, which strengthens confidence in their identification. Disclosure source indicates whether a candidate was identified via TSE self-declaration only, VOTE LGBT matching only, or both. The panels below show the distribution across these categories.

render_identification <- function(data, tab_name) {
  lgbtq_pos <- data %>% filter(lgbtq_candidate)

  pal_disclosure <- c("TSE" = "#3498DB", "VOTE" = "#E74C3C", "Both" = "#9B59B6")

  # --- Disclosure Source ---
  cat("### Disclosure Source\n\n")

  lgbtq_pos %>%
    count(disclosure_source, sort = TRUE) %>%
    mutate(pct = format_pct(n / sum(n))) %>%
    rename(`Disclosure Source` = disclosure_source, N = n, `%` = pct) %>%
    cat_kable(align = c("l", "r", "r"))

  p_disc <- lgbtq_pos %>%
    count(disclosure_source) %>%
    mutate(
      pct = n / sum(n),
      label = paste0(format_n(n), "\n(", format_pct(pct), ")")
    ) %>%
    ggplot(aes(x = reorder(disclosure_source, n), y = n, fill = disclosure_source)) +
    geom_col(alpha = 0.9, show.legend = FALSE, width = 0.6) +
    geom_text(aes(label = label), hjust = -0.1, size = 4) +
    coord_flip() +
    scale_y_continuous(expand = expansion(mult = c(0, 0.25))) +
    scale_fill_manual(values = pal_disclosure) +
    labs(
      x = NULL, y = "Number of Candidates",
      title = paste0("Disclosure Source of LGBTQ+ Candidates (", tab_name, ")"),
      subtitle = "TSE self-declaration vs. VOTE LGBT identification project"
    )

  cat_plot(p_disc, paste0("02-disclosure-source-", pos_suffix(tab_name)), height = 5)

  cat("::: {.callout-note}\n")
  cat("## Key Finding: Multiple Identification Pathways\n")
  cat("The overlap between TSE self-declaration and VOTE LGBT identification validates the identification strategy. Candidates identified by both sources provide a high-confidence subset, while those identified by only one source remind us that no single method captures all LGBTQ+ candidates.\n")
  cat(":::\n\n")

  # --- Disclosure by Ideology ---
  cat("### Disclosure by Ideology\n\n")

  cat("Do different disclosure channels capture different types of LGBTQ+ candidates? The tables below cross-tabulate disclosure source with party ideology and geographic region to reveal whether TSE self-declaration and VOTE LGBT identification reach distinct populations.\n\n")

  disc_ideo <- lgbtq_pos %>%
    filter(!is.na(ideology_category), !is.na(disclosure_source)) %>%
    count(disclosure_source, ideology_category) %>%
    group_by(disclosure_source) %>%
    mutate(pct = round(n / sum(n) * 100, 1)) %>%
    ungroup()

  disc_ideo %>%
    mutate(label = paste0(format_n(n), " (", pct, "%)")) %>%
    select(disclosure_source, ideology_category, label) %>%
    pivot_wider(names_from = ideology_category, values_from = label, values_fill = "0") %>%
    rename(Source = disclosure_source) %>%
    cat_kable(align = c("l", "r", "r", "r"))

  p_ideo <- ggplot(disc_ideo, aes(x = disclosure_source, y = n, fill = ideology_category)) +
    geom_col(position = "fill", alpha = 0.85) +
    scale_y_continuous(labels = scales::percent) +
    scale_fill_manual(values = pal_ideology) +
    labs(
      x = "Disclosure Source", y = "Proportion",
      fill = "Ideology",
      title = paste0("Ideological Composition by Disclosure Channel (", tab_name, ")")
    )

  cat_plot(p_ideo, paste0("02-disclosure-ideology-", pos_suffix(tab_name)))

  # --- Disclosure by Region ---
  cat("### Disclosure by Region\n\n")

  disc_region <- lgbtq_pos %>%
    filter(!is.na(region), !is.na(disclosure_source)) %>%
    count(disclosure_source, region) %>%
    group_by(disclosure_source) %>%
    mutate(pct = round(n / sum(n) * 100, 1)) %>%
    ungroup()

  disc_region %>%
    mutate(label = paste0(format_n(n), " (", pct, "%)")) %>%
    select(disclosure_source, region, label) %>%
    pivot_wider(names_from = region, values_from = label, values_fill = "0") %>%
    rename(Source = disclosure_source) %>%
    cat_kable(align = c("l", rep("r", 5)))

  cat("::: {.callout-note}\n")
  cat("## Two Channels, Distinct Populations\n")
  cat("TSE self-declaration captures candidates who voluntarily check the LGBTQ+ box on the official registration form, while VOTE LGBT identifies candidates through civil society networks. Differences in ideological and regional composition across channels suggest these sources are complementary rather than redundant.\n")
  cat(":::\n\n")
}

render_position_tabset(render_identification, df)

City Councilors

2.1.1 Disclosure Source

Disclosure Source	N	%
TSE	2153	70.8%
VOTE + TSE	607	19.9%
VOTE	283	9.3%

Key Finding: Multiple Identification Pathways

The overlap between TSE self-declaration and VOTE LGBT identification validates the identification strategy. Candidates identified by both sources provide a high-confidence subset, while those identified by only one source remind us that no single method captures all LGBTQ+ candidates.

2.1.2 Disclosure by Ideology

Do different disclosure channels capture different types of LGBTQ+ candidates? The tables below cross-tabulate disclosure source with party ideology and geographic region to reveal whether TSE self-declaration and VOTE LGBT identification reach distinct populations.

Source	Left	Center	Right
TSE	667 (31%)	734 (34.1%)	752 (34.9%)
VOTE	162 (57.2%)	90 (31.8%)	31 (11%)
VOTE + TSE	373 (61.4%)	145 (23.9%)	89 (14.7%)

2.1.3 Disclosure by Region

Source	North	Northeast	Center-West	Southeast	South
TSE	225 (10.5%)	755 (35.1%)	154 (7.2%)	696 (32.3%)	323 (15%)
VOTE	11 (3.9%)	45 (15.9%)	18 (6.4%)	154 (54.4%)	55 (19.4%)
VOTE + TSE	30 (4.9%)	136 (22.4%)	42 (6.9%)	261 (43%)	138 (22.7%)

Two Channels, Distinct Populations

TSE self-declaration captures candidates who voluntarily check the LGBTQ+ box on the official registration form, while VOTE LGBT identifies candidates through civil society networks. Differences in ideological and regional composition across channels suggest these sources are complementary rather than redundant.

Mayors & Vice-Mayors

2.1.4 Disclosure Source

Disclosure Source	N	%
TSE	62	68.1%
VOTE + TSE	21	23.1%
VOTE	8	8.8%

Key Finding: Multiple Identification Pathways

The overlap between TSE self-declaration and VOTE LGBT identification validates the identification strategy. Candidates identified by both sources provide a high-confidence subset, while those identified by only one source remind us that no single method captures all LGBTQ+ candidates.

2.1.5 Disclosure by Ideology

Do different disclosure channels capture different types of LGBTQ+ candidates? The tables below cross-tabulate disclosure source with party ideology and geographic region to reveal whether TSE self-declaration and VOTE LGBT identification reach distinct populations.

Source	Left	Center	Right
TSE	34 (54.8%)	15 (24.2%)	13 (21%)
VOTE	6 (75%)	1 (12.5%)	1 (12.5%)
VOTE + TSE	14 (66.7%)	6 (28.6%)	1 (4.8%)

2.1.6 Disclosure by Region

Source	North	Northeast	Center-West	Southeast	South
TSE	7 (11.3%)	17 (27.4%)	3 (4.8%)	21 (33.9%)	14 (22.6%)
VOTE	0	1 (12.5%)	1 (12.5%)	3 (37.5%)	3 (37.5%)
VOTE + TSE	2 (9.5%)	6 (28.6%)	0	9 (42.9%)	4 (19%)

Two Channels, Distinct Populations

TSE self-declaration captures candidates who voluntarily check the LGBTQ+ box on the official registration form, while VOTE LGBT identifies candidates through civil society networks. Differences in ideological and regional composition across channels suggest these sources are complementary rather than redundant.

All Candidates

Note

This tab pools city councilors (proportional representation) and mayors/vice-mayors (plurality). Position-specific results in the other tabs may be more informative.

2.1.7 Disclosure Source

Disclosure Source	N	%
TSE	2215	70.7%
VOTE + TSE	628	20.0%
VOTE	291	9.3%

Key Finding: Multiple Identification Pathways

The overlap between TSE self-declaration and VOTE LGBT identification validates the identification strategy. Candidates identified by both sources provide a high-confidence subset, while those identified by only one source remind us that no single method captures all LGBTQ+ candidates.

2.1.8 Disclosure by Ideology

Do different disclosure channels capture different types of LGBTQ+ candidates? The tables below cross-tabulate disclosure source with party ideology and geographic region to reveal whether TSE self-declaration and VOTE LGBT identification reach distinct populations.

Source	Left	Center	Right
TSE	701 (31.6%)	749 (33.8%)	765 (34.5%)
VOTE	168 (57.7%)	91 (31.3%)	32 (11%)
VOTE + TSE	387 (61.6%)	151 (24%)	90 (14.3%)

2.1.9 Disclosure by Region

Source	North	Northeast	Center-West	Southeast	South
TSE	232 (10.5%)	772 (34.9%)	157 (7.1%)	717 (32.4%)	337 (15.2%)
VOTE	11 (3.8%)	46 (15.8%)	19 (6.5%)	157 (54%)	58 (19.9%)
VOTE + TSE	32 (5.1%)	142 (22.6%)	42 (6.7%)	270 (43%)	142 (22.6%)

Two Channels, Distinct Populations

TSE self-declaration captures candidates who voluntarily check the LGBTQ+ box on the official registration form, while VOTE LGBT identifies candidates through civil society networks. Differences in ideological and regional composition across channels suggest these sources are complementary rather than redundant.

3 Demographic Comparison

How do LGBTQ+ and non-LGBTQ+ candidates compare on core demographic characteristics? The panels below provide a side-by-side comparison of mean age, gender composition (% female), racial breakdown (% White, Black, Brown), and educational attainment (% College+). All variables are defined as in Chapter 1.

render_demographics <- function(data, tab_name) {
  d <- data %>% filter(!is.na(lgbtq_candidate))

  # --- Summary table ---
  cat("### Demographic Summary\n\n")

  demo_summary <- d %>%
    group_by(lgbtq_label) %>%
    summarise(
      N = format_n(n()),
      `Mean Age` = as.character(round(mean(age, na.rm = TRUE), 1)),
      `SD Age` = as.character(round(sd(age, na.rm = TRUE), 1)),
      `% Female` = format_pct(mean(female, na.rm = TRUE)),
      `% White` = format_pct(mean(race_simple == "White", na.rm = TRUE)),
      `% Black` = format_pct(mean(race_simple == "Black", na.rm = TRUE)),
      `% Brown` = format_pct(mean(race_simple == "Brown", na.rm = TRUE)),
      `% College+` = format_pct(mean(education_simple == "College+", na.rm = TRUE)),
      .groups = "drop"
    ) %>%
    rename(Group = lgbtq_label)

  demo_summary %>%
    pivot_longer(-Group, names_to = "Variable", values_to = "Value") %>%
    pivot_wider(names_from = Group, values_from = Value) %>%
    cat_kable(align = c("l", "r", "r"))

  # --- Age distribution ---
  cat("### Age Distribution\n\n")

  if (use_simplified(data)) {
    # Small N: boxplot + jitter
    p_age <- d %>%
      filter(!is.na(age)) %>%
      ggplot(aes(x = lgbtq_label, y = age, fill = lgbtq_label)) +
      geom_boxplot(alpha = 0.5, outlier.shape = NA, width = 0.5) +
      geom_jitter(data = . %>% filter(lgbtq_label == "LGBTQ+"),
                  alpha = 0.4, width = 0.15, size = 1.5, color = pal_lgbtq["LGBTQ+"]) +
      scale_fill_manual(values = pal_lgbtq, guide = "none") +
      labs(x = NULL, y = "Age (years)",
           title = paste0("Age Distribution (", tab_name, ")"),
           subtitle = "Box plot with individual LGBTQ+ observations")
  } else {
    # Large N: density
    medians <- d %>%
      filter(!is.na(age)) %>%
      group_by(lgbtq_label) %>%
      summarise(med = median(age), .groups = "drop")

    p_age <- d %>%
      filter(!is.na(age)) %>%
      ggplot(aes(x = age, fill = lgbtq_label, color = lgbtq_label)) +
      geom_density(alpha = 0.35, linewidth = 0.8) +
      geom_vline(data = medians, aes(xintercept = med, color = lgbtq_label),
                 linetype = "dashed", linewidth = 0.7, show.legend = FALSE) +
      scale_fill_manual(values = pal_lgbtq, name = NULL) +
      scale_color_manual(values = pal_lgbtq, name = NULL) +
      labs(x = "Age (years)", y = "Density",
           title = paste0("Age Distribution (", tab_name, ")"),
           subtitle = "Dashed lines mark group medians")
  }
  cat_plot(p_age, paste0("02-age-", pos_suffix(tab_name)))

  # Age test
  if (sum(d$lgbtq_label == "LGBTQ+") >= 5 && sum(d$lgbtq_label == "Non-LGBTQ+") >= 5) {
    t_age <- t.test(age ~ lgbtq_label, data = d)
    cat(sprintf("Mean age: LGBTQ+ = %.1f years vs. Non-LGBTQ+ = %.1f years (difference: %.1f years, *p* %s).\n\n",
                t_age$estimate[2], t_age$estimate[1],
                abs(diff(t_age$estimate)),
                ifelse(t_age$p.value < 0.001, "< 0.001", paste0("= ", round(t_age$p.value, 3)))))
  }

  # --- Race comparison ---
  cat("### Race/Ethnicity\n\n")

  p_race <- d %>%
    filter(!is.na(race_simple)) %>%
    count(lgbtq_label, race_simple) %>%
    group_by(lgbtq_label) %>%
    mutate(pct = n / sum(n)) %>%
    ungroup() %>%
    ggplot(aes(x = race_simple, y = pct, fill = lgbtq_label)) +
    geom_col(position = position_dodge(width = 0.7), alpha = 0.9, width = 0.6) +
    geom_text(aes(label = format_pct(pct)),
              position = position_dodge(width = 0.7), vjust = -0.5, size = 3.5) +
    scale_y_continuous(labels = percent, expand = expansion(mult = c(0, 0.15))) +
    scale_fill_manual(values = pal_lgbtq, name = NULL) +
    labs(x = NULL, y = "Proportion",
         title = paste0("Race/Ethnicity by LGBTQ+ Status (", tab_name, ")"))

  cat_plot(p_race, paste0("02-race-", pos_suffix(tab_name)))
}

render_position_tabset(render_demographics, df)

City Councilors

3.0.1 Demographic Summary

Variable	Non-LGBTQ+	LGBTQ+
N	428,962	3,043
Mean Age	47	38.5
SD Age	11.5	10.6
% Female	35.3%	52.3%
% White	45.8%	38.4%
% Black	11.7%	23.5%
% Brown	40.9%	36.3%
% College+	31.0%	48.5%

3.0.2 Age Distribution

Mean age: LGBTQ+ = 38.5 years vs. Non-LGBTQ+ = 47.0 years (difference: 8.6 years, p < 0.001).

3.0.3 Race/Ethnicity

Mayors & Vice-Mayors

3.0.4 Demographic Summary

Variable	Non-LGBTQ+	LGBTQ+
N	31,505	91
Mean Age	50.6	40.1
SD Age	11.5	10.7
% Female	19.3%	37.4%
% White	61.4%	57.1%
% Black	5.3%	19.8%
% Brown	32.2%	22.0%
% College+	57.0%	79.1%

3.0.5 Age Distribution

Mean age: LGBTQ+ = 40.1 years vs. Non-LGBTQ+ = 50.6 years (difference: 10.5 years, p < 0.001).

3.0.6 Race/Ethnicity

All Candidates

Note

This tab pools city councilors (proportional representation) and mayors/vice-mayors (plurality). Position-specific results in the other tabs may be more informative.

3.0.7 Demographic Summary

Variable	Non-LGBTQ+	LGBTQ+
N	460,467	3,134
Mean Age	47.3	38.5
SD Age	11.5	10.6
% Female	34.2%	51.9%
% White	46.9%	38.9%
% Black	11.2%	23.4%
% Brown	40.3%	35.9%
% College+	32.8%	49.4%

3.0.8 Age Distribution

Mean age: LGBTQ+ = 38.5 years vs. Non-LGBTQ+ = 47.3 years (difference: 8.8 years, p < 0.001).

3.0.9 Race/Ethnicity

Racial Composition Matters

Differences in racial composition between LGBTQ+ and non-LGBTQ+ candidates have implications for understanding intersectional disadvantage. If LGBTQ+ candidates are disproportionately nonwhite, they may face compounded barriers — a question we explore in depth in Chapter 6.

4 Political Profile

Party ideology scores are drawn from Bolognesi et al.’s expert survey, which rates each party on a 0–10 left-right scale. We classify scores below 4.0 as Left, 4.0–7.1 as Center, and above 7.1 as Right.

render_political <- function(data, tab_name) {
  d <- data %>% filter(!is.na(lgbtq_candidate))

  # --- Ideology distribution table ---
  cat("### Ideology Distribution\n\n")

  d %>%
    filter(!is.na(ideology_category)) %>%
    count(lgbtq_label, ideology_category) %>%
    group_by(lgbtq_label) %>%
    mutate(pct = format_pct(n / sum(n))) %>%
    ungroup() %>%
    pivot_wider(
      names_from = lgbtq_label,
      values_from = c(n, pct),
      names_glue = "{lgbtq_label}_{.value}"
    ) %>%
    rename(Ideology = ideology_category) %>%
    select(Ideology,
           `Non-LGBTQ+ N` = `Non-LGBTQ+_n`, `Non-LGBTQ+ %` = `Non-LGBTQ+_pct`,
           `LGBTQ+ N` = `LGBTQ+_n`, `LGBTQ+ %` = `LGBTQ+_pct`) %>%
    cat_kable(align = c("l", "r", "r", "r", "r"))

  # --- Ideology density ---
  cat("### Ideology Score Distribution\n\n")

  if (use_simplified(data)) {
    p_ideo <- d %>%
      filter(!is.na(ideology_score)) %>%
      ggplot(aes(x = lgbtq_label, y = ideology_score, fill = lgbtq_label)) +
      geom_boxplot(alpha = 0.5, outlier.shape = NA, width = 0.5) +
      geom_jitter(data = . %>% filter(lgbtq_label == "LGBTQ+"),
                  alpha = 0.4, width = 0.15, size = 1.5, color = pal_lgbtq["LGBTQ+"]) +
      geom_hline(yintercept = c(4, 7.1), linetype = "dashed", color = "gray40", linewidth = 0.4) +
      scale_fill_manual(values = pal_lgbtq, guide = "none") +
      labs(x = NULL, y = "Ideology Score (0 = Left, 10 = Right)",
           title = paste0("Ideology Distribution (", tab_name, ")"))
  } else {
    p_ideo <- d %>%
      filter(!is.na(ideology_score)) %>%
      ggplot(aes(x = ideology_score, fill = lgbtq_label, color = lgbtq_label)) +
      geom_density(alpha = 0.35, linewidth = 0.8, bw = 0.5) +
      geom_vline(xintercept = c(4, 7.1), linetype = "dashed", color = "gray40", linewidth = 0.4) +
      annotate("text", x = 2, y = Inf, label = "Left", vjust = 2, color = "#E74C3C", fontface = "bold") +
      annotate("text", x = 5.5, y = Inf, label = "Center", vjust = 2, color = "#95A5A6", fontface = "bold") +
      annotate("text", x = 8.5, y = Inf, label = "Right", vjust = 2, color = "#3498DB", fontface = "bold") +
      scale_fill_manual(values = pal_lgbtq, name = NULL) +
      scale_color_manual(values = pal_lgbtq, name = NULL) +
      labs(x = "Ideology Score (0 = Far Left, 10 = Far Right)", y = "Density",
           title = paste0("Party Ideology Distribution (", tab_name, ")"),
           caption = "Based on Bolognesi et al. expert survey scores.")
  }
  cat_plot(p_ideo, paste0("02-ideology-", pos_suffix(tab_name)))

  # --- Party LGBTQ+ share ---
  cat("### Top Parties: LGBTQ+ Share\n\n")

  n_parties <- if (use_simplified(data)) 10 else 20

  top_parties <- data %>%
    count(party_abbrev, sort = TRUE) %>%
    head(n_parties) %>%
    pull(party_abbrev)

  party_lgbtq <- data %>%
    filter(party_abbrev %in% top_parties, !is.na(lgbtq_candidate)) %>%
    group_by(party_abbrev) %>%
    summarise(
      n_total = n(),
      n_lgbtq = sum(lgbtq_candidate),
      pct_lgbtq = n_lgbtq / n_total,
      .groups = "drop"
    )

  overall_pct <- mean(data$lgbtq_candidate, na.rm = TRUE)

  p_party <- party_lgbtq %>%
    ggplot(aes(x = pct_lgbtq, y = reorder(party_abbrev, pct_lgbtq))) +
    geom_segment(
      aes(xend = 0, yend = reorder(party_abbrev, pct_lgbtq)),
      color = "gray70", linewidth = 0.5
    ) +
    geom_point(aes(size = n_total), color = "#E74C3C", alpha = 0.8) +
    geom_text(aes(label = paste0("n=", n_lgbtq)),
              hjust = -0.3, size = 3, color = "gray30") +
    geom_vline(xintercept = overall_pct, linetype = "dashed", color = "gray40") +
    annotate("text", x = overall_pct, y = 1,
             label = paste0("Overall: ", format_pct(overall_pct)),
             hjust = -0.1, vjust = -0.5, size = 3.5, color = "gray40") +
    scale_x_continuous(labels = percent, expand = expansion(mult = c(0, 0.15))) +
    scale_size_continuous(name = "Total Candidates\nin Party", labels = comma, range = c(2, 8)) +
    labs(x = "LGBTQ+ Share of Party's Candidates", y = NULL,
         title = paste0("LGBTQ+ Candidate Share by Party (", tab_name, ")"),
         subtitle = paste0("Top ", n_parties, " parties by size"),
         caption = "Dashed line = overall LGBTQ+ share.")

  cat_plot(p_party, paste0("02-party-share-", pos_suffix(tab_name)), height = 8)
}

render_position_tabset(render_political, df)

City Councilors

4.0.1 Ideology Distribution

Ideology	Non-LGBTQ+ N	Non-LGBTQ+ %	LGBTQ+ N	LGBTQ+ %
Left	54557	12.7%	1202	39.5%
Center	154581	36.0%	969	31.8%
Right	219824	51.2%	872	28.7%

4.0.2 Ideology Score Distribution

4.0.3 Top Parties: LGBTQ+ Share

Mayors & Vice-Mayors

4.0.4 Ideology Distribution

Ideology	Non-LGBTQ+ N	Non-LGBTQ+ %	LGBTQ+ N	LGBTQ+ %
Left	4917	15.6%	54	59.3%
Center	11174	35.5%	22	24.2%
Right	15414	48.9%	15	16.5%

4.0.5 Ideology Score Distribution

4.0.6 Top Parties: LGBTQ+ Share

All Candidates

Note

This tab pools city councilors (proportional representation) and mayors/vice-mayors (plurality). Position-specific results in the other tabs may be more informative.

4.0.7 Ideology Distribution

Ideology	Non-LGBTQ+ N	Non-LGBTQ+ %	LGBTQ+ N	LGBTQ+ %
Left	59474	12.9%	1256	40.1%
Center	165755	36.0%	991	31.6%
Right	235238	51.1%	887	28.3%

4.0.8 Ideology Score Distribution

4.0.9 Top Parties: LGBTQ+ Share

4.1 Position Distribution

The table below compares how LGBTQ+ and non-LGBTQ+ candidates are distributed across position types (city councilor vs. mayor). Differences may reflect strategic entry decisions shaped by the electoral system.

df %>%
  filter(!is.na(lgbtq_label)) %>%
  count(lgbtq_label, position_simple) %>%
  group_by(lgbtq_label) %>%
  mutate(pct = format_pct(n / sum(n))) %>%
  ungroup() %>%
  pivot_wider(
    names_from = lgbtq_label,
    values_from = c(n, pct),
    names_glue = "{lgbtq_label}_{.value}"
  ) %>%
  rename(Position = position_simple) %>%
  select(Position,
         `Non-LGBTQ+ N` = `Non-LGBTQ+_n`, `Non-LGBTQ+ %` = `Non-LGBTQ+_pct`,
         `LGBTQ+ N` = `LGBTQ+_n`, `LGBTQ+ %` = `LGBTQ+_pct`) %>%
  kable(align = c("l", "r", "r", "r", "r"))

Table 1: Position Distribution: LGBTQ+ vs Non-LGBTQ+

Position	Non-LGBTQ+ N	Non-LGBTQ+ %	LGBTQ+ N	LGBTQ+ %
City Councilor	428962	93.2%	3043	97.1%
Mayor	15637	3.4%	39	1.2%
Vice-Mayor	15868	3.4%	52	1.7%

Executive vs. Legislative Positions

The distribution of LGBTQ+ candidates across position types is informative: proportional systems (city council races) may offer lower barriers to entry for underrepresented groups, since candidates need only clear a party threshold rather than win a plurality. The data above reveal whether LGBTQ+ candidates are differentially concentrated in one position type.

5 Electoral Outcomes

Election rate is the proportion of candidates who won their race (elected = 1), including candidates elected outright and those elected by average in the proportional system.

render_outcomes <- function(data, tab_name) {
  d <- data %>% filter(!is.na(elected), !is.na(lgbtq_label))

  # --- Election rate table with CIs ---
  cat("### Election Rate Comparison\n\n")

  election_comparison <- d %>%
    group_by(lgbtq_label) %>%
    summarise(
      N = n(),
      Elected = sum(elected),
      Rate = mean(elected),
      .groups = "drop"
    ) %>%
    rowwise() %>%
    mutate(
      ci = list(binom.test(Elected, N)$conf.int),
      CI_low = ci[1],
      CI_high = ci[2]
    ) %>%
    ungroup() %>%
    select(-ci)

  election_comparison %>%
    mutate(
      `Election Rate` = format_pct(Rate),
      `95% CI` = paste0("[", format_pct(CI_low), ", ", format_pct(CI_high), "]"),
      N = format_n(N),
      Elected = format_n(Elected)
    ) %>%
    select(Group = lgbtq_label, N, Elected, `Election Rate`, `95% CI`) %>%
    cat_kable(align = c("l", "r", "r", "r", "r"))

  # --- Formal test ---
  ct <- d %>% with(table(lgbtq_label, elected))
  test_str <- safe_chisq_str(ct)
  cat(paste0("Difference in election rates: ", test_str, ".\n\n"))

  # --- Election rate bar chart ---
  p_rate <- election_comparison %>%
    ggplot(aes(x = lgbtq_label, y = Rate, fill = lgbtq_label)) +
    geom_col(alpha = 0.9, width = 0.5) +
    geom_errorbar(
      aes(ymin = CI_low, ymax = CI_high),
      width = 0.15, linewidth = 0.7
    ) +
    geom_text(
      aes(label = format_pct(Rate)),
      vjust = -1.5, size = 5, fontface = "bold"
    ) +
    scale_y_continuous(labels = percent, expand = expansion(mult = c(0, 0.2))) +
    scale_fill_manual(values = pal_lgbtq, guide = "none") +
    labs(x = NULL, y = "Election Rate",
         title = paste0("Election Rate (", tab_name, ")"),
         subtitle = "Error bars: 95% confidence intervals (exact binomial)")

  cat_plot(p_rate, paste0("02-election-rate-", pos_suffix(tab_name)), height = 5)
}

render_position_tabset(render_outcomes, df)

City Councilors

5.0.1 Election Rate Comparison

Group	N	Elected	Election Rate	95% CI
Non-LGBTQ+	405,636	57,216	14.1%	[14.0%, 14.2%]
LGBTQ+	2,833	217	7.7%	[6.7%, 8.7%]

Difference in election rates: χ² = 96.2, p < 0.001.

Mayors & Vice-Mayors

5.0.2 Election Rate Comparison

Group	N	Elected	Election Rate	95% CI
Non-LGBTQ+	30,025	10,930	36.4%	[35.9%, 37.0%]
LGBTQ+	87	10	11.5%	[5.7%, 20.1%]

Difference in election rates: χ² = 22.2, p < 0.001.

All Candidates

Note

This tab pools city councilors (proportional representation) and mayors/vice-mayors (plurality). Position-specific results in the other tabs may be more informative.

5.0.3 Election Rate Comparison

Group	N	Elected	Election Rate	95% CI
Non-LGBTQ+	435,661	68,146	15.6%	[15.5%, 15.8%]
LGBTQ+	2,920	227	7.8%	[6.8%, 8.8%]

Difference in election rates: χ² = 135.9, p < 0.001.

5.1 Election Rate by Position (Cross-Reference)

The table below disaggregates election rates across all position types and LGBTQ+ status, providing a single cross-reference.

df %>%
  filter(!is.na(elected), !is.na(lgbtq_label)) %>%
  group_by(position_simple, lgbtq_label) %>%
  summarise(
    N = n(),
    Elected = sum(elected),
    Rate = mean(elected),
    .groups = "drop"
  ) %>%
  rowwise() %>%
  mutate(
    ci = list(binom.test(Elected, N)$conf.int),
    CI_low = ci[1],
    CI_high = ci[2]
  ) %>%
  ungroup() %>%
  select(-ci) %>%
  mutate(
    `Election Rate` = format_pct(Rate),
    `95% CI` = paste0("[", format_pct(CI_low), ", ", format_pct(CI_high), "]"),
    N = format_n(N),
    Elected = format_n(Elected)
  ) %>%
  select(Position = position_simple, Group = lgbtq_label, N, Elected, `Election Rate`, `95% CI`) %>%
  kable(align = c("l", "l", "r", "r", "r", "r"))

Table 2: Election Rates by Position and LGBTQ+ Status

Position	Group	N	Elected	Election Rate	95% CI
City Councilor	Non-LGBTQ+	405,636	57,216	14.1%	[14.0%, 14.2%]
City Councilor	LGBTQ+	2,833	217	7.7%	[6.7%, 8.7%]
Mayor	Non-LGBTQ+	15,015	5,468	36.4%	[35.6%, 37.2%]
Mayor	LGBTQ+	38	2	5.3%	[0.6%, 17.7%]
Vice-Mayor	Non-LGBTQ+	15,010	5,462	36.4%	[35.6%, 37.2%]
Vice-Mayor	LGBTQ+	49	8	16.3%	[7.3%, 29.7%]

Why Confidence Intervals Matter

With a much smaller LGBTQ+ sample, confidence intervals are substantially wider. A point estimate that looks different may not be statistically distinguishable from the non-LGBTQ+ rate once uncertainty is accounted for. Always read the CIs alongside the point estimates.

6 Summary

This chapter documents how LGBTQ+ candidates compare with the general candidate pool across several dimensions:

1. Demographics: The data reveal the extent to which LGBTQ+ candidates differ from non-LGBTQ+ counterparts in age, racial composition, and educational attainment — with implications for understanding who gains access to candidacy.
2. Ideology: The ideological distribution of LGBTQ+ candidates is compared to the non-LGBTQ+ baseline, showing the degree to which LGBTQ+ candidates concentrate in particular parts of the left-right spectrum.
3. Party concentration: LGBTQ+ candidates are not uniformly distributed across parties. Some parties serve as key vehicles for LGBTQ+ candidacies while others have virtually none.
4. Electoral outcomes: The comparison of election rates between LGBTQ+ and non-LGBTQ+ candidates provides a critical descriptive finding that sets the stage for understanding representation gaps.

However, treating “LGBTQ+” as a monolithic category obscures crucial within-group variation. The next chapter disaggregates this umbrella into its constituent identities — Gay, Lesbian, Bisexual+, and Trans — to reveal the heterogeneity that the binary comparison conceals.