Calculate margins

Author

Media Innovation Group

Our goal

To find which Texas House district races were within a 5-point and 10-point margin in both the general and primary election for each year? Our data spans from 2012 to 2024.

Updates

  • April 2026: The previous calculations did not take runoffs into account. This has been remedied.
  • December 2025: Updates from the first attempt which compared the highest votes received to the lowest votes received. What we needed was the difference between those finishing 1st and 2nd in each race, which this now does. It’s a complete refactoring.

Setup and import

Expand this to see code 
library(tidyverse)
library(janitor)

# general <- read_rds("data-processed/01-general-results-rep.rds")
# primary <- read_rds("data-processed/01-primary-results-rep.rds")
reps <- read_rds("data-processed/01-house-totals.rds")

Calculate win margin

This summarise method was developed with the help of chatGPT. It’s explained here.

Expand this to see code 
reps_calcs <- reps |> 
  group_by(year, election, district) |>
  mutate(
    total_votes = sum(candvotes),
    pct = (candvotes / total_votes * 100) |> round(2)
  ) |>
  arrange(desc(candvotes), .by_group = TRUE)  |> 
  summarise(
    first_place = first(name),
    first_place_party = first(party),
    first_place_inc = first(incumbent),
    first_place_pct = first(pct),
    second_place = nth(name, 2),
    second_place_party = nth(party, 2),
    second_place_inc = nth(incumbent, 2),
    second_place_pct = nth(pct, 2),
    vote_margin = (first_place_pct - second_place_pct) |> round_half_up(2),
    .groups = "drop"
  ) |>
  mutate(
    runoff = first_place_pct <= 50,
    winner = if_else(!runoff, first_place, NA_character_)
  ) |>
  arrange(year, election, district)

reps_calcs |> head()

Add flags

Here we add flags if the race is competitive within 5 or 10 points.

Just showing the margin and flags to check.

Expand this to see code 
reps_flags <- reps_calcs |> 
  mutate(flg_5 = if_else(vote_margin <= 5, T, F),
         flg_10 = if_else(vote_margin <= 10, T, F))

# peek at flags
reps_flags |>
  head(20) |> 
  select(vote_margin:flg_10)

Test results

2-candidate race

Doing some manual calculations to compare with prepared data.

Expand this to see code 
reps |> 
  filter(year == "2012", election == "General", district == 1) |> 
  adorn_totals()
Expand this to see code 
reps_flags |> filter(year == "2012", election == "General", district == 1)
  • Lavander: (42063 / 50896) * 100 = 82.64500157
  • Eason: (8833 / 50896) * 100 = 17.35499843
  • Diff: 82.64500157 - 17.35499843 = 65.29000314

Multi-candidate race

A test for 3+ candidate race

Expand this to see code 
reps |> 
  filter(year == "2019", election == "House District 148", district == 148) |> 
  adorn_totals()
Expand this to see code 
reps_flags |> filter(year == "2019", election == "House District 148", district == 148)
  • Eastman: (4212 / 20711) * 100 = 20.33701898
  • LaRotta: (3282 / 20711) * 100 = 15.84665154
  • Diff: 20.33701898 - 15.84665154 = 4.49036744

List of elections

Expand this to see code 
reps_flags |> 
  count(election)

Last look

Glimpse

Expand this to see code 
reps_flags |> glimpse()
Rows: 1,278
Columns: 16
$ year               <chr> "2012", "2012", "2012", "2012", "2012", "2012", "20…
$ election           <chr> "Democratic Primary", "Democratic Primary", "Democr…
$ district           <dbl> 35, 37, 39, 40, 43, 74, 75, 77, 80, 90, 95, 101, 11…
$ first_place        <chr> "Longoria", "Oliveira", "Martinez", "Canales", "Gon…
$ first_place_party  <chr> "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "…
$ first_place_inc    <chr> "N", "Y", "Y", "N", "N", "N", "N", "Y", "Y", "Y", "…
$ first_place_pct    <dbl> 55.15, 62.62, 83.88, 30.46, 70.59, 53.82, 52.07, 62…
$ second_place       <chr> "Ruiz", "Dominguez", "Campos", "Hernandez", "Zamora…
$ second_place_party <chr> "D", "D", "D", "D", "D", "D", "D", "D", "D", "D", "…
$ second_place_inc   <chr> "N", "N", "N", "N", "N", "N", "N", "N", "N", "N", "…
$ second_place_pct   <dbl> 44.85, 37.38, 16.12, 26.21, 29.41, 28.28, 36.05, 37…
$ vote_margin        <dbl> 10.30, 25.24, 67.76, 4.25, 41.18, 25.54, 16.02, 24.…
$ runoff             <lgl> FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FAL…
$ winner             <chr> "Longoria", "Oliveira", "Martinez", NA, "Gonzalez T…
$ flg_5              <lgl> FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FAL…
$ flg_10             <lgl> FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FAL…
Expand this to see code 
reps_flags |> slice_sample(n = 10)
Expand this to see code 
# reps_flags |> names() |> clipr::write_clip()

Exports

Sends files for R to data-processed, and .csv to data-export.

Expand this to see code 
# last assignment (doesn't include flags)
reps_margins <- reps_calcs

# just margins for analysis
reps_margins |> 
  write_rds("data-processed/02-house-margins.rds")

# margins with flags for analysis
reps_flags |> 
  write_rds("data-processed/02-house-margins-flags.rds")

# just margins to use in other projects
reps_margins |> 
  write_csv("data-export/02-house-margins.csv")