Title:

From Ukraine to Gaza: Watching One War Displace Another on the U.S. Cable News Agenda

Authors:

Ken Blake, Jason Bernard Reineke, Jun Zhang, Middle Tennessee State University School of Journalism and Strategic Media

Abstract:

Agenda-setting theory expects issues that appear more often in media coverage to hold greater prominence in the public mind, especially when large segments of the public take interest in, and feel uncertainty about, those issues. The theory assumes limitations on the informational capacity of both the public mind and media coverage. Questions remain, however, about how those limitations operate. Based on empirical analysis of data drawn from a comprehensive cable news content archive using custom code developed by the authors, this paper documents how the volume of coverage between early 2022 and late 2024 by MSNBC, CNN, and Fox News about the war in Ukraine dropped when the networks scrambled to report on the outbreak of war between Israel and Hamas, then subsequently rebounded, although only somewhat, as the volume of coverage about the Israel-Hamas conflict moderated. The findings offer theoretical insights into how issues compete for shares of the cable news content stream’s limited capacity. Download a .pdf of the full paper.

Figures:

Descriptives

Descriptive statistics
Period count mean sd min max med
Period 1 336 296.41667 189.62240 2 850 272.5
Period 2 1464 62.13525 67.63695 0 826 44.0
Period 3 321 22.05919 19.02612 0 153 18.0
Period 4 792 39.35606 46.96791 0 375 25.0

ANOVA results

(Note: Kruskal-Wallis test & Dunn’s post hoc test were comparable)

ANOVA Results
term df sumsq meansq statistic p.value
Period 3.000 18,584,024.925 6,194,674.975 874.810 0.000
Residuals 2,909.000 20,599,120.354 7,081.169 NA NA
Tukey HSD Post-Hoc Test
Comparison Diff Lower Upper p_value
Period 2-Period 1 −234.281 −247.366 −221.197 0.000
Period 3-Period 1 −274.357 −291.240 −257.475 0.000
Period 4-Period 1 −257.061 −271.144 −242.978 0.000
Period 3-Period 2 −40.076 −53.407 −26.745 0.000
Period 4-Period 2 −22.779 −32.321 −13.238 0.000
Period 4-Period 3 17.297 2.985 31.609 0.010

R Code:

if (!require("tidyverse"))
  install.packages("tidyverse")
if (!require("plotly"))
  install.packages("plotly")
library(tidyverse)
library(plotly)

# Defining date range

startdate <- "20220214"
enddate <- "20241114"

# Defining query: Russia

query <- "(russia%20OR%20ukrain)"

# Building the volume dataframe

vp1 <- "https://api.gdeltproject.org/api/v2/tv/tv?query="
vp2 <- "%20market:%22National%22&mode=timelinevol&format=csv&datanorm=raw&startdatetime="
vp3 <- "000000&enddatetime="
vp4 <- "000000"
text_v_url <- paste0(vp1, query, vp2, startdate, vp3, enddate, vp4)
v_url <- URLencode(text_v_url)
v_url
Ukraine <- read_csv(v_url)
Ukraine <- Ukraine %>%
  rename(Date = 1, Ukraine = 3)

# Defining query: Gaza

query <- "(gaza%20OR%20israel)"

# Building the volume dataframe

vp1 <- "https://api.gdeltproject.org/api/v2/tv/tv?query="
vp2 <- "%20market:%22National%22&mode=timelinevol&format=csv&datanorm=raw&startdatetime="
vp3 <- "000000&enddatetime="
vp4 <- "000000"
text_v_url <- paste0(vp1, query, vp2, startdate, vp3, enddate, vp4)
v_url <- URLencode(text_v_url)
v_url
Gaza <- read_csv(v_url)
Gaza <- Gaza %>%
  rename(Date = 1, Gaza = 3)

AllData <- left_join(Ukraine, Gaza)

# Filter AllData for Fox News, CNN, and MSNBC

AllData <- AllData %>%
  arrange(Date) %>%
  filter(Series == "FOXNEWS" |
           Series == "CNN" |
           Series == "MSNBC")

# Add "WeekOf" variable to the data frame

if (!require("lubridate"))
  install.packages("lubridate")
library(lubridate)

AllData$WeekOf <- round_date(AllData$Date,
                             unit = "week",
                             week_start = getOption("lubridate.week.start", 1))

# Add "Period" variable to data frame

AllData <- AllData %>% 
  mutate(Period = case_when(
    (Date < ymd("2022-06-06")) ~ "Period 1",
    (Date < ymd("2023-10-07")) ~ "Period 2",
    (Date < ymd("2024-01-22")) ~ "Period 3",
    TRUE ~ "Period 4"
  ))

### Total volume ###

UkrData <- AllData %>% 
  group_by(WeekOf) %>% 
  summarize (Volume = sum(Ukraine)) %>% 
  mutate(Topic = "Ukraine")

GazData <- AllData %>% 
  group_by(WeekOf) %>% 
  summarize (Volume = sum(Gaza)) %>% 
  mutate(Topic = "Gaza") 

UkrGazData <- rbind(UkrData, GazData)

# Graph the group distributions and averages

Boxplot <- ggplot(AllData, aes(x = Period,
                               y = Ukraine)) +
  geom_boxplot() +
  stat_summary(fun=mean, geom="point", shape=20, size=3, color="red", fill="red") +
  theme_minimal() +
  labs(y = "Coverage (15-sec. clips)",
       title = "Ukraine coverage volume by period, all outlets",
       subtitle = "(Red dots show averages)")

Boxplot

##### Stacked volume chart ###############

### Total volume ###

UkrData <- AllData %>% 
  group_by(WeekOf) %>% 
  summarize (Volume = sum(Ukraine)) %>% 
  mutate(Topic = "Ukraine")

GazData <- AllData %>% 
  group_by(WeekOf) %>% 
  summarize (Volume = sum(Gaza)) %>% 
  mutate(Topic = "Gaza") 

UkrGazData <- rbind(UkrData, GazData)

Figure <- ggplot(UkrGazData, aes(x = WeekOf, y = Volume, fill = Topic)) +
  geom_area() +
  ylab("Volume - total") +
  xlab("Week") +
  geom_vline(xintercept = as.numeric(as.Date("2022-06-06")), linetype = "longdash") +
  geom_vline(xintercept = as.numeric(as.Date("2023-10-07")), linetype = "longdash") +
  geom_vline(xintercept = as.numeric(as.Date("2024-01-22")), linetype = "longdash") +
  annotate(
    "text",
    x = as.Date("2022-03-15"),
    y = max(UkrGazData$Volume) * 1,
    label = "Per. 1",
    color = "black",
    size = 2.5,
    fontface = "bold") +
  annotate(
    "text",
    x = as.Date("2023-03-15"),
    y = max(UkrGazData$Volume) * 1,
    label = "Per. 2",
    color = "black",
    size = 2.5,
    fontface = "bold") +
  annotate(
    "text",
    x = as.Date("2023-12-01"),
    y = max(UkrGazData$Volume) * 1,
    label = "Per. 3",
    color = "black",
    size = 2.5,
    fontface = "bold") +
  annotate(
    "text",
    x = as.Date("2024-07-15"),
    y = max(UkrGazData$Volume) * 1,
    label = "Per. 4",
    color = "black",
    size = 2.5,
    fontface = "bold") +
  theme_minimal() +
  labs(title = "Coverage volume by week, all outlets",
       y = "Coverage (15-sec. clips)",,
       x = "Week & periods")

Figure

# Means comparison

if (!require("gtExtras"))
  install.packages("gtExtras")
library(gtExtras)

mydata <- AllData

# Specify the DV and IV
mydata$DV <- mydata$Ukraine
mydata$Period <- mydata$Period

# Calculate and show the group counts, means, standard
# deviations, minimums, maximums, and medians

Descriptives <- group_by(mydata, Period) %>%
  summarise(
    count = n(),
    mean = mean(DV, na.rm = TRUE),
    sd = sd(DV, na.rm = TRUE),
    min = min(DV, na.rm = TRUE),
    max = max(DV, na.rm = TRUE),
    med = median(DV, na.rm = TRUE)
  )

Desc_Table <- gt(Descriptives) %>% 
  tab_header("Descriptive statistics") %>%
  cols_align(align = "left") %>%
  gt_theme_538

Desc_Table

# Kruskal-Wallis test & Dunn's post hoc test

if (!require("FSA"))
  install.packages("FSA")
library(FSA)
options(scipen=0)
kruskal.test(DV ~ Period, data = mydata)
dunnTest(DV ~ Period, data = mydata)

# ANOVA & Tukey post hoc test

options(scipen = 999)
oneway.test(mydata$DV ~ mydata$Period,
            var.equal = FALSE)

anova_1 <- aov(mydata$DV ~ mydata$Period)
TukeyHSD(anova_1)

# Show ANOVA results in table form

# Load necessary libraries
library(gt)
library(gtExtras)
library(dplyr)
library(broom)

# Run ANOVA
anova_1 <- aov(DV ~ Period, data = mydata)
anova_results <- tidy(anova_1)

# Run Tukey HSD test
tukey_results <- TukeyHSD(anova_1)
tukey_df <- tidy(tukey_results)

# Format ANOVA table with gt
anova_table <- anova_results %>%
  gt() %>%
  tab_header(title = "ANOVA Results") %>%
  fmt_number(columns = where(is.numeric), decimals = 3) %>%
  gt_theme_538() 

# Format Tukey HSD results with gt

tukey_table <- tukey_df %>%
  select(contrast, estimate, conf.low, conf.high, adj.p.value) %>%
  rename(Comparison = contrast, Diff = estimate, Lower = conf.low, Upper = conf.high, p_value = adj.p.value) %>%
  gt() %>%
  tab_header(title = "Tukey HSD Post-Hoc Test") %>%
  fmt_number(columns = where(is.numeric), decimals = 3) %>%
  data_color(columns = "p_value",
             colors = scales::col_numeric(palette = c("darkgray", "lightgray", "white"), domain = c(0, 1))) %>%
  gt_theme_538() # Adding a theme for styling

# Display tables

anova_table
tukey_table