Tone of Coverage

Rendered at 10:38 AM, Apr. 17, 2023

Ryan Thornburg (ryan.thornburg@unc.edu)

Overall, most media coverage was neutral.

0 = neutral

1 = positive

2 = negative

3 = includes both positive and negative

Code

all_mentions %>% 
  select(2:5, 1, tenor_tone_of_article_re_candidate) %>%
  filter(!is.na(tenor_tone_of_article_re_candidate)) %>%
  #filter(tenor_tone_of_article_re_candidate %in% c("0","1","2")) %>%
  tabyl(tenor_tone_of_article_re_candidate) %>%
  adorn_pct_formatting() %>%
  datatable(options = list(dom='t'),
          colnames = c('Tone', 'Articles', 'Percent'),
        rownames = FALSE)

Article Tone by Incumbents vs. Challengers

Code

all_mentions %>%
  select(2:5, 1, tenor_tone_of_article_re_candidate) %>%
  filter(!is.na(tenor_tone_of_article_re_candidate)) %>%
  #filter(tenor_tone_of_article_re_candidate %in% c("0","1","2")) %>%
  group_by(tenor_tone_of_article_re_candidate, candidate_status) %>%
  summarize(Articles=n()) %>%
  pivot_wider(names_from = "candidate_status", values_from = Articles) %>%
  adorn_percentages(denominator = 'col') %>%
  adorn_pct_formatting() %>%
  datatable(options = list(dom='t'),
          colnames = c('Tone', 'Challengers', 'Incumbents'),
        rownames = FALSE)

Article Tone for Each Candidate

For this, I re-coded invalid codes as “both positive and negative.” Devinney is not included in this table because the candidate received no article mentions.

Code

all_tone<- all_mentions %>%
  select(2:5, 1, tenor_tone_of_article_re_candidate) %>%
  mutate(tenor_tone_of_article_re_candidate = 
           case_when(
             str_detect(tenor_tone_of_article_re_candidate,"0; 1; 2") ~ "3",
            str_detect(tenor_tone_of_article_re_candidate,"3") ~ "3",
            str_detect(tenor_tone_of_article_re_candidate,"1") ~ "1",
            str_detect(tenor_tone_of_article_re_candidate,"2") ~ "2",
            str_detect(tenor_tone_of_article_re_candidate,"0") ~ "0"
                     )) %>%
  filter(!is.na(tenor_tone_of_article_re_candidate)) %>%
           group_by(state, contest, candidate_name, candidate_status,
                    tenor_tone_of_article_re_candidate) %>%
            summarize(Articles =n()) %>%
  pivot_wider(names_from = tenor_tone_of_article_re_candidate,
              names_prefix = "code_",
              names_sort = TRUE,
              values_from = Articles) %>%
  mutate_if(is.numeric, replace_na, 0)

all_tone %>% 
  datatable(
          colnames = c('State', 'Contest', 'Canddiate', 'Status', 'Neutral Articles', 'Mostly Positive', 'Mostly Negative', 'Both'),
        rownames = FALSE)

Candidates who received the most negative coverage.

Code

all_tone %>%
    left_join(competitiveness, b = c("state", "contest")) %>%
  select(3:4,type,contains("code"), 1:2) %>%
    adorn_percentages(denominator = 'row') %>%
  filter(code_2 > .05) %>%
  arrange(desc(code_2) ) %>%
  adorn_pct_formatting() %>%
    datatable(list(pageLength = 16),
          colnames = c('Candidate', 'Cand. Status', 'Compete Type', 'Neutral Articles', 'Mostly Positive', 'Mostly Negative', 'Both', 'State', 'Contest'),
        rownames = FALSE)