In bryanvictor/BibWrangleR: Software Tool for Converting Bibliographic Text Files into an Analyzable Data Frame

The data for this paper comes from the Web of Science database. Only disciplinary social work journals with articles from all 5 years (2005-2009) were included. (18 journals; 3066 articles)

setwd("C:/Users/bvictor/Desktop/Git/BibWrangleR/Disseminating high impact social work scholarship_Hodge et al 2017")

library(dplyr)
library(stringr)

DF <- read.csv("full_df.csv")

names(DF) <- tolower(names(DF))
names(DF) <- gsub("\\.", "_", names(DF))

#Assign article IDs
DF <- DF %>%
  mutate(articleID = 1:nrow(DF))

#Remove journals without five years of data

fivefull <- DF %>%
  group_by(source_title) %>%
  summarise(N=length(unique(publication_year))) %>%
  filter(N==5)

ids <- fivefull$source_title

DF <- DF[DF$source_title %in% ids,]

DF <- mutate(DF, articleID_wos=1:nrow(DF))

duplicates <- DF$title
duplicates <- tolower(duplicates)
duplicates <- gsub(" ", "", duplicates)
duplicates <- gsub("[[:punct:]]", "", duplicates)
duplicates <- substring(duplicates, 1, 50)
duplicates <- duplicated(duplicates)


DF <- cbind(DF, duplicates)

DF <- filter(DF, duplicates==FALSE)

Variable construction

This section creates xxx variables that will be used in the analysis. They include:

1. Number of authors per article
2. Citations year 1 - year 5
3. Total citations 5 full calendar years after publication
4. 5 year h-index (2005-2009) of the journal in which the article was published

# 1. Number of authors per article
DF <- DF %>%
  mutate(n_authors = str_count(authors, ";") + 1)

DF <-DF %>%
  mutate(coauthored = ifelse(n_authors>1, 1, 0))

#2. Citation counts years 1 - 5

cites05 <- DF %>%
  filter(publication_year==2005) %>%
  select(articleID, year0=x2005, year1=x2006, year2=x2007, year3=x2008, year4=x2009, year5=x2010) %>%
  mutate(total_cites = year0+year1+year2+year3+year4+year5)


cites06 <- DF %>%
  filter(publication_year==2006) %>%
  select(articleID, year0=x2006, year1=x2007, year2=x2008, year3=x2009, year4=x2010, year5=x2011) %>%
  mutate(total_cites = year0+year1+year2+year3+year4+year5)

cites07 <- DF %>%
  filter(publication_year==2007) %>%
  select(articleID, year0=x2007, year1=x2008, year2=x2009, year3=x2010, year4=x2011, year5=x2012) %>%
  mutate(total_cites = year0+year1+year2+year3+year4+year5)

cites08 <- DF %>%
  filter(publication_year==2008) %>%
  select(articleID, year0=x2008, year1=x2009, year2=x2010, year3=x2011, year4=x2012, year5=x2013) %>%
  mutate(total_cites = year0+year1+year2+year3+year4+year5)

cites09 <- DF %>%
  filter(publication_year==2009) %>%
  select(articleID, year0=x2009, year1=x2010, year2=x2011, year3=x2012, year4=x2013, year5=x2014) %>%
  mutate(total_cites = year0+year1+year2+year3+year4+year5)

cites <- rbind(cites05, cites06, cites07, cites08, cites09)

DF <- left_join(DF, cites)

DF <- within(DF, quartile <- as.integer(cut(total_cites, quantile(total_cites, probs=0:4/4), include.lowest=TRUE)))

DF <- mutate(DF, quartile = 5-quartile)

DF$cited <- ifelse(DF$total_cites == 0, 0, 1)

#h-index using articles published between 2005-2009 and their citations 5 years after publication

hindex <- DF %>%
  group_by(source_title) %>%
  arrange(desc(total_cites)) %>%
  mutate(ranking = 1:length(source_title)) %>%
  filter(total_cites>=ranking) %>%
  summarise(hindex=max(ranking))

DF <- left_join(DF, hindex)

#Page count

#DF$beginning_page <- as.numeric(DF$beginning_page)
#DF$ending_page <- as.numeric(DF$ending_page)

#DF <- DF %>%
  #mutate(page_count = ending_page - beginning_page + 1)

library(ggplot2)

tiff(file = "figure 2.tif", width = 2800, height = 2800, units = "px", res = 400) 

ggplot(DF, aes(total_cites)) +
  geom_histogram(col="grey10", 
                 fill="grey20", 
                 alpha = .8,
                 binwidth = 1) +
  theme_bw() +
  scale_x_continuous(breaks=seq(0,120,15), limits = c(0,120)) +
  xlab("5-Year Total Citations") +
  ylab("Count") +
  geom_segment(aes(x = 13, y = 489, xend = 2.5, yend = 489), 
               arrow = arrow(length = unit(0.25, "cm"))) +
  geom_segment(aes(x = 15, y = 314, xend = 4.5, yend = 314), 
               arrow = arrow(length = unit(0.25, "cm"))) +
  geom_segment(aes(x = 17, y = 217, xend = 6.5, yend = 217), 
               arrow = arrow(length = unit(0.25, "cm"))) +
  annotate("text", x = 25, y = 315, label = "median = 3", size = 4) +
  annotate("text", x = 30, y = 218, label = "3rd quartile = 5", size = 4) +
  annotate("text", x = 25, y = 490, label = "1st quartile = 1", size = 4)


dev.off() 

The page number obtained via Web of Science appear to be incorrect in a large number of cases calling into question the validity of this variable in the Web of Science data. Page data is therefore pulled from the Social Work Research Database.

library(tidyr)

ebsco_full <- readRDS("ebsco_full.RDS")

ebsco <- filter(ebsco_full, attributes=="article")


ebsco$ebsco.titles <- ebsco$record
ebsco$ebsco.titles <- tolower(ebsco$ebsco.titles)
ebsco$ebsco.titles <- gsub("[[:punct:]]", "", ebsco$ebsco.titles)
ebsco$ebsco.titles <- gsub(" ", "", ebsco$ebsco.titles) 
ebsco$ebsco.titles <- substring(ebsco$ebsco.titles, 1, 45)
ebsco$ebsco.dup <- duplicated(ebsco$ebsco.titles)

ebsco <- ebsco %>%
  filter(ebsco.dup==FALSE) %>%
  select(-ebsco.dup)

ebsco.titles <- ebsco$ebsco.titles

wos <- DF%>%
  select(articleID_wos, title)

wos$wos.titles <- tolower(wos$title)
wos$wos.titles <- gsub("[[:punct:]]", "", wos$wos.titles)
wos$wos.titles <- gsub(" ", "", wos$wos.titles) 
wos$wos.titles <- substring(wos$wos.titles, 1, 45)
wos.dup <- duplicated(wos$wos.titles)

wos <- cbind(wos, wos.dup)

wos <- wos %>%
  filter(wos.dup==FALSE) %>%
  select(articleID_wos, ebsco.titles=wos.titles, -wos.dup)

wos <- left_join(wos, ebsco)

temp_ids <- wos$articleID

ebsco_wos <- ebsco_full[ebsco_full$articleID %in% temp_ids, ]
ebsco_wos <- filter(ebsco_wos, attributes=="pages")

pages <- select(ebsco_wos, articleID, record)

nums <- pages %>% 
  separate(record, into = paste(c("first_page", "last_page"), sep = "-"))

nums$first_page<- gsub("[[:alpha:]]", "", nums$first_page)

nums$first_page <- as.numeric(nums$first_page)
nums$last_page <- as.numeric(nums$last_page)
nums <- nums %>%
  mutate(count = last_page - first_page + 1)

nums <- mutate(nums, count = ifelse(is.na(count), first_page + 1, count))

wos <- left_join(wos, nums)

wos <- select(wos, articleID_wos, page_count=count)

DF <- left_join(DF, wos)

DF <- DF %>%
  mutate(page_count = ifelse(page_count > 100, NA, page_count))

missing <- filter(DF, is.na(page_count))

PQ_full <- readRDS("proquest.RDS")

PQ <- filter(PQ_full, attributes=="article")

PQ$PQ.titles <- PQ$record
PQ$PQ.titles <- tolower(PQ$PQ.titles)
PQ$PQ.titles <- gsub("[[:punct:]]", "", PQ$PQ.titles)
PQ$PQ.titles <- gsub(" ", "", PQ$PQ.titles) 
PQ$PQ.titles <- substring(PQ$PQ.titles, 1, 45)
PQ$PQ.dup <- duplicated(PQ$PQ.titles)

PQ <- PQ %>%
  filter(PQ.dup==FALSE) %>%
  select(-PQ.dup)

PQ.titles <- PQ$PQ.titles

wos <- missing%>%
  select(articleID_wos, title)

wos$wos.titles <- tolower(wos$title)
wos$wos.titles <- gsub("[[:punct:]]", "", wos$wos.titles)
wos$wos.titles <- gsub(" ", "", wos$wos.titles) 
wos$wos.titles <- substring(wos$wos.titles, 1, 45)
wos.dup <- duplicated(wos$wos.titles)

wos <- cbind(wos, wos.dup)

wos <- wos %>%
  filter(wos.dup==FALSE) %>%
  select(articleID_wos, PQ.titles=wos.titles, -wos.dup)

wos <- left_join(wos, PQ)

temp_ids <- wos$articleID

PQ_wos <- PQ_full[PQ_full$articleID %in% temp_ids, ]
PQ_wos <- filter(PQ_wos, attributes=="pages")

pages <- select(PQ_wos, articleID, record)

nums <- pages %>% 
  separate(record, into = paste(c("first_page", "last_page"), sep = "-"))

nums$first_page<- gsub("[[:alpha:]]", "", nums$first_page)

nums$first_page <- as.numeric(nums$first_page)
nums$last_page <- as.numeric(nums$last_page)
nums <- nums %>%
  mutate(count = last_page - first_page + 1)

nums <- mutate(nums, count = ifelse(is.na(count), first_page + 1, count))

wos <- left_join(wos, nums)

wos <- select(wos, articleID_wos, page_count_temp=count)

DF <- left_join(DF, wos)

DF <- DF %>%
  mutate(page_count = ifelse(is.na(page_count), page_count_temp, page_count))

no_page <- filter(DF, is.na(page_count))

no_page_titles <- no_page$title

ebsco <- filter(ebsco_full, attributes=="article")

match_vector <- 99

for(i in 1:length(no_page_titles)) {
  match <- agrep(no_page_titles[i], ebsco$record)
  if(length(match)==0){
    match <- NA
  }
  match_vector <- rbind(match_vector, match)
}

match_vector <- match_vector[2:length(match_vector),]

match_vector <- as.data.frame(match_vector)

match_filter <- match_vector %>%
  na.omit %>%
  arrange(match_vector)

ebsco_filter <- ebsco %>%
  filter(row_number()%in% match_vector$match_vector)

ebsco_filter <- cbind(ebsco_filter, match_filter)

ebsco_ids <- ebsco_filter$articleID

ebsco_wos <- ebsco_full[ebsco_full$articleID %in% ebsco_ids, ]

ebsco_wos <- filter(ebsco_wos, attributes=="pages")

pages <- select(ebsco_wos, articleID, record)

nums <- pages %>% 
  separate(record, into = paste(c("first_page", "last_page"), sep = "-"))

nums$first_page<- gsub("[[:alpha:]]", "", nums$first_page)

nums$first_page <- as.numeric(nums$first_page)
nums$last_page <- as.numeric(nums$last_page)
nums <- nums %>%
  mutate(count = last_page - first_page + 1)

nums <- mutate(nums, count = ifelse(is.na(count), first_page + 1, count))

wos <- no_page %>%
  select(articleID_wos, title)

temp_df <- cbind(no_page, match_vector)

temp_df <- select(temp_df, -articleID)

temp_df <- left_join(temp_df, ebsco_filter)

temp_df <- select(temp_df, articleID_wos, articleID)

temp_df <- left_join(temp_df, nums)

new_pages <- select(temp_df, articleID_wos, count)

DF <- left_join(DF, new_pages)

DF <- DF %>%
  mutate(page_count = ifelse(is.na(page_count), count, page_count))

DF <- select(DF, -count)

DF <- DF %>%
  mutate(page_count = ifelse(page_count > 100, NA, page_count))

no_page <- filter(DF, is.na(page_count))

no_page_titles <- no_page$title

PQ <- filter(PQ_full, attributes=="article")

match_vector <- 99

for(i in 1:length(no_page_titles)) {
  match <- agrep(no_page_titles[i], PQ$record)
  if(length(match)==0){
    match <- NA
  }
  match_vector <- rbind(match_vector, match)
}

match_vector <- match_vector[2:length(match_vector),]

match_vector <- as.data.frame(match_vector)

match_filter <- match_vector %>%
  na.omit %>%
  arrange(match_vector)

PQ_filter <- PQ %>%
  filter(row_number()%in% match_vector$match_vector)

PQ_filter <- cbind(PQ_filter, match_filter)

PQ_ids <- PQ_filter$articleID

PQ_wos <- PQ_full[PQ_full$articleID %in% PQ_ids, ]

PQ_wos <- filter(PQ_wos, attributes=="pages")

pages <- select(PQ_wos, articleID, record)

nums <- pages %>% 
  separate(record, into = paste(c("first_page", "last_page"), sep = "-"))

nums$first_page<- gsub("[[:alpha:]]", "", nums$first_page)

nums$first_page <- as.numeric(nums$first_page)
nums$last_page <- as.numeric(nums$last_page)
nums <- nums %>%
  mutate(count = last_page - first_page + 1)

nums <- mutate(nums, count = ifelse(is.na(count), first_page + 1, count))

wos <- no_page %>%
  select(articleID_wos, title)

temp_df <- cbind(no_page, match_vector)

temp_df <- select(temp_df, -articleID)

temp_df <- left_join(temp_df, PQ_filter)

temp_df <- select(temp_df, articleID_wos, articleID)

temp_df <- left_join(temp_df, nums)

new_pages <- select(temp_df, articleID_wos, count)

DF <- left_join(DF, new_pages)

DF <- DF %>%
  mutate(page_count = ifelse(is.na(page_count), count, page_count))

DF <- select(DF, -count)

page_count_manual <- read.csv("page_count.csv")

DF <- left_join(DF, page_count_manual)

rm(page_count_manual)

DF <- mutate(DF, page_count = ifelse(is.na(page_count), page_count_manual, page_count))

library(BibWrangleR)

SWA <- ebscoBWR.f(path="./SWA")

SWA <- SWA %>%
  filter(attributes == "article")

SWA_titles <- select(SWA, record)

SWA_titles$record <- tolower(SWA_titles$record)
SWA_titles$record <- gsub(" ", "", SWA_titles$record)
SWA_titles$record <- gsub("[[:punct:]]", "", SWA_titles$record)
SWA_titles$record <- substring(SWA_titles$record, 1, 50)
SWA.dup <- duplicated(SWA_titles$record)

SWA_titles <- cbind(SWA_titles, SWA.dup)
SWA_titles <-filter(SWA_titles, SWA.dup==FALSE)

wos <- DF%>%
  select(articleID_wos, title)

wos$wos.titles <- tolower(wos$title)
wos$wos.titles <- gsub("[[:punct:]]", "", wos$wos.titles)
wos$wos.titles <- gsub(" ", "", wos$wos.titles) 
wos$wos.titles <- substring(wos$wos.titles, 1, 50)
wos.dup <- duplicated(wos$wos.titles)
table(wos.dup)

SWA_match <- wos$wos.titles%in%SWA_titles$record
wos <- cbind(wos, SWA_match)
wos <- select(wos, articleID_wos, SWA_match)

DF <- left_join(DF, wos)

not_indexed <- filter(DF, SWA_match == FALSE)

ni_titles <- not_indexed$title

match_vector <- 99


for(i in 1:length(ni_titles)) {
  match <- agrep(ni_titles[i], SWA$record)
  if(length(match)==0){
    match <- NA
  }
  match_vector <- rbind(match_vector, match)
}

match_vector <- match_vector[2:length(match_vector),]

match_vector <- as.data.frame(match_vector)

not_indexed <- cbind(not_indexed, match_vector)

not_indexed <- select(not_indexed, articleID_wos, match_vector)

DF <- left_join(DF, not_indexed)

DF <- mutate(DF, indexed_SWA = ifelse(SWA_match == TRUE | match_vector>=1, 1, 0))


DF <- select(DF, -match_vector)
rm(match_vector)

psyc <- ebscoBWR.f(path = "./psycINFO")

psyc <- psyc %>%
  filter(attributes == "article")

psyc_titles <- select(psyc, record)

psyc_titles$record <- tolower(psyc_titles$record)
psyc_titles$record <- gsub(" ", "", psyc_titles$record)
psyc_titles$record <- gsub("[[:punct:]]", "", psyc_titles$record)
psyc_titles$record <- substring(psyc_titles$record, 1, 50)
psyc.dup <- duplicated(psyc_titles$record)

psyc_titles <- cbind(psyc_titles, psyc.dup)
psyc_titles <-filter(psyc_titles, psyc.dup==FALSE)

wos <- DF%>%
  select(articleID_wos, title)

wos$wos.titles <- tolower(wos$title)
wos$wos.titles <- gsub("[[:punct:]]", "", wos$wos.titles)
wos$wos.titles <- gsub(" ", "", wos$wos.titles) 
wos$wos.titles <- substring(wos$wos.titles, 1, 50)
wos.dup <- duplicated(wos$wos.titles)
table(wos.dup)

psyc_match <- wos$wos.titles%in%psyc_titles$record
wos <- cbind(wos, psyc_match)
wos <- select(wos, articleID_wos, psyc_match)

DF <- left_join(DF, wos)

not_indexed <- filter(DF, psyc_match == FALSE)

ni_titles <- not_indexed$title

match_vector <- 99


for(i in 1:length(ni_titles)) {
  match <- agrep(ni_titles[i], psyc$record)
  if(length(match)==0){
    match <- NA
  }
  match_vector <- rbind(match_vector, match)
}

match_vector <- match_vector[2:length(match_vector),]

match_vector <- as.data.frame(match_vector)

not_indexed <- cbind(not_indexed, match_vector)

not_indexed <- select(not_indexed, articleID_wos, match_vector)

DF <- left_join(DF, not_indexed)

DF <- mutate(DF, indexed_psyc = ifelse(psyc_match == TRUE | match_vector >= 1, 1, 0))

DF <- select(DF, -match_vector)
rm(match_vector)

pubMed_clean <- read.csv("./PubMed/PubMed.csv", stringsAsFactors = FALSE, fileEncoding="latin1")

pubMed_clean <- filter(pubMed_clean, year >= 2004 & year <= 2010)

pubMed <- select(pubMed_clean, title)

pubMed$title <- tolower(pubMed$title)
pubMed$title <- gsub(" ", "", pubMed$title)
pubMed$title <- gsub("[[:punct:]]", "", pubMed$title)
pubMed$title <- substring(pubMed$title, 1, 50)
pubMed.dup <- duplicated(pubMed$title)

pubMed <- cbind(pubMed, pubMed.dup)
pubMed<-filter(pubMed, pubMed.dup==FALSE)

wos <- DF%>%
  select(articleID_wos, title)

wos$wos.titles <- tolower(wos$title)
wos$wos.titles <- gsub("[[:punct:]]", "", wos$wos.titles)
wos$wos.titles <- gsub(" ", "", wos$wos.titles) 
wos$wos.titles <- substring(wos$wos.titles, 1, 50)
wos.dup <- duplicated(wos$wos.titles)
table(wos.dup)

pubMed_match <- wos$wos.titles%in%pubMed$title
wos <- cbind(wos, pubMed_match)
wos <- select(wos, articleID_wos, pubMed_match)

DF <- left_join(DF, wos)

not_indexed <- filter(DF, pubMed_match == FALSE)

ni_titles <- not_indexed$title

match_vector <- 99


for(i in 1:length(ni_titles)) {
  match <- agrep(ni_titles[i], pubMed_clean$title)
  if(length(match)==0){
    match <- NA
  }
  match_vector <- rbind(match_vector, match)
}

match_vector <- match_vector[2:length(match_vector),]

match_vector <- as.data.frame(match_vector)

not_indexed <- cbind(not_indexed, match_vector)

not_indexed <- select(not_indexed, articleID_wos, match_vector)

DF <- left_join(DF, not_indexed)

DF <- mutate(DF, indexed_pubMed = ifelse(pubMed_match == TRUE | match_vector >= 1, 1, 0))

model_frame <- DF %>%
  select(articleID=articleID_wos,
         title,
         authors,
         n_authors,
         coauthored,
         journal=source_title,
         hindex,
         pub_year=publication_year,
         page_count,
         year0,
         year1,
         year2,
         year3,
         year4,
         year5,
         total_cites, 
         quartile,
         indexed_pubMed,
         indexed_psycINFO = indexed_psyc,
         indexed_SWA,
         cited)

model_frame$indexed_pubMed[is.na(model_frame$indexed_pubMed)] <- 0
model_frame$indexed_psycINFO[is.na(model_frame$indexed_psycINFO)] <- 0
model_frame$indexed_SWA[is.na(model_frame$indexed_SWA)] <- 0

model_frame <- model_frame %>%
  mutate(authors_factor = ifelse(n_authors >=4, "4+", n_authors)) %>%
  mutate(database_factor = indexed_pubMed + indexed_psycINFO + indexed_SWA)

model_frame$database_factor <- as.factor(model_frame$database_factor)


coverage <- model_frame %>%
  group_by(journal) %>%
  summarise(N = length(journal),
            coauthored = round(sum(coauthored)/length(journal)*100, digits=1),
            SWA_coverage = round(sum(indexed_SWA/length(journal))*100, digits=1),
            psycINFO_coverage = round(sum(indexed_psycINFO/length(journal))*100, digits=1),
            pubMed_coverage = round(sum(indexed_pubMed/length(journal))*100, digits=1))

#saveRDS(model_frame, file="analytic_frame.RDS")

model_frame <- readRDS("analytic_frame.RDS")

### NEED TO BE UPDATED TO REFLECT CHANGES IN MODEL FRAME


table1 <- model_frame %>%
  group_by(journal, hindex, pubMed, psycINFO) %>%
  summarise(N=length(journal),
            percent_coauth = mean(coauthored),
            percent_no_cite = mean(no_cite)) 

table1 <- select(table1, N, hindex, percent_coauth, pubMed, psycINFO)
table1$percent_coauth <- round(table1$percent_coauth*100, digits=0)

table2 <- model_frame %>%
  mutate(authors = ifelse(n_authors >=4, "4+", n_authors)) %>%
  group_by(authors) %>%
  summarise(N=length(authors),
            avg=mean(total_cites),
            sd=sd(total_cites))

table3 <- model_frame %>%
   mutate(authors = ifelse(n_authors >=4, "4+", n_authors)) %>%
  group_by(pub_year, authors) %>%
summarise (avg=mean(total_cites)) 

table3$avg = round(table3$avg, digits=1)

ggplot(table3, aes(x=authors, y=avg, fill=authors)) +
  geom_bar(stat="identity") +
  facet_grid(~pub_year) +
  scale_fill_grey(start = .6, end = .25) +
  theme_bw() +
  xlab("Authors per Article") +
  ylab("Average Citations 5 Years after Publication") +
  geom_text(data=table3, aes(x=authors, y=avg+0.2, label=avg, size=3)) +
  theme(legend.position="none")

table4 <- model_frame %>%
   mutate(authors = ifelse(n_authors >=4, "4+", n_authors)) %>%
  group_by(pub_year, authors) %>%
summarise (avg=1-mean(cited)) 

table4$avg = round(table4$avg*100, digits=1)

ggplot(table4, aes(x=authors, y=avg, fill=authors)) +
  geom_bar(stat="identity") +
  facet_grid(~pub_year) +
  scale_fill_grey(start = .6, end = .25) +
  theme_bw() +
  xlab("Authors per Article") +
  ylab("Percentage of articles not cited five years after publication") +
  geom_text(data=table4, aes(x=authors, y=avg+0.8, label=avg, size=3)) +
  theme(legend.position="none")


cor05 <- model_frame %>%
  filter(pub_year==2005)

cor06 <- model_frame %>%
  filter(pub_year==2006)

cor07 <- model_frame %>%
  filter(pub_year==2007)

cor08 <- model_frame %>%
  filter(pub_year==2008)

cor09 <- model_frame %>%
  filter(pub_year==2009)

cor.test(cor05$total_cites, cor05$n_authors)
cor.test(cor06$total_cites, cor06$n_authors)
cor.test(cor07$total_cites, cor07$n_authors)
cor.test(cor08$total_cites, cor08$n_authors)
cor.test(cor09$total_cites, cor09$n_authors)
cor.test(model_frame$total_cites, model_frame$n_authors)


ANOVA <- model_frame %>%
  mutate(authors = ifelse(n_authors >=4, "4+", n_authors))

fit <- aov(total_cites~authors, data=ANOVA)

TukeyHSD(fit)

library(lme4)

mymodel <-
glmmadmb(total_cites ~ authors_factor + pub_year + page_count + database_factor + years_active
(1 | journal), data=model_frame, zeroInflation=TRUE, 
                     family="nbinom")

summary(logit.fit)

index <- model_frame %>%
  group_by(journal) %>%
  mutate(yesPsyc = )

bryanvictor/BibWrangleR documentation built on May 13, 2019, 8:11 a.m.