The successor to the meta-analyses-focused FORRT Replication Database (FReD) is a library of replications and reproductions that does not include methodological features such as sample and effect sizes.
-
Notes: This is heavily based on the fred-mini creator. Currently, it includes code to download crossref keywords. It is, however, also possible to attach the already downloaded keywords instead of getting them all from scratch.
-
-
TO DOs
-
-
exclude meta papers without URLs (Boyce, Soto, possible others)
-
add further variables
-
-
-
library(dplyr)
-
-
-Attache Paket: 'dplyr'
-
-
-
Die folgenden Objekte sind maskiert von 'package:stats':
-
- filter, lag
-
-
-
Die folgenden Objekte sind maskiert von 'package:base':
-
- intersect, setdiff, setequal, union
Das folgende Objekt ist maskiert 'package:purrr':
-
- flatten
-
-
library(glue)
-library(stringr)
-library(rlang)
-
-
-Attache Paket: 'rlang'
-
-
-
Die folgenden Objekte sind maskiert von 'package:jsonlite':
-
- flatten, unbox
-
-
-
Die folgenden Objekte sind maskiert von 'package:purrr':
-
- %@%, flatten, flatten_chr, flatten_dbl, flatten_int, flatten_lgl,
- flatten_raw, invoke, splice
-
-
library(httr)
-
-reported_success <-read_csv(
-"https://docs.google.com/spreadsheets/d/e/2PACX-1vSlPVImL7kjzMfpfX08JtAikvEiBg8uQ3HyEJs-YvuQMUa2RMvip-4v8-DilWl2PdwOD_UTD8gJ5wKA/pub?gid=984458430&single=true&output=csv",
-skip =1,
-show_col_types =FALSE
-) %>%
-# ✅ validate DOI prefix: starts with 10., has 4–9 digits, and then a slash
-filter(str_detect(doi_o, regex("^10\\.[0-9]{4,9}/", ignore_case =TRUE))) %>%
-filter(str_detect(doi_r, regex("^10\\.[0-9]{4,9}/", ignore_case =TRUE)))
-
-
Warning: One or more parsing issues, call `problems()` on your data frame for details,
-e.g.:
- dat <- vroom(...)
- problems(dat)
-
-
-
-
-
Select relevant fields and deduplicate (we do not report study level)
get_openalex_keywords <-function(doi, mailto =NULL, fallback_to_concepts =TRUE, n_concepts =10) {
- base_url <-"https://api.openalex.org/works/"
-# Encode only the DOI part so slashes become %2F but keep "doi:" literal
- encoded_doi <-URLencode(doi, reserved =TRUE)
- url <-paste0(base_url, "doi:", encoded_doi)
-
- q <-list(select ="id,doi,display_name,keywords,concepts")
-if (!is.null(mailto)) q$mailto <- mailto
-
- resp <-GET(url, query = q, user_agent("R script (OpenAlex lookup)"))
-if (http_error(resp)) {
-stop(sprintf("HTTP %s: %s", status_code(resp), content(resp, as ="text", encoding ="UTF-8")))
- }
- x <-fromJSON(content(resp, as ="text", encoding ="UTF-8"), simplifyVector =TRUE)
-
-# Try explicit keywords first
- kws <-NULL
-if (!is.null(x$keywords)) {
-if (is.character(x$keywords)) {
- kws <- x$keywords
- } elseif (is.data.frame(x$keywords)) {
- nm <-names(x$keywords)
- pick <-intersect(c("display_name", "keyword", "name", "value"), nm)
-if (length(pick)) kws <-unique(na.omit(as.character(x$keywords[[pick[1]]])))
- } elseif (is.list(x$keywords)) {
-# list of lists
- kws <-unique(na.omit(as.character(unlist(lapply(
- x$keywords,
-function(k) k[["display_name"]] %||% k[["keyword"]] %||% k[["name"]] %||% k
- )))))
- }
- }
-
-# Fallback to top-N concepts if no explicit keywords
-if ((is.null(kws) ||length(kws) ==0) && fallback_to_concepts &&!is.null(x$concepts)) {
- cons <- x$concepts
-if (is.data.frame(cons) &&"display_name"%in%names(cons)) {
-if ("score"%in%names(cons)) {
- cons <- cons[order(cons$score, decreasing =TRUE), , drop =FALSE]
- }
- kws <-head(unique(as.character(cons$display_name)), n_concepts)
- } elseif (is.list(cons)) {
-# list of concepts
- dn <-unlist(lapply(cons, function(cn) cn[["display_name"]]))
- sc <-unlist(lapply(cons, function(cn) cn[["score"]]))
- ord <-order(ifelse(is.na(sc), 0, sc), decreasing =TRUE)
- kws <-head(unique(as.character(dn[ord])), n_concepts)
- }
- }
-
-# Clean up
- kws <-unique(trimws(kws))
- kws[nzchar(kws)]
-}
-
-`%||%`<-function(a, b) if (!is.null(a) &&length(a) >0) a else b
-
-
-dois <-unique(fred_mini$doi_o)
-dois <- dois[!is.na(dois) &grepl("^10", dois)]
-dois <- dois[!is.na(dois)]
-
-# dois is vector of DOIs
-kwdat <-data.frame(
-doi_o = dois,
-keywords =NA_character_,
-error =NA_character_,
-stringsAsFactors =FALSE
-)
-
-# # download from openalex
-# for (idx in seq_along(dois)) {
-# res <- tryCatch(
-# {
-# kw <- get_openalex_keywords(dois[idx], mailto = "lukas.roeseler@uni-muenster.de")
-# list(kw = kw, err = NA_character_)
-# },
-# error = function(e) list(kw = character(0), err = conditionMessage(e))
-# )
-#
-# kwdat$keywords[idx] <- if (length(res$kw) > 0) paste(res$kw, collapse = ", ") else NA_character_
-# kwdat$error[idx] <- res$err
-# }
-
-# open already downloaded keywords
-kwdat <-read.csv("https://raw.githubusercontent.com/forrtproject/FReD-data/refs/heads/main/data%20augmentation/doio_keywords.csv")
-kwdat <- kwdat[ ,-1]
-
-fred_mini <-merge(fred_mini, kwdat[, 1:2], by ="doi_o")
-
-
-
-
Manually coded values
-
-
# code any of these entries as meta papers
-metapapers <-c(
-"10.1038--s41562-024-02062-9",
-"10.3389/fcomm.2022.1048896", # 10 cases of sensory research
-"10.1073/pnas.2103313118", # scarcity effects
-"10.1177/0956797619831612", # Soto
-"10.1007/s13164-018-0400-9", # X-PHI
-"10.1038/s41562-018-0399-z", # Camerer 2018
-"10.1098/rsos.231240", # Boyce
-"10.1027/1864-9335/a000178", # ML1
-"10.1177/2515245918810225", # ML2
-"10.1016/j.jesp.2015.10.012", # ML3
-# "10.1177/2515245920958687", # ML5 NOT because it is focusing on a single study
-"10.1126/science.aac4716", # RP:P
-"10.1038/s41562-024-02062-9", # Holzmeister et al. 2024
-"10.1177/08902070221094216", # SVO development
-"10.1126/science.aaf0918"# Camerer 2016
-)
-
-
-# determine metapaper_r = 1 for metapapers
-fred_mini$metapaper_r <-as.integer(fred_mini$doi_r %in% metapapers)
-
-# excluding cases
-
-# keep only those with url_r non missing: Boyce, Soto, XPHI
-# TO DO: [check others]