A crosswalk from the names of speakers and chairs, as written in the debate
transcripts (see get_speeches), to person ids, by parliamentary session.
get_speeches adds these links to its output by default (link = TRUE). With
link = FALSE, the dataset can be joined to the output by speaker_name and
session_id (speakers) or by chair_name and session_id (chairs).
Format
A data frame with 4 columns and 8098 rows
- speaker_name
Name as written in the transcripts (
speaker_nameorchair_namein get_speeches)- session_id
Id of the parliamentary session
- linked_person_id
Id of the person (see get_mp), or
NAwhen the name cannot be linked- link_method
How the name was linked: on full name, first and last name, or as a changed name, to an MP roster and/or a minister spell; "ambiguous" when several persons match; "substitute only, no party or title" when not linked for that reason
Source
Built by data-raw/speaker_links.R from the transcripts
(https://data.stortinget.no/dokumentasjon-og-hjelp/publikasjon/), the MPs of each parliamentary
period (https://data.stortinget.no/dokumentasjon-og-hjelp/representanter/), and a list of ministers
and their periods in office from regjeringen.no, supplemented by the coded biographies
(https://data.stortinget.no/dokumentasjon-og-hjelp/kodet-personbiografi/).
Details
Names are linked to MPs and substitutes of the session's parliamentary period
(get_parlperiod_mps) and to ministers in office during the session. A name
is linked only when exactly one person matches, first on the full name and
then on first and last name. Names with no such match can still be linked as
a changed name (e.g. a surname added or dropped at marriage): when the first
name is the same, every word of the shorter name is in the longer one, exactly
one MP in the period fits, and the name is written with that MP's party
throughout the session. A name that only matches a substitute registration,
and is never written with a party, a title, or as chair in the session (e.g. a
witness in a hearing), is not linked. There are no manual corrections: names
that cannot be resolved this way (e.g. witnesses in hearings, misspelled names,
and some changed names) have linked_person_id NA.
The links are made from names only and do not use the person_id given in
the transcripts from 2016-2017 onward, which is sometimes wrong.
Ministers and their periods in office come from a list of ministers from 1945 to January 2024, supplemented by the cabinet posts in get_mp_bio for ministers in office after that. The minister data will be updated in a later version.
The dataset covers the transcripts available in the API when it was built (sessions 1998-99 onward). The API's list of transcripts for 2008-2009 fails, so these transcripts were found by their ids, constructed from get_session_meetings; open hearings in that session are therefore missing. Four listed transcripts (s031205, s050526k, s602081, o006051) could not be retrieved from the API.
Examples
if (FALSE) { # \dontrun{
speeches <- get_speeches("s140213", link = FALSE)
speeches <- merge(speeches, speaker_links, by = c("speaker_name", "session_id"), all.x = TRUE)
} # }