Branch: master
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

OBV: Old Bailey Voices 1780-1880

This dataset is released under a Creative Commons Attribution-ShareAlike 4.0 International Licence

License: CC BY-SA 4.0


OBV is derived from two sources: the Old Bailey Corpus (version 2) (OBC) and the Old Bailey Online (OBO). It contains data from all single defendant trials (21023 defendants) in 227 sessions of the Old Bailey Proceedings between 1780 and 1880 which have had linguistic markup added by the Old Bailey Corpus project.

The dataset has been created in order to explore the Voices of Authority research theme of the Digital Panopticon project. It's made available here for other researchers who may find it of interest.

For this project it was essential to correctly associate defendants with their spoken words (not a concern for OBC), as we intend to trace their histories and long-term outcomes using the Digital Panopticon's record linkage. The difficulties involved in ensuring this was done correctly in trials with multiple defendants (in fact, it's quite often impossible!), led to the decision to restrict this dataset to single-defendant trials.

The data has two components:

  • a new version of the tagged speech data from OBC with some additional tagging and OBO defendant IDs.
  • summary data for each defendant (obv_defendants_trials.tsv) containing biographical and trial information

This is version 2.0 of the data (OBV2), January 2017.

Update, May 2017

I have added data extracted from OBC for scribes, publishers, printers and editors for each session (obc2_producers.tsv).

Summary data


This contains data for all single defendants in OBC-tagged sessions, not just OBC-tagged trials.

It includes information about whether there is OBC-tagged speech a) for the trial and b) for the defendant; word counts for the total text of a trial, all OBC-tagged speech, and for the defendant.

For all trials it contains defendant name, gender and age and occupation (if tagged in OBO); offence, verdict and sentence; and the OBO defendant ID

OBO offence, verdict and sentence data is simplified. In an OBO trial there may be multiple offences or outcomes per defendant; in order to flatten this, OBV data represents the most 'serious' offence and outcome (eg if sentenced to both imprisonment and a fine, it retains only imprisonment).

data fields

Field label Description
o2dtid unique ID for data table (has no other meaning)
obo_trial OBO trial ID
obo_deftid OBO defendant ID
sess_date OBO session date yyyymmdd
year OBO session year
trial_tagged Whether the trial is tagged in OBC as containing speech (1=yes, 0=no)
def_spk Whether there is tagged defendant speech (short code): 1=yes, 0=no, NULL=no tagged speech
speech Whether there is tagged defendant speech ("deft_speaks", "deft_silent", or untagged trial="no_speech")
trial_u_count Count of OBC "utterances" (tagged trials only)
trial_speech_wc Total wordcount for OBC-tagged speech (tagged trials only)
trial_total_wc Total wordcount for trial report in OBO (all trials)
deft_u_count Count of OBC utterances by defendant
deft_total_wc Total wordcount for OBC-tagged speech by defendant
deft_u_q Count of questions asked by defendant
deft_u_a Count of answers by defendant
deft_u_d Count of defence statements by defendant
deft_u_s Count of other statements by defendant
deft_given Defendant given name (as tagged in OBO)
deft_surname Defendant surname (OBO)
deft_gender Defendant gender (OBO)
deft_age Defendant age (OBO) (NULL if no tagged age)
deft_occupation Defendant occupation as tagged in OBO
deft_offcat Offence category as tagged in OBO
deft_offsub Offence sub-category as tagged in OBO
deft_vercat Verdict category as tagged in OBO
deft_versub Verdict sub-category as tagged in OBO
deft_puncat Sentence category as tagged in OBO
deft_punsub Sentence sub-category as tagged in OBO

Words data

The OBC XML data was converted to tabular format (in a MySQL database) for data preparation.

One row of data = one OBC tagged utterance (<u> tags). Please note that there are 217,000 rows of data and I cannot give any guarantees that your favourite spreadsheet software can handle this amount of data very well.

For the subset of single-defendant trials:

  • the original speaker roles assigned in OBC were checked and corrected/filled in where necessary, particularly focusing on accurately identifying defendants.
  • broad speech categories were added to each tagged utterance: question, answer, defendant's defence statement; other statement

data fields

Field label Description
obv2wid dataset unique ID
sess_date OBO session date
year OBO session year
obo_trial OBO trial ID
obo_deftid OBO defendant ID
obc_u_no OBC utterance number in trial
obc_event OBC event ID
obc_speaker OBC speaker ID
obc_sex OBC sex of speaker
obc_hiscoLabel OBC hisco data for speaker
obc_hiscoCode OBC hisco data for speaker
obc_class OBC hisco data for speaker
obc_role OBC speaker role
obv_role OBV speaker role: def=defendant; wv= witness or victim; lj = lawyer or judge; jur=juror
words text of words
obv_words_type OBV assigned words type: q=question; a=answer; d=prisoner's defence statement; s=other statement
words_count OBC word count
defendant name of defendant in trial

Offence, verdict and sentence categories and subcategories


offcat offsubcat
breakingPeace assault
breakingPeace barratry
breakingPeace libel
breakingPeace other
breakingPeace riot
breakingPeace threateningBehaviour
breakingPeace vagabond
breakingPeace wounding
damage arson
damage other
deception bankrupcy
deception forgery
deception fraud
deception other
deception perjury
kill infanticide
kill manslaughter
kill murder
kill other
kill pettyTreason
miscellaneous concealingABirth
miscellaneous conspiracy
miscellaneous habitualCriminal
miscellaneous illegalAbortion
miscellaneous kidnapping
miscellaneous other
miscellaneous pervertingJustice
miscellaneous piracy
miscellaneous returnFromTransportation
royalOffences coiningOffences
royalOffences other
royalOffences religiousOffences
royalOffences seditiousLibel
royalOffences seditiousWords
royalOffences seducingAllegiance
royalOffences taxOffences
royalOffences treason
sexual assaultWithIntent
sexual assaultWithSodomiticalIntent
sexual bigamy
sexual indecentAssault
sexual keepingABrothel
sexual other
sexual rape
sexual sodomy
theft animalTheft
theft burglary
theft embezzlement
theft extortion
theft gameLawOffence
theft grandLarceny
theft housebreaking
theft mail
theft other
theft pettyLarceny
theft pocketpicking
theft receiving
theft shoplifting
theft simpleLarceny
theft stealingFromMaster
theft theftFromPlace
violentTheft highwayRobbery
violentTheft other
violentTheft robbery


vercat versubcat
guilty chanceMedley
guilty insane
guilty lesserOffence
guilty manslaughter
guilty pleadedGuilty
guilty pleadedPartGuilty
guilty theftunder100s
guilty theftunder1s
guilty theftunder40s
guilty theftunder5s
guilty withRecommendation
miscVerdict noAgreement
miscVerdict postponed
miscVerdict unfitToPlead
notGuilty accidentalDeath
notGuilty directed
notGuilty fault
notGuilty noEvidence
notGuilty nonComposMentis
notGuilty noProsecutor
notGuilty selfDefence
notGuilty notGuilty noEvidence


puncat punsubcat
corporal pillory
corporal privateWhipping
corporal publicWhipping
corporal whipping
death burning
death deathAndDissection
death drawnAndQuartered
death executed
death hangingInChains
death respited
death respitedForPregnancy
imprison hardLabour
imprison houseOfCorrection
imprison insanity
imprison newgate
imprison otherInstitution
imprison penalServitude
imprison preventiveDetention
miscPunish branding
miscPunish fine
miscPunish forfeiture
miscPunish militaryNavalDuty
miscPunish sureties
noPunish pardon
noPunish sentenceRespited