Outline Andre-Michel´ Guerry and the Rise of Moral Introduction Essai sur la statistique moral de la France Challenges for Multivariable Guerry’s Life and Work Guerry’s Life Michael Friendly Guerry’s Data Guerry’s Works Psychology Department Guerry’s Methods York University Multivariate Analyses: Data-centric Views May 2006 Bivariate displays Joint Statistical Meetings Reduced-rank displays

Multivariate Mapping: Map-centric Views Glyph maps Blended color maps Conditioned choropleth maps

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 1/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 2/59

Introduction Statistique moral de la France Introduction Statistique moral de la France Essai sur la statistique moral de la France The discovery of “social facts” The launching pad of modern social science Stability and Variation Guerry’s results were both compelling and startling: I Rates of crime and suicide remained remarkably invariant over time, yet varied sytematically by region, sex of accused, type of crime, etc. I Presented to Academie des Sciences I In any given French city or department, almost the same number Franc¸ais July 2, 1832 committed suicide, stole, gave birth out of wedlock, etc. I First systematic analysis of comprehensive data on crime, suicide, Year 1826 1827 1828 1829 1830 Avg and other social variables. Sex All accused (%) I Along with Quetelet (1831, 1835), Male 79 79 78 77 78 78 established the study of “moral statistics” Female 21 21 22 23 22 22 7→ modern social science, criminology, Age Accused of Theft (%) sociology 16–25 37 35 38 37 37 37 25–25 31 32 30 31 32 31 Crime Committed in summer (%) Indecent assault . 36 36 35 38 36 Assault & battery . 28 27 27 27 28

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 5/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 6/59 Introduction Statistique moral de la France Introduction Statistique moral de la France The discovery of “social facts” Social context of crime Social laws a´ la physical laws

What to do about crime? Do crime and other moral variables represent: I Crime a serious concern: Explosive growth in Paris, widespread unemployment, emergence of the “dangerous classes.” I structural, lawful characteristics of society, or are they I Liberal (philanthrope) view: increase education, better prison conditions, I simply indicants of individual behaviour? religious instruction, better diet (bread and soup!) I Conservative view: build more prisons, harsher treatment for recidivists! Guerry argued: Each year sees the same number of crimes of the same degree Guerry’s results were startling reproduced in the same regions.(Guerry, 1833, p.10) I Crimes against persons and crimes against property showed different ... We are forced to recognize that the facts of the moral order are distributions over departments of France subject, like those of the physical order, to invariable laws (Guerry, I Crimes against persons unrelated to literacy 1833, p14) I Crimes against property increased with literacy!

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 7/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 8/59

Guerry’s Life and Work Guerry’s Life Guerry’s Life and Work Guerry’s Life Guerry’s Life Guerry’s Life

Sources Accomplishments I Primary: Larousse, Grand Dictionnaire, 1866; necrology and notices by I Three major works on moral statistics: A. Maury, H. Diard, E. Vinet, 1867 I 1829: Statistique comparee´ de l’etat´ de l’instruction...; I Secondary: Whitt, translation of Guerry (1833), Beirne (1993, Ch 4), The I 1833: Essai sur la statistique morale...; Social of Crime, brief mentions, often in relation to Quetelet I 1864: Statistique morale de l’Angleterre comparee...´ by criminologists (Radzinowicz), sociologists (Lazarsfeld), historians I 1833 & 1864 awarded the Moynton Prize for work in statistics by the (Porter, Hacking), ... Academie des Sciences I Invented the ordonateur statistique, to facilitate calculation and tabulation Basic facts of these data (no details survive). [“Ordonateur” adopted by IBM France to avoid the franglais “computeur.”] I Born: Tours, 25 Dec 1802; Died: Paris, 9 Apr 1866 I I Father: Michel Guerry, enterpreneur First analysis of the motives for suicide (content analysis of all suicide I Studied law, literature and physiology at Univ. Poitiers notes found by police in Paris) [Adopted, with little credit, by Durkheim] I Admitted to the bar in Paris, becomes Advocat Royale I Developed methodological ideas for defining indicators of moral variables I 1827: Assigned to work with the crime data collected by the Ministry of I How to measure crime: # of accused? convictions? Justice I National standards for statistics on literacy, education? I 1830: Appointed Director of Criminal Statistics in Ministry of Justice I How to determine relations between variables?

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 11/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 12/59 Guerry’s Life and Work Guerry’s Data Guerry’s Life and Work Guerry’s Works Guerry’s data 1829: Statistique comparee´ de l’etat´ de l’instruction...

I Compte gen´ eral´ de l’administration de la justice criminelle en France I Done with Adriano Balbi I The first national compilation of official justice data (1825) I Single-sheet set of three I detailed data on all charges and disposition shaded maps (darker = I collected quarterly in all 86 departments. worse) I Other sources: Bureau de Longitudes (illegitimate births); Parent-Duchateletˆ I (prostitutes in Paris); Compte du ministere du guerre (military desertions); ... Crime against persons, property (pop per crime) I Moral variables: Scaled so ’more’ is ’better’ I Instruction (# male school Crime pers Population per Crime against persons children) Crime prop Population per Crime against property Donations Donations to the poor Infants Population per illegitimate birth Literacy Percent who can read & write Suicides Population per suicide I Other variables: Ranks by department: wealth, commerce, ...

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 14/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 16/59

Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Works 1829: Statistique comparee´ de l’etat´ de l’instruction... 1833: Essai sur la statistique morale de la France

I Divided the 86 I First shaded thematic departments into 5 maps of crime data regions I First comparative maps of I Supplemented data from social data the Compte gen´ eral´ with: I 7→ crime against persons I Suicides in Paris, seemed inversely related 1794–1832 to crime against property! I Prostitutes in Paris I Instruction: 7→ France (Parent-Duchatelet)ˆ obscure and France I Wealth (taxes per eclair´ ee´ (Dupin, 1826) inhabitant) I Distribution of clergy I North of France highest in I ... education, but also in I First study to use crime property crime! data to ‘test’ hypotheses I Attracted widespread interest in Europe Guerry’s 1833 map of literacy in France

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 17/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 18/59 Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Works 1833: Essai sur la statistique morale de la France 1833: Essai sur la statistique morale de la France

Reproduced maps (good=blue, middle=yellow, bad=red) First comprehensive, national data on moral variables I Crimes against persons I Crimes against property I Literacy Population per Crime against persons Population per Crime against property Per cent who can Read and Write

59 69 5 26 62 56

83 38 53 39 66 85 2 19 60 52 64 78 72 31 68 80 34 24 62 10 42 7 12 65 9 52 65 64 86 71 27 68 15 33 77 79 1322 12 3 1 14 7082 76 73 56 54 21 56 68 76 37 7 79 52 49 46 34 4 6 62 6 68 73 76 55 67 30 68 22 83 74 5 66 13 12 82 60 84 36 49 55 16 5 32 50 37 57 29 47 60 73 51 24 26 48 78 81 67 57 64 20 17 74 47 9 36 22 26 85 54 64 76 53 3 14 65 51 85 50 82 75 43 44 77 29 8 38 40 25 35 11 48 22 70 48 3 77 84 42 86 86 17 30 45 28 8 3 56 43 63 32 71 18 83 80 81 45 40 12 31 26 44 82 56 1 31 53 38 29 72 79 85 10 35 15 61 8 19 33 41 73 46 52 47 26 50 80 2 6 63 23 20 26 58 7 15 61 32 40 35 35 58 21 20 42 35 23 17 25 59 50 29 22 47 16 27 14 42 14 20 18 75 78 17 69 44 56 44 17 31 42 58 39 60 35 11 28 71 74 66 39 3 70 10 4 45 35

Rank 1 - 9 10 - 19 20 - 28 1 Rank 1 - 9 10 - 19 20 - 28 10 Rank 1 - 8 10 - 17 20 - 26 62 29 - 38 39 - 47 48 - 57 29 - 38 39 - 47 48 - 57 29 - 35 38 - 47 48 - 56 58 - 66 67 - 76 77 - 86 58 - 66 67 - 76 77 - 86 58 - 66 68 - 76 77 - 86

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 19/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 20/59

Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Works 1833: Essai sur la statistique morale de la France 1864: Statistique morale de l’Angleterre comparee...´ Dayenu!

Reproduced maps (good=blue, middle=yellow, bad=red) I Proposes to replace simple “moral I Illigitimate births (infants naturelles) statistics” (tables) with “analytical statistics” I Donations to poor I calculation, graphic display I Suicide I 7→ general, abstract results Population per Illegitimate birth Donations to the poor Population per Suicide I 17 large color plates (56 × 39 cm): 8 4 51 55 22 17 17 43 13 I 3 26 37 58 62 56 7 12 43 data for France (1825–1855), England 48 49 3 22 5 36 45 20 46 85 74 34 65 70 49 16 15 41 11 15 28 27 6 23 38 1 23 2935 80 1 2 29 61 39 64 48 53 4 83 67 78 (1834–1855) 75 25 50 29 84 39 25 31 40 20 8 50 84 46 60 30 65 33 49 27 5 54 45 34 80 10 7 81 20 42 54 46 10 42 12 36 73 11 63 I 19 50 19 crimes against persons and property 21 51 35 61 37 9 44 51 24 41 18 59 21 21 64 69 54 66 68 32 47 59 15 55 86 56 60 79 71 26 75 39 36 74 82 83 63 38 35 decomposed in various ways 43 69 82 81 45 56 24 70 79 55 2 78 3 52 27 I 71 53 66 78 76 77 52 22 26 42 80 77 first attempt to delineate multivariate 52 14 82 32 66 59 63 33 79 41 33 12 58 81 86 9 85 40 68 44 17 13 57 31 69 37 25 58 73 47 4 68 9 relations among moral variables 47 64 67 38 18 11 62 83 18 44 1 30 32 70 76 75 40 16 57 67 28 77 23 76 31 62 6 28 14 2 7 8 72 48 5 14 I Voluminous data: 16 34 19 6 73 71 13 57 54 10 85 74 65 24 84 30 72 61 I 85,564 suicide records (1836–1860), Rank 1 - 9 10 - 19 20 - 28 72 Rank 1 - 9 10 - 19 20 - 28 86 Rank 1 - 9 10 - 19 20 - 28 60 29 - 38 39 - 47 48 - 57 29 - 38 39 - 47 48 - 57 29 - 38 39 - 47 48 - 57 58 - 66 67 - 76 77 - 86 58 - 66 67 - 76 77 - 86 58 - 66 67 - 76 77 - 86 classified by motive I 226,224 accused of personal crime I numbers, in a line → 1170 meters!

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 21/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 22/59 Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Works 1864: Statistique morale de l’Angleterre comparee...´ Comparing France and England Comparing France and England

Maps for: I crimes against persons, I crimes against property, I murder, I rape, I larceny by servants (vol domestique), I arson, I instruction, I suicide (only for France)

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 23/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 24/59

Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Works 1864: Statistique morale de l’Angleterre comparee...´ 1864: Statistique morale de l’Angleterre comparee...´ Statistique analytique Statistique analytique I Special symbols & annotations used to mark noteworthy patterns, circumstances (↑, ↓ shows increase/decrease) I Surrounding line graphs designed to decompose overall facts or relate to other facts

Figure: Detail from Plate I: of (L) Crimes, (R) Condemmed to death

Figure: Detail from Plate II: Increase and Decrease in Crime

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 25/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 26/59 Guerry’s Life and Work Guerry’s Works Guerry’s Life and Work Guerry’s Methods 1864: Statistique morale de l’Angleterre comparee...´ Methodological Questions Statistique analytique

I How to define valid and reliable social indicators? I Education I Reported levels of instruction (# male children in primary school) were suspect. I More uniform data from Ministry of War: exams for new recruits 7→ % who could read and write. I Crimes I Number of convictions (condamnes´ ) subject to factors that affect juries (severity of punishment, place where accused is judged) Crimes against persons by I Better to use number of indictments (accuses´ ): Indictment doesn’t necessarily month 7→ guilt, but it reasonably 7→ a crime was committed. (Detail from Plate II) I Migration and bias: Can one attribute crimes in a department to inhabitants (vs. strangers)? I Data from 1828+ 7→ 72% of accused either born or lived in each department I Only 3% committed by foreigners I Interpretive issues: I Shows some awareness of issues related to ecological fallacy and spurious correlation: e.g., literacy of actual prisoners

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 27/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 29/59

Guerry’s Life and Work Guerry’s Methods Guerry’s Life and Work Guerry’s Methods Graphical Comparisons Graphical Comparisons I Guerry worked before ideas of correlation, regression, scatterplots 7→ No evidence for relation between crime and literacy I 7→ Used direct comparison of pairs of maps or ranked lists I Corsica highest on crime, ∼ middle on literacy I Literacy lowest in west and central France, but crime varies considerably I “Clearly the relationship people talk about does not exist” (Guerry, 1833, p. 90)

(c) Crimes against (a) Literacy (b) Ranked lists persons Figure: Comparison of crimes against persons with literacy (% who can read and write) 7→ similar analyses for other variables (suicides, illigitimate births, ...)

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 30/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 31/59 Guerry’s Life and Work Guerry’s Methods Guerry’s Life and Work Guerry’s Methods 1833: Semi-graphic tables 1833: Semi-graphic tables How does type of crime vary with age? Crimes against persons I Indecent assault on adults (viol sur des adultes) decreases with age I 7→ Used ranked tables of crime/1000 connected by colored lines I Indecent assault on children increases with age (top for 70+) I First instance of modern parallel coordinates plot I Paricide rises to max at age 60–70

Figure: of crimes against persons at different ages Figure: Relative ranking of crimes at different ages

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 32/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 33/59

Guerry’s Life and Work Guerry’s Methods Guerry’s Life and Work Guerry’s Methods 1864: Statistique Analytique: Influence of Age Statistique Analytique: General Causes of Crime Plate XVII: M. Guerry’s Magnum Opus I Age distributions of criminals in I England vs. France Analysis of the factors associated with crimes and their geographic I 10 categories of crime broken down by subtype distribution I I Viol → rape, indecent exposure Rows: 23 crimes, ordered by I Vol → burglary, housebreaking, frequency and seriousness ... I keeping baudy house, bigamy, I Assassinat → murder, cattle stealing, ... manslaughter, shooting, I ... fraud, rape, murder stabbing, ... I Cols: Rank order of degree of criminality of English counties I Entries: Symbols for associated moral aspects I Population (% Irish, agricultural, domestics, ...) I Criminality (Male, young, ...) I Religion (Anglicans, “dissenters”, ...) I Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 34/59 MichaelCurves: Friendly (York positive University) and negative Andre-Michel´ Guerry JSM,Aug.2006 35/59 coincidences Multivariate Analyses: Data-centric Views Multivariate Analyses: Data-centric Views Guerry’s Challenge Graphical methods for multivariate data

I What can we learn from reanalysis of Guerry’s data? I : Can be enhanced to show statistical relations I Bivariate displays What could we do for Guerry as a consultant? more clearly and effectively Statistical historiography I Scatterplots with data (concentration) ellipses and smoothed (loess) curves I Scatterplot matrices I Understanding through reproduction I Reduced-rank displays: Multivariate visualization techniques can I What was he thinking? show the statistical data in simple ways, using dimension reduction techniques. I - show variables and observations in space accounting for greatest I Canonical discriminant plots - show variables and observations in space I How to visualize and understand relations among many variables? accounting for greatest between-group variation I How to relate these to geographic information? Reduced rank

Summary

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 37/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 38/59

Multivariate Analyses: Data-centric Views Bivariate displays Multivariate Analyses: Data-centric Views Bivariate displays Bivariate plots: Data ellipse and smoothing Bivariate plots: Region differences

Scatterplot with 40% and 68% Summary ellipses for each region 40000 data ellipses, and smoothed 40000 I Biggest differences (C, W) vs. Creuse (loess) curve Creuse Ardennes Ardennes (N, E) on literacy Indre Cote-d’Or I No linear relation between Indre Cote-d’Or I South low on both crime and 30000 crimes against persons and 30000 literacy Haute-Marne Jura literacy Haute-MarneJura Meuse CW N Meuse I No indication of non-linear E 20000 relation 20000 I Substantial number of unusual S departments Pop. per Crime against persons Pop. per Crime against persons Doubs 10000 Doubs 10000

Haut-Rhin Haut-Rhin Ariege Ariege

Corsica Corse 0 0

10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80 Percent Read & Write Percent Read & Write

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 40/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 41/59 Multivariate Analyses: Data-centric Views Bivariate displays Multivariate Analyses: Data-centric Views Reduced-rank displays Bivariate plots: Scatterplot matrices Multivariate analyses: Reduced rank displays

37014 I Crime vs. persons and property: %! Crime_pers I I Literacy & most others 2199 Multivariate visualization techniques can show the statistical data in 20235 I Suicide % out of wedlock simple ways, using dimension reduction techniques. I Biplots - show variables and departments in space accounting for greatest Crime_prop variance 1368 I Canonical discriminant plots - show variables and departments in space 74 accounting for greatest between-region variation Literacy I Show geographic location by color coding or other visual attributes. I Color code by region 12 163241 I Data ellipse to summarize regions I Suicides → Data-centric displays: The multivariate data is shown directly; geographic relations indirectly

3460 62486

Infants

2660

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 42/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 44/59

Multivariate Analyses: Data-centric Views Reduced-rank displays Multivariate Analyses: Data-centric Views Reduced-rank displays Biplots Biplots: Guerry data

1.2 I Dim 1 (34.5%): France Crime_pers I Biplots represent both variables (attributes) and observations obscure vs. France 0.8 Creuse (departments) in the same plot— a low-rank (2D) approximation to a data eclair´ ee´ Crime_prop Cote-d’Or matrix I ArdennesSarthe Dim 2 (20.8%): Personal Somme d Haute-Marne Ain 0.4 Saone-et-Loire ? MeurtheMeuseJura MancheMayenne Loire T T Maine-et-Loire Haute-Loire = crime vs. Donations Gers Y ≈ AB ak bk Literacy NordOiseGirondeVosgesE Cantal X Rhone Basses-PyreneesLot-et-GaronneC N Haute-GaronneIsereOrneHerault DordogneAude (benevolence?) Pas-de-CalaisSeine-et-Marne Charente k=1 Aisne Var IndreHautes-PyreneesPuy-de-Dome MarneEure-et-LoirHaute-SaoneYonneLoir-et-CherLoire-Inferieure 0.0 Bouches-du-RhoneDoubsAubeLoiretHautes-AlpesDrome S Allier Suicides Seine Vaucluse NievreWCher Basses-AlpesIndre-et-Loire Seine-Inferieure Tarn-et-Garonne Gard Seine-et-Oise Cotes-du-Nord I Haut-Rhin AveyronCorrezeAriege Variables usually represented by vectors from origin () Moselle Landes Lot Tarn Eure Lozere Ille-et-VilaineArdeche I Charente-Inferie Morbihan -0.4 Bas-Rhin Observations usually represented by points Pyrenees-OrientaDeux-SevresHaute-Vienne Dimension 2 (20.8%) VienneFinistere I Can show clusters of observations by data ellipses Infants

Calvados I Vendee Properties: -0.8 I Angles between vectors show correlations (r ≈ cos(θ)) Donations I ∼ Length of variable vectors % variance accounted for -1.2 I ≈ T yij ai bj : projection of observation on variable vector I Dimensions are uncorrelated overall (but not necessarily within group) -1.6 Corse

-1.2 -0.8 -0.4 0.0 0.4 0.8 1.2 1.6

Dimension 1 (35.4%)

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 45/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 46/59 Multivariate Analyses: Data-centric Views Reduced-rank displays Multivariate Analyses: Data-centric Views Reduced-rank displays Canonical discriminant plots Canonical discriminant plots: Guerry data, by Region

I Dim 1 (61.9%): France obscure vs. France I Project the variables into a low-rank (2D) space that maximally eclair´ ee´ discrimates among regions I Dim 2 (28.8%): Donations; I Visual summary of a MANOVA (S, E) vs. (N, C, W) I Canonical dimensions are linear combinations of the variables with maximum univariate F-statistics. I Vectors from the origin (grand mean) for the observed variables show the correlations with the canonical dimensions I Properties: I Canonical variates are uncorrelated I 2 − Circles of radius qχ2(1 α)/ni give confidence regions for group . I Variable vectors show how variables discriminate among groups I Lengths of variable vectors ∼ contribution to discrimination

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 47/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 48/59

Multivariate Mapping: Map-centric Views Multivariate Mapping: Map-centric Views Glyph maps Multivariate mapping: Map-centric displays Star maps

I Star map of Guerry Variables Bigger 7→ better Crime_pers Crime_prop I Variables ordered as in Literacy Suicides Donations Infants How to generalize choropleth maps to many variables? I 7→ identify unusual depts. I Star maps: Show multivariate data on the map using star icons, variable ∼ length of ray I Blended RGB displays: (V1, V2, V3) 7→ (R, G, B) shading I Conditioned choropleth maps: Stratify by two variables, show conditioned maps in a trellis-like display

Summary

Departments colored by Region

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 50/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 52/59 Multivariate Mapping: Map-centric Views Glyph maps Multivariate Mapping: Map-centric Views Blended color maps Star maps: by region RGB map

Median Ranks by Region Crime against persons, Crime against property, Literacy Crime_pers Crime_prop RGB map: R=Crime_pers, G=Crime_prop, B=Literacy Literacy Suicides 62 Donations Infants 59

80 8 76 2 60 57 50 14 27 51 55 75 54 67 78 61 77 22 29 28 10 88 35 53 52 72 68 56 45 89 70 41 21 44 49 25 37 58 18 I 39 N: Hi lit & personal crime 85 36 71 79 86 3 1 23 69 I 87 Creuse ( 23 ), Ain ( 1 ): Hi 42 17 16 63

19 38 pop/crime, low lit. 43 24 15 33 7 26 5 46 48 47 12 4 82 30 84 40 81 32 34 13 83

64 31 Crime_pers 65 11 9 66

200

Literacy Crime_prop

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 53/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 55/59

Multivariate Mapping: Map-centric Views Conditioned choropleth maps Summary Conditioned choropleth maps Summary Crime against property | % Literacy and Wealth (rank)

1Wealth 45 1Wealth 45

I Mdn: 8294 74 Mdn: 5521 Crime: good = blue; bad = n: 17 n: 31 red Guerry’s place in history I Control for two Literacy I (Along with Quetelet) Guerry’s work established the empirical study of background variables moral statistics 7→ “Moral statistics movement” 7→ modern criminology, I e.g., UR: High Lit, Wealth 37 sociology, social science. 7→ (N, High crime) I The 1833 Essai broke new ground in thematic cartography and data I Dynamic version: choose Literacy Pop/Crime 53 61 68 75 Pop/Crime 53 61 68 75 visualization. (00s) 81 89 95 (00s) 81 89 95 12 39 ranges with sliders 42Wealth 86 42Wealth 86

Mdn: 8326 74 Mdn: 7289 I n: 31 n: 16 The 1864 comparative study contemplated multivariate explanations beyond available theory and methods. I Guerry’s questions, methods and data still present challenges for Literacy multivariate spatial data visualization today. 37 Literacy

Pop/Crime 53 61 68 75 Pop/Crime 53 61 68 75 (00s) 81 89 95 (00s) 81 89 95 12 39

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 57/59 Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 58/59 Summary References I

Balbi, A. and Guerry, A.-M. Statistique comparee´ de l’etat´ de l’instruction et du nombre des crimes dans les divers arrondissements des academies´ et des cours royales de France. Jules Renouard, Paris, 1829. BL:Tab.597.b.(38); BNF: Ge C 9014 . Beirne, P. Inventing Criminology. State University of Albany Press, Albany, NY, 1993. Dupin, C. Carte figurativ de l’instruction populaire de la France. Jobard, 1826. BNF: Ge C 6588 (Funkhouser (1937, p. 300) incorrectly dates this as 1819). Guerry, A.-M. Essai Sur La Statistique Morale de la France. Crochard, Paris, 1833. English translation: Hugh P. Whitt and Victor W. Reinking, Lewiston, N.Y. : Edwin Mellen Press, 2002. Guerry, A.-M. Statistique morale de l’Angleterre comparee´ avec la statistique morale de la France, d’apres´ les comptes de l’administration de la justice criminelle en Angleterre et en France, etc. J.-B. Bailliere` et fils, Paris, 1864. BNF: GR FOL-N-319; SG D/4330; BL: Maps 32.e.34; SBB: Fe 8586; LC: 11005911.

Michael Friendly (York University) Andre-Michel´ Guerry JSM,Aug.2006 59/59