Bangor Business School Working Paper BBSWP/10/001 PUBLIC PERCEPTION AND POLICE DELIVERY By Y. Altunbas and S. P. Chakravarty Division of Financial Studies Bangor Business School and M. Y. Feilzer School of Social Sciences Bangor University February 2010 Bangor Business School Bangor University Hen Goleg College Road Bangor Gwynedd LL57 2DG United Kingdom Tel: +44 (0) 1248 382277 E-mail: [email protected]
26
Embed
Bangor Business School Working Paper · Keywords: British Crime Survey (BCS), Public confidence in police, Efficient policing, Measuring performance, Victims, Non -victims . 3 BBSWP/10/001
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Bang or Bus ines s School
Working Paper
BBSWP/10/001
PUBLIC PERCEPTION AND POLICE DELIVERY
By
Y. Al tunbas and S. P . Chakravarty
Divis ion of Financia l Studies
Bang or Bus ines s School
and
M. Y. Fe i lzer
School of Socia l Sciences
Bang or Univers i ty
February 2010
Bang or Bus ines s School
Bang or Univers i ty
Hen Goleg
Col l eg e Road
Bang or
Gw ynedd LL57 2DG
United King dom
Tel: +44 (0 ) 1248 38227 7
E-mai l : bus ines s @bang or.a c.uk
2
BBSWP/10 /001
Publ ic percept ion a nd pol i ce de l ivery
Abs tract
On e of th e d ifficu lt ies of m easu r in g p er for m an ce in th e d eliver y of p olice ser vices fr om
su r vey s is t h e p r oblem of r econ cilin g views abou t th e qu ality of policin g of d is t in ct
g r ou p s of r espon d en ts w ith d iffer en t p r ior ou t look abou t policin g . Th is p aper su g g es ts
a m eth od for addr ess in g th e p r oblem of th e d iffer en ces in th e d em og r ap h ic com posit ion
of vict im s an d n on -vict im s of cr im e to d r aw r obu st con clu s ion s . It appea r s th a t
per cep t ion s of vict im s of cr im e abou t th e r espon siven ess of th e police to r equ es ts for
h elp a r e m or e pess im is t ic th an th e p er cep t ion s of th ose wh o ex p r ess an op in ion abou t
th e ser vice th ey wou ld r eceive in th e h ypoth et ica l ca se of th em selves becom in g vict im s
of cr im e.
J EL cla ss ifica t ion : H11, H40, K40
Keywor d s: Br it ish Cr im e Su r vey (BCS), Pu blic con fiden ce in p olice, Efficien t policin g ,
Measu r in g p er for m an ce, Vict im s , Non -vict im s
3
BBSWP/10 /001
Publ ic percept ion a nd pol i ce de l ivery
1 . Introduct ion
Th is pap er look s a t a n eg lected a sp ect of th e Br it ish Cr im e Su r vey (BCS) da ta , t h e
cor r espon den ce between th e p er cep t ion s of th e police ser vice by th ose wh o h ave r eason
to com e in to con tact with th e police a s vict im s of cr im e an d th ose wh o h ave n ot h ad an y
su ch occasion for d ir ect con tact . A par t icu la r q u est ion th e an swer s to wh ich a r e
ex am in ed h er e con cer n s pu blic p er cep t ion abou t th e ability of th e police to r esp on d to
vict im s‘ appeals to tu r n u p a t th e scen e of cr im e. Th e ex p er ien ce of p r eviou s con tact
h as a s ig n ifican t im pact on th a t per cep t ion , a s ex p la in ed below.
Data r epor t ed in BCS m ay in for m th e views of som e of th ose wh o s tu dy th e qu ality of
policin g in Br ita in . A r eview of th e r eg ion al for ces cla im s th a t ―…var ia t ion in th e cr im e
r a t es in s im ila r a r eas m ig h t th er efor e be down to th e efficien cy or wor k in g p r act ices of,
for ex am ple, p olicin g in th ose a r eas‖ (Cr ou ch an d Mar d le, 2007). However , t h e r an k in g
of r eg ion al for ces th a t m ig h t be d er ived fr om BCS data h as to be appr oach ed with
cau t ion becau se th e da ta a r e n ot r an dom ised a t r eg ion al p olice for ce a r ea level. An
ex am in at ion of th e d iver g en ce in th e p u blic con fiden ce in th e Nor th Wales an d Dy fed
Powy s p olice for ces su g g est s th a t som e of th a t d iver g en ce m ay be du e to d iffer en ces in
th e g eog r ap h ica l d is t r ibu t ion (r u r a l-u r ban ) of su r vey r espon den t s between th e two
r eg ion s (Feilzer , 2009).
Th e sh or t com in g s of su r vey d a ta , especia lly abou t per cep t ion s con cer n in g th e ex ten t of
cr im in al beh aviou r in society or th e efficien cy with wh ich th e police r esp on d to cr im e,
a r e n u m er ou s . For ex am ple, a vis ible in cr ease in police n u m ber s m ay en g en d er a
g r ea ter sen se of secu r ity even if i t tu r n ed ou t to be th e ca se th a t th e p r even t ion of cr im e
is bet ter ach ieved by a sm aller n u m ber of bet t er -t r a in ed police officer s . Th e fea r of
cr im e a s m easu r ed in su r veys m ay con fla te ir r a t ion a l an x iet ies of r esp on den ts with th e
object ive r ea lity on th e g r ou n d . Th e fear of s t r an g er s m ig h t en ter in to p er cep t ion s of
th e r isk of becom in g a vict im of cr im e (Dit ton an d Fa r r a ll, 2007; Far r a ll et a l., 2000;
Lee, 2007). To m ak e m at ter s of in ter p r eta t ion of su r vey d a ta fu r th er com plica t ed , it
appea r s th a t th e fear of cr im e am on g st old er people ―does n ot a lway s t r an s la te in to
wor r ies of [th em selves ] bein g vict im ised‖ (Pan tazis , 2000). Th er e is a lso eviden ce th a t
you n g m ales wh o sp en d ―con sid er able pa r t of t h eir leisu r e t im e ou t s ide th e h om e [ar e]
m ost a t r isk fr om bein g vict im s of per son al cr im e‖, bu t th is is a dem og r aph ic g r ou p
4
BBSWP/10 /001
th a t adm it s to ―ver y lit t le fear ‖ of bein g in su ch dan g er (Bor ooah an d Car ach , 1997).
Th is is con sis t en t with th e fin d in g s in th e US as r ep or ted in th e Abs t r act of a
g over n m en t r ep or t t h a t vu ln er ability to cr im e ―d id n ot app ea r to be r ela ted to cr im e
r ed u ct ion effor t s‖ (Sk og an an d Max field 1981) . Th er e is a p r oblem in u s in g su r vey
da ta on per cep t ion s th a t g oes beyon d th e p ecu lia r it ies of BCS. At t itu d es su r vey s fa r too
oft en based on ―qu es t ion in g th e p u blic on su bject s wh ich a r e in su fficien t ly m ean in g fu l
to th em is lik ely to r esu lt , a t bes t , in r an dom an d , u su a lly , in in d u ced an swer s‖
(Rober t , 1978).
Desp it e th e in cr eas in g r ea lisa t ion th a t Br it ish Cr im e Su r vey da ta is n ot p ar t icu lar ly
r eliable in ju d g in g th e efficacy of p olicin g , it is a r g u ed h er e th a t som e r obu s t
con clu s ion s m ay s t ill be poss ible to der ive. Th is a r g u m en t is illu s t r a ted by an a lys in g
r espon ses to a pa r t icu lar qu es t ion th a t is a sk ed in th e above su r vey : ―Th ey (th e police
in th is a r ea) can be r elied on to be th er e wh en y ou n eed th em ‖
All th ose wh o ex pr ess a view in an swer s to th e above q u est ion h ave n ot n ecessar ily
com e in con tact with t h e police, an d m ay n ot h ave been vict im s of cr im e. Th e
dem og r ap h ic com posit ion of th ose th a t h ad been vict im s an d oth er s a r e n ot n ecessa r ily
th e sam e. For ex am ple, you n g m ales a r e m or e vu ln er able to ph y sica l a t t ack , sp en din g
a g r ea ter fr act ion of th eir t im e ou t of h om e. Th is par t icu la r is su e of d iffer en ces in th e
dem og r ap h ic com posit ion is add r essed h er e. A r obu st con clu s ion em er g es , t h a t t h e
n eg a t ive per cep t ion abou t th e police as r epor ted in an swer s to th e qu est ion above
can n ot be d ism issed s im ply as a con seq u en ce of dem og r ap h ic d iffer en ces in th e two
sam ples .
Th e p r oblem of d iffer en ces in th e dem og r aph ic com posit ion , especia lly s in ce
dem og r ap h ic va r iables in for m th e s t r a t ifica t ion in BCS, g ives r ise to th e t ech n ica l is su e
of in dex n u m ber s . A s im ilar ity in d ex , m easu r in g th e deg r ee of s im ila r ity or d is -
s im ila r ity between th ese two g r ou ps , is con sid er ed h er e.1 Ag g r eg at in g r espon ses fr om
differ en t g r ou p s in society in to a s in g le m easu r e of th e ‗pu blic r espon se‘ en ta ils
cr ea t in g an in dex th a t depen ds on th e com posit ion of th e r espon din g g r ou ps . Th e
in dex n u m ber p r oblem , ag g r eg at ion p r oblem as it som et im es ca lled , is n ot su fficien t ly
em p h asised in th e cr im in olog y lit er a tu r e. Th e p r oblem is flag g ed u p pr om in en t ly h er e
befor e delvin g in to th e su bstan ce of th e pa p er .
1 Th e in dex u sed h er e der ives fr om th e idea of s im ilar ity between ex por ts of d iffer en t cou n tr ies , an d it is
ex pla in ed in th e liter a tu r e on in ter n at ion al t r ade (F in g er an d Kr ein in 1979).
5
BBSWP/10 /001
Th e p aper is or g an ised a s follows. Sect ion 2 g ives a br ief su m m ar y of th e back g r ou n d
to th e p ar t icu lar BCS qu es t ion th a t is an a lysed h er e. Sect ion 3 r a ises th e p r oblem of
ag g r eg at in g r espon ses fr om d iffer en t in d ivid u als , g r ou ped in to ca teg or ies by r efer en ce
to ag e, eth n icity , g en d er etc, t o obta in an over a ll view of a t t it u de to th e police. Th e
con cept of a s im ilar ity in dex in th e d em og r aph ic com posit ion of r espon den t s —br ok en
down in to two m u tu ally ex clu s ive g r ou p s , t h ose th a t h ave n ever been vict im s of cr im e
an d th u s h ad n o occasion to seek h elp fr om th e p olice an d th ose th a t h ave a t lea s t on ce
been a vict im of cr im e an d m ay h ave su bsequ en t ly com e in con tact with th e police in
con sequ en ce—is in t r odu ced in Sect ion 4 . Th eir d iffer en t r esp on ses a r e com p ar ed an d
con t r a s t ed in Sect ion 5 , an d Sect ion 6 con clu d es .
2 . Bri t i s h crime s urvey
Th e Br it ish Cr im e Su r vey (BCS) is an an n u al su r vey based on a sam ple of a lm os t 50
th ou san d in ter views con du cted du r in g th e y ea r (Nich ola s et a l . 2007). Th e da ta set in
th is pap er r efer s to th e 2 006/2007 su r vey r esu lt s . BCS beg an a s p ar t of an a t t em pt to
u n d er s tan d th e n a tu r e of an d ch a llen g es to policin g in Br it a in , bu t it h a s g r adu ally
becom e a tool for m easu r es of p er for m an ce an d or g an isa t ion a l com peten ce of th e police
for ce in En g lan d an d Wales (Allen 2007; Hou g h et a l., 2007) in En g lan d an d Wales .
Scot lan d an d Nor th er n Ir elan d ar e n ot in clu ded in BCS. Th e u se of BCS d ata for th e
pu r poses of m easu r in g p olice per for m an ce an d com par in g th e or g an is a t ion a l
com peten ce of police au th or it ies r em ain s con t r over s ia l, a s ex p la in ed ear lier .
Th e sam ple is r an dom ised over En g lan d an d Wales an d th e r an dom isa t ion is based on
postcod es . Adu lt s between th e ag es of 16 -74 livin g in p r iva te h ou seh olds a r e sam pled .
Cr im e ag ain s t bu sin esses an d wh ite-colla r cr im e ar e n ot cover ed . Th e da ta r epor t ed in
BCS2006/07 con ta in s in t er views with 47 ,203 people between Apr il 2006 an d Mar ch
2007 . Th ese in t er views las t ed on aver ag e for 50 m in u tes , bu t lon g er t im e was d evoted
to vict im s th an n on -vict im s . Th e types of qu es t ion s th a t a r e a sk ed a t in t er view ar e
d iscu ssed in th e liter a tu r e (Nich ola s et a l ., 2007). Th e m eth od of con du ct in g in t er views
is a lso a m a t t er of pu blic r ecor d (Bollin g et a l., 2 007).
3 . Ag g reg at ion of res pons es
St r a t ifica t ion of th e p op u la t ion in th e ch oice of sam ples en ta ils t h e ch oice of r ele van t
p r ox ies by wh ich to id en t ify d iffer en ces in a t t itu de to th e p olice by d iffer en t g r ou p s of
people. Socio-econ om ic var iables , bu t on ly th r ou g h p r ox ies (g en d er , eth n icity an d ag e),
6
BBSWP/10 /001
ar e em ploy ed in BCS to iden t ify th e g r ou p s to su r vey . Ag g r eg at ion of an sw er s r epor t ed
in th e su r vey en ta il r esor t g r ou p weig h t s with in each sam ple pop u la t ion . In com par in g
r espon ses fr om d iffer en t sam ples in a su r vey —wh eth er between two police a r eas or
vict im s ver su s n on -vict im s of cr im e—differ en ces in th e com posit ion of th e sa m ples by
socio-econ om ic or d em og r ap h ic g r ou ps h ave to be tak en in to accou n t . It is r ea son able
to ask if t h e ag g r eg at ion pr ocedu r e u n accep tably d is tor ts t h e sam ple r esp on ses to
r en d er com p ar ison u n wise. We pr opose to ex am in e th is qu es t ion with ou t im m edia te
r esor t t o r eg r ess ion an a lys is . In s t ead , we follow th e pa th of ag g r eg a t ion an d in d ex
n u m ber s .
To illu s t r a te th e a r g u m en t , con sider two sam ples – Sam ple I an d Sam ple II – wh ich ar e
com posed of d iffer en t ty pes of people, for ex am ple in t er m s of ag e, eth n icit y , an d
g en der . Th e two sam ples m ay h ave seen th e police fr om d iffer en t van tag e poin t s , for
ex am ple a s vict im s of cr im e an d n on -vict im s of cr im e, r esp ect ively . Th e r espon den t s in
each sam ple a r e ask ed to ex pr ess a bin a r y a t t itu d e, favou r able an d u n favou r abl e, abou t
th e q u ality of ser vice th ey r eceive fr om th e police. Sam ple popu la t ion s com pr ise two
m u tu ally ex clu s ive an d ex h au s t ive ca t eg or ies , A an d B in each sam ple. Th e ca teg or ies
m ig h t be defin ed by com bin in g ag e, g en d er , eth n icity et c in to a set of m u tu a l ly
ex clu s ive an d ex h au st ive a t t r ibu tes . Su ppose fu r th er th a t t h e a t t it u d e towar d s th e
police va r ies n ot ju s t between th e two sam ples , vict im s an d n on -vict im s , bu t a lso
in dep en d en t ly between ca teg or ies of th e pop u la t ion . For ex am ple, Ca teg or y A
r espon den t s a r e m or e cr it ica l in Sam ple I th an in Sam ple II, bu t Ca teg or y A
r espon den t s a r e g en er a lly les s cr it ica l, ea s ier people to sa t is fy , t h an Ca teg or y B in both
th e sam ples . Categ or y B r esp on d en t s a r e les s cr it ica l in Sam ple I t h an in Sam ple II bu t
Categ or y B r espon d en t s a r e g en er a lly m or e cr it ica l in t em p er am en t , th an m em ber s of
Categ or y A in both th e sam ples . A n u m er ica l ex am ple is p r esen ted in pan els (a ) an d
(b) of Table 1 .
Let u s su ppose th a t we wish to k n ow if p eople r epr esen ted by Sam ple I a r e les s cr it ica l
of th e ser vice th ey r eceived th an p eople r ep r esen ted by Sam ple II. We ag g r eg a te th e
r espon se of th e two ca teg or ies u s in g th eir r ela t ive weig h t s in th e pop u la t ion of th e two
r esp ect ive a r eas , as r epor t ed in pan el (c) of Table 1 .
It m ig h t be in fer r ed th a t people r ep r esen ted in Sam ple I a r e les s cr it ica l t h an th ose
r ep r esen ted in Sam ple II. Now su ppose th a t Sam ple I com p r ise vict im s of cr im e an d
Sam ple II com pr ise th ose wh o h ad n ever th em selves been vict im s. Th en it wou ld be
tem pt in g to con clu d e th a t bein g a vict im of cr im e (wh ich m ig h t br in g th em in to
7
BBSWP/10 /001
con tact with th e police) m ak es p eople t ak e less cr it ica l of t h e ser vice th a t is p r ovided by
th e police for ce. Th e p u blic a t t it u de to th e police ser vice is less cr it ica l th an th e
m easu r es of a t t itu d e d er ived fr om in ter view with th ose th a t h ave n ot ex p er ien ced cr im e
a t fir s t h an d . Th is lin e of r eason in g wou ld be m is lead in g u n less th e r obu stn ess of th e
con clu s ion th a t Sam ple I is les s cr it ica l of t h e ser vice th an Sam ple II is t es ted with
r esp ect to th e ca teg or ica l com posit ion of th e two sam ples . Th is ag g r eg a t ion issu e is
illu s t r a t ed below.
It m ig h t be a r g u ed th a t t h e above r an k in g of pu blic views with th e ser vice between
Sam ple I an d Sam ple II is an u n r eliable in d ica tor of th e per ceived q u ality of policin g
becau se Sam ple II h as a g r ea ter p er cen tag e of Ca teg or y B in d ividu als wh o a r e
g en er a lly m or e cr it ica l. Su ppose we u se th e sam e ag g r eg at ion weig h ts , t h e weig h t s for
Sam ple I, in der ivin g th e ag g r eg a te r espon ses for both th e sam ples . Fr om th e r esu lt s
r ep or ted in p an el (d ) of Table 1 , it appear s th a t Sam ple II r espon den t s r a t e th e ser vice
bet t er t h an Sam ple I r espon den t s . Ran k in g of th e r espon se a s r ep or ted in pan el (c) a r e
r ever sed th r ou g h a d iffer en t ag g r eg at ion p r oced u r e.
For a r obu s t com par ison between th e r espon ses of Sam ple I an d Sam ple II popu la t ion s
to be m ade, t h e r an k in g of r espon ses between th e two sam ples h as to be in depen den t of
th e ag g r eg at ion weig h ts . Com p ar ison is a lso eas ier if th e d is t r ibu t ion of th e ca t eg or ies
of r espon d en ts is s im ilar between th e two sam ples .
4 . S imila ri ty index
Th e q u es t ion an a lysed h er e con cer n s th e per cep t ion of th e ability an d willin g n ess of
th e police to tu r n u p wh en r eq u ir ed . A view m ig h t be for m ed based on pr ox y va r iables ,
su ch a s th e vis ibility of police on th e s t r eet s an d in ciden t s h ig h lig h ted in th e p r ess . Th e
view m ay be r ein for ced or m odified wh en th er e is occas ion to ca ll th e p olice. Th e
com posit ion of r espon den t s to th e r elevan t BCS q u est ion n air e of th ose wh o h ad been
vict im s an d m ig h t h ave h ad con tact with th e police (Sam ple A) an d th ose th a t h ad
n ever been vict im s (Sam ple B) is n ot id en t ica l. A s im ila r ity in d ex for th e com posit ion of
th ese g r ou ps is su g g es ted h er e, followin g d iscu ss ion abou t in dex n u m ber s for
com par in g ex por t s im ila r ity (F in g er an d Kr ein in 1979).
Th er e a r e 42 police for ce a r eas (PFA) in En g lan d an d Wales . In each of th ese PFA
r eg ion s th er e a r e two sam ples , A an d B, of r espon den t s . Su ppose th a t th er e a r e 3
s t r a t ifica t ion cr it er ia for each sam ple of r espon den t s : ag e (You n g , Middle-ag ed , Old),
8
BBSWP/10 /001
eth n icity (Wh ite/Non -wh ite), an d g en d er Male/Fem ale). Th ese can be com bin ed to der ive
12 m u tu a lly ex clu s ive an d ex h au s t ive ca t eg or ies of r espon d en t s . Th e idea is to com p ar e
th e s im ila r ity between th e sam ples by r efer en ce to th e 12 m u tu a lly ex clu s ive an d
ex h au s t ive ca t eg or ies of th e r espon d en ts in each PFA an d in En g lan d an d Wales a s a
wh ole. Th e tech n ica l d eta ils a r e in th e app en dix below. Th e s im ilar ity in d ices a r e
tabu la t ed in Table 2 .
5 . Att i tude to pol i ce
Su r vey s abou t a t t it u des to th e p olice a r e n otor iou sly d ifficu lt to in t er pr et becau se
qu est ion s m ay be in ter p r eted d iffer en t ly by d iffer en t r espon den t s , even if th eir con tact
with th e police wer e for s im ila r r eason s , for ex am ple as vict im s of cr im e. Som e
r espon den t s wh en ask ed abou t th eir con fid en ce in th e police m ig h t base th eir an swer
on th e ou tcom e of th eir con tact in t er m s of h avin g th e p er p et r a tor s cau g h t . Oth er s m ay
base th eir an swer s on som eth in g less t an g ible lik e fa ir n ess th a t d ep en d s on h ow th ey
wer e t r ea t ed wh en th ey r ep or ted a cr im e. Th er e is eviden ce fr om oth er su r veys th a t
―posit ively r eceived en cou n ter s‖ h ave a posit ive im pact on th e pu blic p er cep t ion of
fa ir n ess in policin g (Br adfor d et a l. 2009). Sk og an (2006) h as a t t em pted to an a ly se
h ow th e above two d iffer en t factor s im p act on th e pu blic per cep t ion of th e effect iven ess
of th e p olice for ce. He con clu des th a t vict im s a r e oft en les s con cer n ed abou t h avin g th e
per p et r a tor s cau g h t or g oods r etu r n ed ―th an th ey ar e abou t h ow pr oper ly or
im pr op er ly th ey a r e t r ea ted by th e au th or it ies‖ (Sk og an , 2006). Th e r esp on den t s a r e
m or e con cer n ed abou t th e p r ocess th an th e ou tcom e. We ex am in e h er e on e of th e BCS
qu est ion s focu sin g on th e p r ocess to see if vict im s of cr im e h ave g r ea ter con fid en ce
th an th ose wh o h ave n ot been vict im s abou t th e effect iven ess of th e police in h an dlin g
cr im e.
Effect ive policin g en ta ils th e police tu r n in g u p wh en n eed ed . On e of th e qu est ion s
ask ed in BCS is con cer n ed with th is a spect of policin g . T h e qu est ion in th e BCS th a t is
ex am in ed h er e is d escr ibed below.
Var iable = pola t t1
Qu es t ion : Th e police in th is a r ea can be r elied on to be th er e wh en you n eed
An swer s to th e q u est ion ar e g iven n u m er ic va lu es a s ex p la in ed below.
Valu e = 1 Label = St r on g ly ag r ee
9
BBSWP/10 /001
Valu e = 2 Label = Ten d to ag r ee
Valu e = 3 Label = Neith er ag r ee n or d isag r ee
Valu e = 4 Label = Ten d to d isag r ee
Valu e = 5 Label = St r on g ly d isag r ee
Valu e = 8 Label = Refu sed
Valu e = 9 Label = Don 't k n ow
Th e above qu est ion is a sk ed to g r ou p s of r espon den ts A (vict im s o f cr im e) an d B (n on -
vict im s) an d th e cu m u la t ive per cen tag es of r espon den t s wh o eith er h old a favou r able
op in ion (St r on g ly ag r ee p lu s Ten d to ag r ee) or h old a n on -n eg at ive op in ion (favou r able
p lu s Neith er ag r ee n or d isag r ee) a r e r ecor d ed in Table 3 . Wh a t is r em ar k able in Table 3
is t h e level of con fiden ce in th e p olice bein g able an d willin g to tu r n u p wh en n eed ed
th a t th e vict im s of cr im e h ave is lower in vir tu a lly ever y PFA th an th ose wh o h ave n ot
been vict im s of cr im e an d h ad pr esu m ably less n eed to seek h elp fr om th e police.
To ex am in e if t h is appa r en t ly sy s t em a t ic d iver g en ce of views between sam ples A an d B
can be a t t r ibu ted to d iffer en ces in th e com posit ion of th e two g r ou p s of r espon den t s , we
h ave ex am in ed th e cor r ela t ion between th e Sim ilar ity In dex in Table 2 separ a tely with
each set of va lu es of A-B in Table 3 over th e ser ies span n in g a ll th e 42 p olice for ce
a r eas . Th e cor r ela t ion coefficien t is -0 .16 between FAV (A-B) an d Sim ilar ity In d ex . Th e
cor r espon din g coefficien t for Non -n eg lig ible (A-B) an d Sim ila r ity In d ex is -0 .22 , an d
th u s th e per s is ten ce of lower op in ion of th e p olice a s descr ibed above h eld by th e
vict im s of cr im e can n ot be d ism issed r ead ily a s an ag g r eg at ion p r oblem of th e da ta .
Wh ils t t h e cor r ela t ion coefficien t s above a r e sm all, t h ey do n ot r u le ou t t h e poss ibility
th a t over -r ep r esen ta t ion of ca teg or ies (Appen dix Table 1 ) of in d ivid u als wh o m ig h t
g en er a lly h old a lower op in ion of th e police in Sam ple A cou ld be th e ex p lan a t ion of th e
r esu lt s r epor ted in Table 3 . Wh ils t th e sam ple s izes in in d ividu al police for ce
au th or it ies a r e too sm all to war r an t r an k in g au th or it ies in Table 3 , t h e p r oposit ion
abou t th e im pact of th e ag g r eg a t ion p r oced u r e can be t es ted u s in g th e la r g er d a ta set in
PFA 43 , th e wh ole of En g lan d an d Wales .
To p r oceed with th is lin e of a r g u m en t , we ex am in e th e BCS da ta on ―pola t t1‖ by th e 12
m u tu ally in clu s ive an d ex h au st ive ca teg or ies of r esp on d en ts in PFA 43 . At t it u des of
th e two sam ples , A an d B, a r e r ecor d ed in Tables 4 an d 5 , an d com par ed in Table 6 . Th e
com par ison elabor a tes p r esen ts th e sam e in for m at ion Table 3 , an d th e con clu s ion is th e
sam e. Th e vict im sam ple is m or e scep t ica l of th e police th an th e oth er sam ple.
10
BBSWP/10 /001
Com bin in g r esu lt s in Tables 4 an d 5 , we can com par e th e r espon ses of th e two sam ples ,
A an d B. As we can see fr om Table 5 , an d a lso fr om th e la s t r ow in Table 3 , t h e
per cen tag e of r espon den ts r espon din g eith er favou r ably or n on -n eg a t ively is lower for
A th an it is for B. Su pp ose it is th e ca se th a t th e d is t r ibu t ion of g r ou p s com pr is in g
Sam ple A is s u ch th a t t h er e is a bia s ag a in s t t h e police. To r em ove th a t p oss ibility
in flu en cin g th e d iffer en ce in r espon ses by A an d B, we s im u la te a d is t r ibu t ion for
Sam ple A th a t follows th e d is t r ibu t ion of g r ou p s in Sam ple B. Th en we com par e
r espon ses . Th e p r oced u r e is a s follows. Fir s t con sider th e d is t r ibu t ion of r espon den t s in
PFA 43 between th e 12 ca teg or ies for each of th e sam ples A (Vict im s) an d B (Non -
vict im s). Now s im u la te a d iffer en t d is t r ibu t ion by ch an g in g th e n u m ber s of
r espon den t s by g r ou p s 1 to 12 for Sam ple A to r eflect th e d is t r ibu t ion in Sam ple B. We
n ote th a t Favou r able or Non -n eg a t ive r espon ses fr om A will ch an g e between th e
r ecor d ed va lu es an d va lu es ca lcu la t ed u s in g th e s im u la ted d is t r ibu t ion of th e 12
m u tu a lly ex clu s ive an d ex h au s t ive g r ou p s ca t eg or is in g th e r espon den t s . Th u s Table 4
n ow ch an g es as sh own below in Table 9 . Th e a t t it u des of th e two sam ples in th is
s im u la ted case a r e com p ar ed in Table 9 an d su m m ar ised in pan el (a ) of Table 10 . Th e
m ain fin d in g , th a t r esp on den t s belon g in g to Sam ple A a r e m or e scep t ica l of r eceivin g
h elp fr om th e police th an th ose belon g in g to Sam ple B, h old s for both Table 6 , wh ich
was based on th e or ig in a l da ta an d pan el (a ) of Table 10 , based on th e s im u la t in g th e
d is t r ibu t ion of sam ples between ca teg or ies or r espon den t s in Sam ple A to fit th ose of
Sam ple B in th e or ig in a l da ta . A t es t of r obu stn ess of th e com par ison between A an d B
a lso en ta ils a com par ison based on a fu r th er s im u la t ion , d is t r ibu t in g th e ca t eg or ies in
Sam ple B to follow th e or ig in a l d is t r ibu t ion of ca teg or ies of r esp on d en ts in Sam ple A.
Th e ou tcom e is r epor t ed in pan el (b) of Table 10 .2
A com par ison of Tables 6 an d 10 add s r obu stn ess to ou r con cer n th a t t h e vict im s of
cr im e m ay be less con fiden t abou t bein g able to ca ll u pon th e police wh en n eed ed th an
th ose th a t h ave n ever been vict im s.
6 . Conclus ion s
Th er e a r e p r oblem s of in ter pr et in g BCS da ta , esp ecia lly s in ce it is collected for a ver y
wide r an g e p u r poses with ou t focu s on th e r elevan t s t r a t ifica t ion cr it er ia to be u sed in
sam plin g . Desp it e r eser va t ion s abou t th e q u ality of th e BCS d ata , som e con clu s i on s can
be d er ived , if on ly a s a p oin ter for fu r th er in ves t ig a t ion , fr om th is da ta . It appear s th a t
2 To con s t r u ct pan el (b) of Table 10 , we follow s teps s im ilar to th ose ou t lin ed in g oin g fr om Table 6 to pan el
(a) of Table 10 , wh er e Sam ple A was s im u la ted to r esem ble th e d is t r ibu t ion of ca teg or ies as in Sam ple B. We
n ow h old d is t r ibu t ion of A a t it s or ig in a l va lu es an d s im u la te Sam ple B to r esem ble th e d is tr ibu t ion of
ca teg or ies in Sam ple A.
11
BBSWP/10 /001
r espon den t s wh o h ave been vict im s of cr im e a r e m or e scep t ica l abou t th e ability of th e
police to tu r n u p wh en r equ ir ed th an th ose wh o h ave n ever been vict im s. It is
dem on st r a ted h er e th a t t h is con clu s ion can n ot be d ism issed a d a ta ag g r eg at ion
pr oblem du e to th e dem og r aph ic com posit ion of th e sam ples for vict im s an d n on -
vict im s bein g d iffer en t fr om each oth er . Th e pap er su g g est s a way of t es t in g for
r obu stn ess ag a in s t th e above p r oblem of d em og r aph ic com p osit ion , bu t a n ote of
cau t ion is n eeded . For ex am ple, it is n ot clea r th a t police for ce a r eas r an dom ise th e da ta
an d h en ce com par ison between d iffer en t for ce a r eas u s in g th is da ta is n ot pa r t icu la r ly
r eliable.
Wh a t th e r esu lt s d em on s t r a t e is a n eed to in vest ig a te, with sp ecia l focu s on sam ple
select ion , wh y cer ta in g r ou p s of p eople, especia lly if t h ey h ave been vict im s of cr im e,
h old d iffer en t views fr om oth er s abou t th e abilit y of th e police to d ea l p r om pt ly with
th eir n eed s . Th e d iffer en ces a r e n ot ex p la in ed p u r ely by th e d iffer en ces in th e
s t r a t ifica t ion var iables in th e sam ple com posit ion . Th e h ypoth esis th a t p er son al con tact
with th e police h as a n eg at ive im p act on a t t itu d e to th e p olice, as r epor t ed in S k og an
(2006), is too br oad a view to ex p la in ou r obser va t ion . Ou r d a ta does n ot d iffer en t ia t e
between th ose vict im s wh o h ad con tact s an d th ose th a t d id n ot . In an y even t , Br adfor d ,
J ack son an d Stan k o (20 09) r epor t t h a t con tact s , per se, do n ot n ecessar ily lead to
n eg a t ive a t t it u de to th e police.
12
BBSWP/10 /001
References
Allen , J ., 2007 . Su r vey a ssessm en ts of police p er for m an ce in th e Br it ish Cr im e Su r vey ,
in : M. Hou g h an d M. Max field , eds ., Su r vey in g Cr im e in th e 21s t Cen tu r y (Cu llom pton :
Willan ) pp …
Bollin g , K ., C. Gr an t an d P. Sin cla ir , 2007 . 2006 -07 Br it ish Cr im e Su r vey (En g lan d an d
Wales): Tech n ica l R epor t vol 1 (Lon don : Hom e Office0 .
Bor ooah , V.K. an d C.A. Ca r cach , 1997 , Cr im e an d fear , evid en ce fr om Au st r a lia , Br it ish
J ou r n a l of Cr im in olog y 37 , 635 -656 .
Br ad for d , B, J ack son , J ., an d E . Stan k o, 2009 , Con tact an d con fiden ce: r evis it in g th e
im pact of p u blic en cou n ter s with th e police, Policin g an d Society 19 :1 , 20 -46 .
Cr ou ch G. an d S. Mar d le, 2007 . Police an d cr im e s tan da r d s d ir ector a te r eview of
per for m an ce ben ch m ar k in g , Dr aft Rep or t Ver s ion : 1 .0 , Police an d Cr im e Stan d ar d s
Dir ector a te 14 /6 /07 .
J . Dit ton , J an d S. Fa r r a ll, 2007 . Th e Br it ish cr im e su r vey an d fea r of cr im e, in : M.
Hou g h an d M. Max field , ed s ., Su r vey in g Cr im e in th e 21 st Cen tu r y (Cu llom pton :
Willan ) pp …..
Fa r r a ll, S., Ban n is ter , J ., Dit ton , J ., an d E . Gilch r is t , 2000 . Socia l p sy ch olog y an d th e
fea r of cr im e. Br it ish J ou r n a l of Cr im in olog y 40 , 399-413 .
Feilzer , M (2009), ―Not F it for Pu r p ose! Th e (Ab -)u se of th e Br it ish Cr im e Su r vey a s a
Per for m an ce Measu r e for In d ividu al Police For ces‖, Policin g ; doi: 0 .1093/police/pap009
Fin g er , J .M. an d M.E . Kr ein in , 1979 . A Measu r e of ‗ex por t s im ilar ity ‘ an d it s poss ible
u ses . Th e Econ om ic J ou r n a l 89 , 905-912
Hou g h , M., Max field , M., Mor r is , B., an d J . Sim m on s , 2007 . Th e Br it ish cr im e s u r vey
over 25 year s : p r og r ess , p r oblem s, an d p r ospect s , in : M. Hou g h an d M. Max field , ed s .,
Su r vey in g Cr im e in th e 21st Cen tu r y (Cu llom p ton : Willan ), pp…..
Nich ola s , S., Ker sh aw, C., an d A. Walk er , 2007 . Cr im e in En g lan d an d Wales 2006/07
(Lon don : Hom e Office).
Lee, M., 2007 . In ven t in g fea r of cr im e: cr im in olog y an d th e p olit ics of an x iety
(Cu llom pton : Willan ).
Pan tazis , C., 2000 , Fea r of cr im e, vu ln er ability an d pover ty : eviden ce fr om th e Br it ish
cr im e su r vey , Br it ish J ou r n a l of Cr im in olog y 40 , 414 -436 .
Sk og an , W. G an d M. G. Man sfield , 1981 . Copin g with cr im e: in d ividu al an d
n eig h bou r h ood r eact ion s (Th ou san d Oak s , CA: Sag e). Abst r act ava ilable a t :
h t tp ://www.n cjr s .g ov/Ap p/Pu blica t ion s/abs t r act .a spx ?ID=78899 )
Sk og an , W. G., 2006 . Asym m et r y in th e im pact of en cou n ter s with th e police . Policin g
an d Society 16 , 99-126 .
Rober t , P ., 1978 . Pu blic Opin ion in Rela t ion to Cr im e an d Cr im in al J u s t ice. Pap er
p r esen ted a t t h e 13th
Cr im in olog ica l Con fer en ce, St r a sbou r g .