2-

MinSummar
y
measurement,
theprocessofassoci
ati
ngnumber
swi
thphy
sical
quant
it
iesandphenomena.

Measur ementi sfundament altothesci ences;toengineering,

constructi
on,andot hertechnicalfields;andtoalmostal lever
yday
acti
v i
ti
es.Forthatreasont heelement s,condit
ions,
li
mi tati
ons,and
theoreti
calfoundationsofmeasur ementhav ebeenmuchst udied.
Seeal someasur ementsy stem foracompar i
sonofdifferent
systemsandt hehistor
yoft heirdev elopment.

Measur ementsmaybemadebyunai dedhumansenses, inwhich

caset heyareoftencal l
edestimat es,or,morecommonl y,bythe
useofi nstr
ument s,whichmayr angei ncompl exit
yfrom si mple
rul
esf ormeasur i
ngl engthstohighlysophi st
icatedsy stems
designedt odetectandmeasur equant it
iesentirel
ybey ondt he
capabili
ti
esofthesenses, suchasr adi
owav esf rom adi stantst
ar
orthemagnet icmomentofasubat omi cparti
cle.
(Seeinstrumentation.)

Measur ementbegi nswi thadef ini

tionoft hequant i
tyt hati
stobe
measur ed,andi talwaysi nvolvesacompar isonwi thsomeknown
quantit
yoft hesameki nd.I ftheobj ectorquant it
yt obemeasur ed
i
snotaccessi blef ordirectcompar i
son,itisconv ertedor
“t
ransduced”i ntoananal ogousmeasur ementsi gnal .Since
measur emental way sinv ol
vessomei nteracti
onbet weent he
objectandt heobser verorobser vinginstrument ,therei salway
san
exchangeofener gy ,
whi ch,althoughi nev erydayappl icati
onsis
negli
gible,
canbecomeconsi der ableinsomet ypesof
measur ementandt herebyl imitaccur acy.
Measur ementi nstrument sandsy stems

Ingener
al,measuri
ngsystemscompr i
seanumberoff unct i
onal
element
s.Oneelementisrequi
redtodiscriminatetheobjectand
senseit
sdimensionsorfr
equency.Thi
si nformationi
sthen
transmi
tt
edthroughoutt
hesystem byphy sical
signal
s.Ifthe
objectisitselfactive,suchaswat erf l
ow, i
tmaypowert hesignal
;
i
fpassi ve,itmustt riggerthesignal byinteracti
oneitherwithan
energeticpr obe,suchasal ightsour ceorX- r
aytube, orwitha
carri
ersignal .Eventuall
yt hephysical si
gnal i
scompar edwitha
refer
encesi gnalofknownquant itythathasbeensubdi v
idedor
mul t
ipli
edt osuitther angeofmeasur ementr equir
ed.The
refer
encesi gnalisder i
vedfrom object sofknownquant it
ybya
processcal ledcalibrati
on.Thecompar isonmaybe
ananal ogpr ocessi nwhi chsignalsinacont i
nuousdi mensionare
broughtt oequal i
ty.Anal ter
nativecompar i
sonpr ocess
i
squant i
zationbycount ing,i
.e.
,divi
di ngthesignalintopartsof
equal andknownsi zeandaddi ngupt henumberofpar t
s.

Otherf uncti
onsofmeasur ementsystemsf aci
li
tatet hebasic
processdescr i
bedabov e.Ampli
fi
cationensur est hatt hephy si
cal
signalisstrongenought ocompletethemeasur ement .I
nor derto
reducedegr adat i
onoft hemeasurementasi tprogr essesthrough
thesy stem, thesi gnalmaybeconv ertedtocodedordi git
alform.
Magni ficat
ion, enlargi
ngthemeasur ementsi gnal wi t
hout
i
ncr easingitspower ,i
sof t
ennecessar ytomat cht heout putofone
elementoft hesy stem withthei
nputofanot her,suchasmat chi
ng
thesiz eofther eadoutmet erwi
ththedi scerni
ngpowerof
thehumaney e.

Onei mportantty
peofmeasur ementi stheanalysisofresonance,
orthefrequencyofv ari
ationwi thi
naphy si
calsystem.Thisis
determinedbyhar moni canal ysis,
commonl yexhibit
edinthe
sorti
ngofsi gnal
sbyar adi oreceiver.Comput ati
oni sanother
i
mpor tantmeasur ementpr ocess,inwhi chmeasur ementsignals
aremani pulat
edmat hemat icall
y,t
y picall
ybysomef orm ofanalog
ordigit
alcomput er.Comput ersmayal soprovi
deacont rol
functi
oninmoni toringsy stem performance.

Measur ingsy stemsmayal soi ncludedev i

cesf ortr
ansmi t
ting
si
gnal sov ergreatdistances( seet elemet r
y).Allmeasur i
ng
systems, ev enhighlyautomat edones, incl
udesomemet hodof
di
spl ayi
ngt hesi gnaltoanobser ver.Visualdisplaysystemsmay
compr iseacal i
br at
edchar tandapoi nter,
ani ntegr
ateddisplayon
acat hode-raytube, oradigitalreadout .Measur ementsy stems
oftenincludeel ement sforrecording.Acommont y
peut i
li
zesa
writi
ngst ylusthatrecordsmeasur ement sonamov i
ngchar t.
El
ect
ri
calr
ecorder
smayi
ncl
udef
eedbackr
eadi
ngdev
icesf
or
gr
eat
eraccur
acy.

Theact ualperfor manceofmeasur i

nginstrumentsi saffectedby
numerousex ternalandi nt
ernal f
actors.Amongext ernalfactors
arenoiseandi nterf
erence,bothofwhi chtendtomaskordi stor
t
themeasur ementsi gnal.I
nternal f
actorsincl
udelinearity,
resol
ution,preci si
on,andaccur acy ,
allofwhicharechar acterist
ic
ofagiveni nstrumentorsy st
em, anddy namicresponse, dri
ft,
andhyst er
esis, whichareef f
ect sproducedi nthepr ocessof
measur ementi tself.Thegener alquestionoferrorinmeasur ement
rai
sesthet opicofmeasur ementt heory.
Measur ementt heory

Measur ementt heoryisthest udyofhownumber sar eassignedto

objectsandphenomena, andi t
sconcernsincl
udet heki ndsof
thi
ngst hatcanbemeasur ed,howdiff
erentmeasur esrelat
eto
eachot her,
andt hepr obl
em ofer r
orinthemeasur ementpr ocess.
Anygener altheoryofmeasur ementmustcomet ogr i
pswiththree
basicproblems: err
or;representat
ion,whichist
hej usti
fi
cati
onof
numberassi gnment ;anduni queness,whichist
hedegr eetowhi ch
thekindofr epresentati
onchosenappr oachesbei ngt heonlyone
possiblefortheobjectorphenomenoni nquestion.

Varioussy stemsofax ioms, orbasi cr ul esandassumpt i

ons, have
beenf or mul atedasabasi sformeasur ementt heor y.Someoft he
mosti mpor tantt ypesofaxi omsi ncludeaxi omsofor der ,axioms
ofex t
ensi on, axiomsofdi fference, axi omsofconj ointness, and
axiomsofgeomet ry.Ax iomsofor derensur ethatt heor der
i
mposedonobj ect sbyt heassi gnmentofnumber si st hesame
orderat tai
nedi nact ual obser vati
onormeasur ement .Axiomsof
extensiondeal wi t
ht her epresent ationofsuchat t
ributesast ime
durati
on, l
engt h, andmass, whichcanbecombi ned, or
concat enat ed, f
ormul ti
pl eobject sexhi biti
ngt heat tri
butei n
question.Axi omsofdi ffer encegov er nt hemeasur ingofi nter v
als.
Axiomsofconj ointnesspost ulatet hatat tr
ibutest hatcannotbe
measur edempi ricall
y( forexampl e, l
oudness, intelli
gence, or
hunger )canbemeasur edbyobser vingt hewayt hei rcomponent
dimensi onschangei nrelat i
ont oeachot her.Axi omsofgeomet r
y
gov er
nt her epresent ati
onofdi mensi onal l
ycompl exat tr
ibutesby
pairsofnumber s,tri
plesofnumber s, orev enn- t
upl esofnumber s.
Theproblem ofer r
ori soneoft hecent ralconcer nsof
measurementt heor y
.Atonet imei twasbel i
ev edt haterrorsof
measurementcoul dev ent uall
ybeel i
mi natedthr ought he
ref
inementofsci enti
ficpr inciplesandequi pment .Thisbeliefisno
l
ongerheldbymostsci entist
s, andalmostal lphy sical
measurement sr epor tedt odayar eaccompani edbysome
i
ndicati
onoft hel imitationofaccur acyort hepr obabl edegreeof
err
or.Amongt hev arioust ypesofer rort hatmustbet akeni nt
o
accountareerr orsofobser vati
on( whi chincludei nstrument al
err
ors,personal errors, systemat icerrors,andr andom er r
ors),
err
orsofsampl ing,anddi rectandi ndirecterror s(inwhi ch
oneerroneousmeasur ementi susedi ncomput i
ngot her
measurement s) .

Measur ementt heorydat esbackt ot he4t hcentur yBC, whena

theor yofmagni tudesdev elopedbyt heGr eek
mat hemat i
ciansEudoxusofCni dusandThaeat et uswasi ncluded
i
nEucl id’sElement s.Thef ir
stsy stemat i
cwor konobser v
ational
errorwaspr oducedbyt heEngl i
shmat hemat i
cianThomas
Simpsoni n1757, butt hef undament alwor koner rortheorywas
donebyt wo18t h-
centuryFr enchast ronomer s,Joseph- Louis
Lagr angeandPi er
re-SimonLapl ace.Thef ir
stat temptt o
i
ncor por atemeasur ementt heoryint othesoci alsci encesal so
occur r
edi nthe18t hcent ur y,whenJer emyBent ham, aBr i
ti
sh
uti
litarianmor al
ist
,attempt edtocr eateat heor yf orthe
measur ementofv alue.Moder naxiomat i
ctheor i
esof
measur ementder i
v efrom t hewor koft woGer man
scient i
st s,Hermannv onHel mhol t
zandOt t
oHöl der ,and
cont empor aryworkont heappl i
cat i
onofmeasur ementt heoryto
psy chol ogyandeconomi csder i
vesi nlargepar tfrom t hewor k
ofOskarMor gensternandJohnv onNeumann.

Sincemostsoci alt
heor i
esar especul
ati
veinnat ure,attempt st
o
establi
shstandardmeasur i
ngsequencesort echniquesf orthem
havemetwi t
hlimit
edsuccess.Someoft hepr oblemsi nvolvedin
socialmeasurementi ncludethelackofuniversall
yaccept ed
theoret
icalf
ramewor ksandt husofquantifi
ablemeasur es,
sampl i
ngerrors,pr
obl emsassoci atedwit
ht heintrusionoft he
measur erontheobj ectbei ngmeasured,andt hesubj ecti
venature
oftheinformationrecei vedfrom humansubj ects.Economi csis
probablythesocialsci encethathashadt hemostsuccessi n
adopti
ngmeasurementt heori
es,pri
mari
lybecausemany
economicvar
iables(l
ikepri
ceandquant i
ty)canbemeasured
easi
lyandobj
ectivel
y.Demogr aphyhassuccessful
l
yemployed
measurementtechni
quesaswel l
,par
ti
cularl
yint
heareaof
mortal
it
ytabl
es..