Sampl

e:-
Wheny ouconductr
esearchaboutagr oupofpeople,
it’
srar
elypossibl
et o
coll
ectdat
afrom ev
erypersoninthatgroup.I
nstead,
youselectasampl e.
Thesampleisthegroupofindi
vidual
swhowi llact
ual
lypart
ici
pateinthe
resear
ch.

Todrawv al
idconcl
usi
onsfrom yourresult
s,y
ouhavetocaref
ull
ydeci
de
howy ouwill
sel
ectasamplethatisrepresent
ati
veoft
hegroupasawhole.
Ther
ear etwotypesofsampli
ngmet hods:

Pr
obabil
i
t ysampl
ing:-invol
vesrandom sel
ect
ion,al
l
owi
ngy
out
omake
st
rongstati
sti
cal
inferencesaboutthewholegroup.

Non-
probabil
i
tysampling:-i
nvol
vesnon-
random sel
ect
ionbasedon
conv
enienceorothercri
teri
a,al
l
owingyoutoeasil
ycol
lectdat
a.

Popul
ati
on:
-
Apopul ati
onisadisti
nctgroupofindi
viduals,
whethert
hatgroup
compr i
sesanat i
onoragr oupofpeoplewi t
hacommonchar act
erist
ic.I
n
stat
isti
cs,apopulat
ionisthepoolofindivi
dualsf
rom whichastati
sti
cal
sampl eisdrawnforastudy.Thus,
anysel ecti
onofindi
vi
dualsgrouped
togetherbyacommonf eatur
ecanbesai dt obeapopulati
on.

Asampl eisast at
isti
call
ysignif
icantporti
onofapopulati
on,not
anent i
repopul
at i
on.Forthisreason,astati
sti
cal
analysi
sofa
sampl emustreporttheappr oxi
mat estandar
ddev i
ati
on,or
standarderr
or,ofitsresul
tsfrom theentir
epopulat
ion.Onl
yan
analysi
sofanent ir
epopul at
ionwoul dhavenostandarderror
.
I
nmostuses, thewor dpopulationi
mpliesagr oupofpeopl eorat
l
eastagr oupoflivingbeings.Howev er
,stati
sti
ciansreferto
whatevergroupt heyarestudy i
ngasapopul ati
on.Thepopul ati
on
ofastudymi ghtbebabi esbor ninNort
hAmer i
cai n2021,the
tot
alnumberoft echst art
upsi nAsiasi
ncet heyear2000, the
aver
agehei ghtofal l
accountingexaminationcandidates,orthe
meanwei ghtofU. S.taxpayer
s.

Stat
isti
ciansandresearcher
sprefert
oknowt hecharacter
isti
csof
ever
yent it
yinapopul at
ioni
nordertodrawthemostpr ecise
conclusi
onspossible.Thisi
simpossibl
eorimpract
icalmostof
theti
me, however
, si
ncepopulat
ionsetstendtobequitelarge.

Forexample,ifacompanywant edtoknowwhet hermostofits

50,000customer sweresatisfi
edwiththecompany 'sservi
celast
year,i
twouldbei mpracti
cal t
ocalleverycli
entont hephonet o
conductasur vey.Asampl eofthepopul at
ionmustbet akensince
thecharacter
isti
csofeveryindivi
dualinapopul ati
oncannotbe
measur edduet oconstr
aintsofti
me, resources,andaccessibil
it
y.

Lar
gesampl
e:-

Thesampl esizenisgreat
ert han30(n≥30)i
tisknownasl ar
ge
sample.Forlargesamplesthesamplingdist
ribut
ionsofst
ati
sti
c
arenormal(Ztest)
.Astudyofsampl i
ngdistr
ibuti
onofstat
ist
icf
or
l
argesampl eisknownasl argesampletheory.
Smal
lsampl
e:-

Ift
hesampl esizenil
slessthan30(n<30),i
tisknownassmall
sample.Forsmallsamplesthesampli
ngdistri
buti
onsar
et,Fand
χ2distr
ibut
ion.Astudyofsampli
ngdistr
ibuti
onsforsmal
l
samplesisknownassmal lsamplet
heor y.

Nul
lHy
pot
hesi
s(Ho)
:

Ho:Thegi
vensamplewoul
dhav
ecomef
rom apopul
ati
onwi
th
speci
fi
edpr
oport
ionP=Po

Al
ter
nat
iveHy
pot
hesi
s(H1)

H1: Thegiv
ensampl
emaynotbef
rom apopul
ati
onwi
th
specif
iedpr
opor
ti
on

P≠Po(
TwoSi
ded)

P>Po(
Onesi
ded-
ri
ghtsi
ded)

P<Po(
Onesi
ded-
lef
tsi
ded)

Sampl ing:-Asampl ei
sasubsetofi ndi
vidualsf
rom al arger
population.Sampli
ngmeanssel ecti
ngthegr oupthatyouwi ll
actual
lycol l
ectdat
af r
om i
nyourr esear
ch.Forexampl e,ifyouar
e
researchingtheopini
onsofst
udent sinyouruniversi
ty,youcould
surveyasampl eof100st
udents.

Inst
ati
sti
cs,sampl
ingal
lowsy out
otestahy
pot
hesi
saboutt
he
char
acter
ist
icsofapopulat
ion.
Sampl i
ngi sapr ocessusedinstati
sti
calanalysi
sinwhicha
predetermi nednumberofobser v
ati
onsar etakenfrom alar
ger
population.Themet hodologyusedtosampl efrom alar
ger
populationdependsont het y
peofanalysisbeingperfor
med, but
i
tmayi ncludesimplerandom samplingorsystemat i
csampling.

Essenceofsampl
ing:
-Theessent
ial
sofsampl
i
ngar
e:

 Thesampl
emustt
rul
yrepr
esentt
hepopul
ati
on.

I
tssi
zemustbeadequat
e.

 Youmustsel
ectt
hesampl
erandoml
yandi
ndependent
ly.

 Thepopul
ati
onmustbehomogeneous.

Typesofsampli
ng:-Therearefi
vetypesofsampli
ng:
Random,
Sy
stemati
c,Conveni
ence,
Cluster
,andStrat
if
ied.

Random sampl ingisanal ogoust oputt

ingev ery
one'snamei nt
oa
hatanddr awingoutsev eralnames.Eachel ementint he
populat i
onhasanequal chanceofoccur ing.Whilethisisthe
preferredwayofsampl ing, i
tisoftendif
ficultt
odo.I trequir
es
thatacompl eteli
stofev eryelementint hepopulati
onbe
obtained.Comput ergener at
edlistsareoftenusedwi thrandom
sampl ing.Youcangener aterandom number susingt heTI82
calculator.
Syst
emati
csamplingiseasiertodothanr andom sampling.In
syst
emati
csampli
ng,thelistofelement si
s"countedoff".Thati
s,
ever
ykthel
ementistaken.Thisissimilart
olini
ngev er
yoneup
andnumberi
ngoff"1,
2,3,
4;1,2,
3,4;et
c".Whendonenumber i
ng,al
l
peopl
enumbered4woul dbeused.

Conveniencesampli
ngisveryeasyt odo,butit
'spr
obablythe
worsttechni
quetouse.Inconveniencesampl i
ng,r
eadi
lyavail
abl
e
dat
aisused.Thatis,thefi
rstpeoplethesurveyorr
unsinto.

Clustersampli
ngisaccomplishedbydivi
dingthepopulat
ioninto
groups--usual
lygeographi
call
y.Thesegroupsarecal
ledcluster
s
orblocks.Theclust
ersarerandomlysel
ected,andeachelement
i
nt heselect
edcluster
sareused.

Str
atif
iedsampl ingalsodividesthepopulati
onintogroupscall
ed
str
ata.Howev er,thi
st i
mei tisbysomechar act
erist
ic,not
geographical
l
y .Forinstance,thepopulati
onmightbesepar at
ed
i
ntomal esandf emal es.Asampl eistakenfrom eachofthese
str
atausingeitherrandom, syst
ematic,orconv
eniencesampl i
ng.

Er
ror-|t
ype:
-

Thef i
rstkindoferrori
stherej
ect
ionofatr
uenullhypothesi
sas
theresultofatestprocedur
e.Thi
skindofer
rori
scal l
edat y
peI
error(
falseposit
ive)andissometi
mescall
edaner r
orofthefir
st
kind.

Inter
msofthecourtroom example,
aty
peIer
rorcor
respondst
o
convi
cti
nganinnocentdefendant
.
Er
ror-|
|ty
pe:
-
Thesecondkindoferrorist
hefail
uret
orejectaf al
senullhypot
hesisas
theresul
tofatestpr
ocedure.Thi
ssortoferroriscall
edat y
peIIerr
or
(fal
senegat
ive)andi
sal sor
eferr
edtoasaner rorofthesecondkind.

I
nter
msoft hecour
troom ex
ampl
e,at
ypeI
Ier
rorcor
respondst
oacqui
tt
ing
acr
iminal
.

St
andar
der
rorofmean(
SEM):
-
Thest andarderror(
SE)ofast at
isti
cistheappr oxi
mat estandar
ddev iat
ion
ofast ati
sti
calsamplepopulati
on.Thest andarderrorisastati
sti
calterm
thatmeasur estheaccuracywithwhichasampl edistr
ibut
ionrepresentsa
populati
onbyusi ngstandarddeviat
ion.I
nst at
isti
cs,asampl emean
deviat
esf r
om theactualmeanofapopul ati
on;thisdeviat
ionisthe
standarderrorofthemean.

Therelati
onshi pbetweenthestandarderr
orandt hestandarddev i
ati
onis
suchthat,foragi vensamplesize,t
hestandarderrorequalst hestandard
devi
ati
ondi v idedbythesquarerootoft
hesampl esize.Thest andarderr
or
i
salsoi nverselypropor
ti
onaltothesamplesize;thelargerthesampl esi
ze,
thesmal l
ert hestandar
derrorbecausethestati
sticwil
l approacht heact
ual
val
ue.