diypathests

A hupo the sts tS a tentatve and totatble stalement

parposes a nelationship between Vaicbles
{hal Soentfic Investiqation and
Seve au a bsis fos
empizial Validation os falstfication
through to
is Subjcctexpeximcntatiom obsevation
coneming tsting dpothens
Basic Cncapl ot
Null typothesis: method B' about it
Compat melhoc A' with
# we ae sumption that
Supesoits and it we proceed on the
both metros anc equally good, then thu asumption
Is tmd as the nll hypothesi.

we should always have 'equal to' C) sig

"In this, statement about the
"Null thypothi is the specific
pararneter.

hyypotheis u the one Which one wcns to cisprove

Null
to the
Egi- the population mean C) a equalExprrss s
ypothes0sed men Ho) =oo, e can

2 Altesnative typothesis:
CopmPOrelmettod A with 'method B' and
IF we are to supeiot or the
that the method A ès
We mathink as atenative
ethod B is infeio, then it ia esned
huypothes.Concude rejecting
the null hypothesis.is
hat we
Enoun as alenative hypothesis.
"The Set of altenatives to the nl hypothess S
tted to as the altomative hypoth
" A|teonative hypothesis is usually he oe ishes to
Prove
 e may consider 8 posible atteonative
hypothess as foltows:

hypothexis reprcsent he hypothess

ihts anull reett and altomative hypotheis
We aue tying to
Tepsenti all othey possibilities.
3Tpe IErors:
rejecion o hypotheis which
Type I Enor meansacepted
been a
Should have
uee
ne may eject Ho When Ho s
the probabity
The SIze of Type 1 Enos u gven bywhich is denoted
type I Gos plleject llot,) which a
by o qualitycontol oc ès Called
Jh statstical
Procucer's isk
probability o type I Eror u uually au -thu
The (nderstool
cettmincd in advance, and s
lbvel o signifance of tettng the hypothesis,that
t94 type I Enor u fired at 5 it neans
thete aw about 5 chances in (oo that we will
eject H When to is tue.
tixing t at
le can ontiot type TEror just by
0 ot)
a lowes level (ot l'1. ;it would bé

1ype D enor mians accepfing the hypotheis

IWhich should have been vjecte d
hle mnay acept t when in at tbis not bre
the Sie of Type D Enor is given by theo
Pro babitity t bype f Enor p(Acept b/) hhh

In Statisttal andlys B Ucale d as Conunods

 Possible Ypothesis of lest_out comes
Actual situatin
Ho toe Ho fale
pecislon
Alo eoY
Acceptth Poobability (-a Pobabiihy P
Tpe I Eror No EhoY
Raject44 Probabiityoc
Probabili ty -p

Samplen When we toy to vdule

But AHh a fxed ofcomniting THre D
ype J Enot, he probabity
Enor inCcases.
cduced smuttancousu
types of Enos Cannot be
.Botn betwen boo type of Erors
.thoe a a tiace ott probabiuts if making one type
bich means that the if we ae wtling to
only be rdulced
of Enor Can ony he ofher type of éron
nccase he probabituts of mating
bsincss SituatioTg,
n
To daal wrth this trade ott
decision -makuu decide thi apprepnate level oB type J Enor
attached to
o penalities
by tramining -the cos
eos. ri
both type
One-tailed tes t:
5 Two-tailed and
alte mctive hypothesis
Based on the Sign in altemactive
(#>or 2)
We have 3 ditaent tesa
we have
klhun we have sgn in alttmative hypothuis
bwo-taíled test

theiu, we have
khen we have ' Sigr in atte mative hypo
igh tiled teat.

tailed test:
When we have <' signiwe have (ett
not matve Shaldthe od Samples
Impotant
care the one behween
awhether fosmal hypotheses
ofstatement of an intors, ot Same level
due knlhin
altc natuteof Capacity tons. natut a
use l
o with the or of diffeicnce
the thantons lo
H,:> :so. hypothess
Atleynative
an and weshouldl. level level measurme
tel
Valid.
beto
he
of thatConsidering mosé=lo is accompüshed levela Such
to beaning Null hypotheses test on
th:=80
hypothess Cbject sqrihcane < the Samples is
means hypotteSeA also means Hb:hspothesis onsiderati¡n.
under
roblem tested
boo-tailed as the
purpoSe
Cither hypo
in be hypothesis Nhether and aHhect-the
nvolved and load the
hupothesis making must
JAlkmative pratice, of
art Signihcane of
hypotheis (Ho)This stated,
Scems the the of be
be with idicatas. Speofied
be
Should
hypotheses means.
Sample
-hatnagnitudethe (ity
stops hypothesis
(H). problem
TesearchWhlch fosmulation
must
accordanaa a tor thu
g Vavúabi
Settng
in
up Conslsts ypothess Testing Null
cleaty bridqe Which
or
tot Selecting factors sthe nathu
n adopted
henerally
Voious also (ayTheThe Thu
c (asei tailed of
tst This nat be
10the The )
The stap In It ) Thelevel 1S
The (b (c) d
adequat meany Valtie the or
ijecting as
aqenisttyeethesis:
tailed Cett-tuled)
the
of hypothesis
the level
obtain value
Clowercnhal
statistc kor
mechanical WhetherasOn
of fov test yv and l2 Cight not
tosti by Pavarmetric statistic decistongiaboutvalueihealalue cvidence Is aids hy
be hatue wel statistic
eof Selecked When testing to of he ditucCaw
as indicati
mut Vae crifilal value usejul
and
theis nul
hypothsseE hul
statstical
a a
Puctuatio
not
Sigriane hpothetical tet hypothesis
decisleris upper <Cihial tsto intwedthatanly easons
puposehypo theobsevatios test
y
the Simply but
of viewqe
Conducting
andobtain
Naouane tibution
D0)the
and,
of
Value.
Significance the
he
stutistic
>
Statisticeject
statistic the
enaugh be
fn sts
the reasos
theyto
of t
the d nut cannot gf not Explain due
statistic thi
Sample the
ejecting
not test shadkept Eist, othr
the mak rcjectthe test
imitatins the decision-making
levelo we Value
Cátial
(
Stop, and undley dis
vaue.
Citial Compaing test test have be tselk, donot is
the context bew becau
of ot of
We Tescarcher the N) we
Decision Vau Vale de we testshould
diterence
ditfuene
the Nalue, Value value donot
otherwse, making
bne, Tedt propotio this statd Using Test
he using The
At Ale a) (b) (C) we D
ISo0 c) O) )
on Expesied
that hypoth ening
invenee suiüenty
greaBer problensthe
based ihen corect samples techniques
shows of
s the for good
auc beCannotngni7itant,
ifference
he entirelySmall be
in knouledge
test
on o Nigher thatinteoenceof
fexG
test a tuth sanples abatby'
When bascd
he
ef drauing Suggest
such statisicaluy
4he Case adaquat
SgnitcanotuliCestaintby, to he generally
of
the
that infksence
said n he
Sze linitationg
as conceníng Stgnikilance,
wtthuth
and so
chance
sugget be So pro
bahltt÷ the
b
be
Specially
along
to statistial
of Probablities
dttence cannot to
()Rosls, due eidences
leliab:lity
happens
cornbired
theseStatistical
Gmpy t s his,s
fnakr
Utth teet the
not 0 Atl be
p
*o
s