0% found this document useful (0 votes)
104 views16 pages

Hipotese Toetse: Eensteekproef Hypothesis Testing: One Sample

This document discusses hypothesis testing concepts including: - Null and alternative hypotheses, which make opposite statements to be tested - Test statistics, which are functions used to test hypotheses - Critical values, which are cut-off points used to determine if a hypothesis can be rejected or not rejected - Rejection and non-rejection regions, which are areas determined by critical values and alternative hypotheses to classify test statistics - Significance levels, which determine the acceptable probability of Type I errors

Uploaded by

Gregor
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
104 views16 pages

Hipotese Toetse: Eensteekproef Hypothesis Testing: One Sample

This document discusses hypothesis testing concepts including: - Null and alternative hypotheses, which make opposite statements to be tested - Test statistics, which are functions used to test hypotheses - Critical values, which are cut-off points used to determine if a hypothesis can be rejected or not rejected - Rejection and non-rejection regions, which are areas determined by critical values and alternative hypotheses to classify test statistics - Significance levels, which determine the acceptable probability of Type I errors

Uploaded by

Gregor
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

9 Hipotese Toetse:

Eensteekproef
Hypothesis Testing:
One Sample

Statistics are like a bikini.


What they reveal is suggestive, but what they conceal is vital.

A statistician is a person whose lifetime ambition is to be wrong 5% of the


time.

Hipotese Toetse
Hypothesis Testing

In die vorige hoofstuk is vertrouensintervalle gebruik om vir


ons ‘n idee te gee wat die waarde van ‘n populasie
parameter moontlik is.
In the previous chapter confidence intervals were used to
give us an idea of what the value of a population
parameter might be.

In hierdie hoofstuk gaan ons spesifieke stellings oor


populasie parameters toets.
In this chapter we are going to test specific statements
about population parameters.

 Sal meer as 30% van verbruikers my produk koop?


Will more than 30% of consumers buy my product?

 Is alle Coke blikkies gevul met gemiddeld 500ml?


Are all Coke cans filled with an average of 500ml?

 Help dit om in die klas te wees vir IS 314?


Does it help to be in class for IS 314?

 Is operasie A beter as operasie B?


Is operation A better than operation B?

1
Nulhipotese en Alternatiewe Hipotese
Null Hypothesis and Alternative Hypothesis

Vir enige hipotesetoets is daar altyd ‘n gelyk aan stelling


wat ons wil toets, genoem die nulhipotese (H0).
For any hypothesis test there is always an equal statement
that we want to test, called the null hypothesis (H0).

Die alternatiewe hipotese (HA of H1) is die teenoorgestelde


van die nulhipotese.
The alternative hypothesis (HA of H1) is the opposite of the
null hypothesis.

Enige hipotesetoets toets of hierdie gelyk aan stelling kan


korrek wees of nie.
Any hypothesis test tests if this equal statement can be
true or not.

Hoekom kan ons slegs gelyk aan stellings toets?


Why can we only test equal to statements?

Wiskundig is dit moeilik om ‘n kleiner as, groter as, of


ongelyk aan stellings te hanteer.
Mathematically it is difficult to handle a less than, greater
than, or unequal statement.

Ons gaan ‘n wiskundige formule gebruik om die


hipotesetoets te doen, en in daardie formule moet ons
die waarde sit waarvoor ons wil toets.
We will use a mathematical formula to do the hypothesis
test, and in that formula we need to put the value that
we want to test for.

H0: µ = 50  f(50)

H0: σ = 10  f(10)

H0: µ = 34  f(34)

H0: µ > 50  f(???) f(60), f(50.1), f(1000), …


H0: σ ≠ 50  f(???) f(12), f(1), f(149.9), …

2
H0 het dus altyd ‘n = teken in: =, ≥, of ≤
H0 wil always have = sign in it: =, ≥, of ≤

In die alternatief sit ons die teenoorgestelde teken:


In the alternative we put the opposite sign:
H0: =  H1: ≠
H0: ≤ (=)  H1: >
H0: ≥ (=)  H1: <

H0: µ = 50 H0: µ ≤ (=) 102.4 H0: µ ≥ (=) -54


H1: µ ≠ 50 H1: µ > 102.4 H1: µ < -54
Tweekantig Regs-Eenkantig Links-Eenkantig
Two-Sided Right One-sided Left One-Sided

Toetsstatistiek
Test Statistic

Vantevore van ‘n funksie gepraat waarin ons die waarde


van die hipotese sit, hierdie funksie word die
toetsstatistiek genoem.
Previously we spoke about a function that we put the value
of the hypothesis into, this function is called the test
statistic.

Elke verskillende hipotesetoets het ‘n verskillende


toetsstatistiek.
Every different hypothesis test has a different test statistic.

Dit is bloot ‘n funksie waarin ons die steekproef inligting,


die hipotese se waarde, en ander gegewe inligting in sit
om een enkele waarde te kry.
It is simply a function into which we put some sample
information, the hypothesised value, and some other
given information to give us a single value.

Hierdie enkele waarde is die waarde wat ons dan gebruik


om te besluit of die hipotese kan waar wees of nie.
This single value is the value that we’ll use to determine if
they hypothesis can be true or not.

X  0
TS 
S/ n

3
9-2) Vir µ, σ bekend
For µ, σ known

9-3) Vir µ, σ onbekend


For µ, σ unknown

9-4) Vir σ2 en σ
For σ2 and σ

9-5) Vir π
For π

10-1) Vir µ1 teenoor µ2, σ’s bekend


For µ1 versus µ2, σ’s known

10-2) Vir µ1 teenoor µ2, σ’s onbekend


For µ1 versus µ2, σ’s unknown

10-4) Vir µ1 teenoor µ2, afhanklike steekproewe


For µ1 versus µ2, dependent samples

10-5) Vir σ21 teenoor σ22


For σ21 versus σ22

10-6) Vir π1 teenoor π2


For π1 versus π2

Kritieke Waarde
Critical Value

Nadat ons ‘n toetsstatistiek het, moet is dit teenoor ‘n


seker waarde vergelyk om sodoende te besluit wat ons
moet doen – kan die hipotese waar wees of nie.
Once we have a test statistics, we need to compare it to a
value to decide what decision we should make – can the
hypothesis be true or not.

Ons gaan weer van die steekproefverdeling van hoofstuk 7


gebruik maak vir hierdie taak.
We are again going to make use of the sampling
distributions of chapter 7 for this.

4
X ~ Nor(0,n2 )

0

Uit hoofstuk 7 weet ons dat die gemiddelde ‘n spesifieke


normaal verdeling het rondom μ0, indien ons hipotese
korrek is.
From chapter 7 we know that the mean has a specific
normal distribution around μ0, if our hypothesis is
correct.

Indien die waarde wat ons uit ons steekproef kry “naby”
die middel van die verdeling is, dan kan die hipotese oor
μ0 moontlik waar wees.
If the value we get from our sample is “close” to the
middle of the distribution, then the hypothesis about μ0
could be true.

?? 0 =R75 ??

X1  R50 X 2  R65 X 3  R90 X 4  R115

Die kritieke waarde is die afsnypunt tussen sê of die


hipotese dalk kan waar wees en of dat dit eerder vals is.
The critical value is the cut-off point between saying if a
hypothesis could be true or if it is false.

Vir elke hipotesetoets gaan ons ‘n verskillende verdeling


gebruik om die kritieke waarde uit af te lees.
For each difference hypothesis test we will use a different
distribution to read the critical value from.

Z-tabel / table
t-tabel / table
χ2-tabel / table
F-tabel / table

5
Verwerpings en Nie-Verwerpingsgebiede
Rejection and Nonrejection Regions

Hierdie gebiede word bepaal deur die kritieke waarde en


die alternatiewe hipotese.
These areas are determined by the critical value and the
alternative hypothesis.

Indien die toetsstatistiek in die verwerpingsgebied val,


verwerp ons die hipotese en indien dit nie in die
verwerpingsgebied val nie, kan ons nie die hipotese
verwerp nie.
If the test statistic falls in the rejection region we reject the
hypothsis and if it does not fall in the rejection region we
cannot reject the hypothesis.

Tweekantig
Two-Sided H0:  50
H1:  50

KW / CV 50 KW / CV
Verwerpingsgebied Nie-Verwerpingsgebied Verwerpingsgebied
Rejection Region Non-Rejection Region Rejection Region

Regseenkantig
Right One-Sided H0:  50
H1:  50

50 KW / CV
Nie-Verwerpingsgebied Verwerpingsgebied
Non-Rejection Region Rejection Region

6
Linkseenkantig
Left One-Sided H0:  50
H1:  50

KW / CV 50
Verwerpingsgebied Nie-Verwerpingsgebied
Rejection Region Non-Rejection Region

Betekenispeil
Significance Level

Die betekenispeil is weereens die fout wat ons bereid is om


te maak met die hipotese. Dit is identies aan dit wat ons
gedoen het by vertrouensintervalle.
The significance level is again the mistake we are willing to
make with the hypothesis test. It is identical to what we
used with confidence intervals.

Die betekenispeil bepaal hoe groot die oppervlak van die


verwerpingsgebiede moet wees.
The significance level determines how large the area of the
rejection regions should be.

Daar is twee tipe foute wat ons kan maak indien ons ‘n
hipotesetoets doen.
There are two different types of errors we can make when
doing a hypothesis test.

Tipe I   P(Tipe I Fout)


Type I   P(Type I Error)

Tipe II   P(Tipe II Fout)


Type II   P(Type II Error)

7
Tipe I Fout / Type I Error

Ons maak hierdie fout indien ons ‘n hipotese verwerp wat


eintlik korrek is.
We make this error if we reject a hypothsis that is actually
correct

H0: µ ≤ 1000 H1: µ > 1000

Ons sê dat die gemiddelde inkomste van ‘n populasie is


meer as R1000 (H1), terwyl dit eintlik wel R1000 is (H0).
We say that the average income for a population is more
than R1000 (H1), when it is actually equal to R1000 (H0).

Tipe II Fout / Type II Error

Ons maak hierdie tipe fout indien ons ‘n hipotese aanvaar


wat eintlik verkeerd is.
We make this error if we accept a hypothesis that is
actually false.

H0: µ ≤ 1000 H1: µ > 1000

Ons sê dat die gemiddelde inkomste van ‘n populasie wel


R1000 is (H0), terwyl dit eintlik meer as R1000 moet
wees (H1).
We say the the average income of a population is R1000
(H0), when in fact it should be more than R1000 (H1).

Ons besluit wat ons wil hê die waarskynlikheid van die


twee foute moet wees.
We have to decide what we want the probability of the two
errors should be.

 is die waarskynlikheid van ‘n tipe I fout - waarskynlikheid


H0 is reg, indien ons die hipotese verwerp.
 is the probability of a type I error - probability that H0 is
correct, if we reject it.

 is die waarskynlikheid van ‘n tipe II fout - waarskynlikheid


H0 is verkeerd, indien ons nie die hipotese aanvaar nie.
 is the probability of a type II error - probability that H0 is
incorrect, if we do not reject the hypothesis it.

8
H0 :  Tweekantig
H1:  Two-Sided

  0.05

0.025 0.025

KW / CV  Z0.025 0 KW / CV  Z0.975

H0 :  Tweekantig
H1:  Two-Sided

  0.10

0.05 0.05

KW / CV  Z0.05 0 KW / CV  Z0.95

H0 :  Links Eenkantig
H1:  Left One-Sided

  0.05

0.05

KW / CV  Z0.05 0

9
H0 :  Regs Eenkantig
H1:  Right One-Sided

  0.07

0.07

0 KW / CV  Z0.93

P-Waarde van ‘n toets (tipe I fout, α)


P-Value of a test (type I error, α)

“Nuwe” en meer akkurate manier van hipotesetoetse doen.


“New” and more accurate way of doing hypothesis testing.

Met die kritieke waarde metode van hipotesetoets doen


vergelyk ons bloot die toetsstatistiek met die kritieke
waarde en indien dit in ‘n verwerpingsgebied val verwerp
ons andersins kan ons nie verwerp nie.
With the critical value approach for hypothesis testing we
compare the test statistics to the critical value and if it
falls in a rejection region we reject otherwise we cannot
reject.

KW KW

TS3 TS1 0 TS2 TS4

10
TS1 Nie Verwerp
Not Reject
TS2 Verwerp
Reject
TS3 Verwerp, maar baie naby ???
Reject, but very close ???
TS4 Verwerp, maar vergelyking met TS2 ???
Reject, but compared to TS2 ???

Beter manier nodig om te onderskei tussen TS2, TS3, en


TS4.
We need a better way to distinguish between TS2, TS3, and
TS4.

Die p-waarde van ‘n hipotesetoets is die presiese tipe I fout


wat ons sal maak indien ons verwerp.
The p-value for a hypothesis test is the exact type I error
we will make if we reject.

Indien hierdie fout klein genoeg is (  ) sal ons verwerp,


andersins kan ons nie verwerp nie.
If this error is small enough (  ) we will reject, otherwise
we cannot reject.

Alle hipotesetoetse behoort met ‘n p-waarde gedoen te


word, ongelukkig is dit nie altyd maklik om dit te
bereken nie. Met die hand slegs vir normaal verdeling.
All hypothsis tests should be done with a p-value,
unfortunately not that easy to calculate. By hand we can
only do it for the normal distribution.

Die p-waarde se berekening verskil na gelang van die


alternatiewe hipotese.
The calculation of the p-value differs depending on the
alternative hypothesis.

H1: µ ≠ p = 2 x P(Z > |TS|)

H1: µ < p = P(Z < TS)

H1: µ > p = P(Z > TS)

Verwerp indien p  waarde  


Reject if p  value  

11
TS = 2.76 H1: µ ≠
p = 2 x P(Z > |2.76|)
= 2 x {1 – P(Z < 2.76)} = …
= 0.0058

TS = -2.43 H1: µ <


p = P(Z < -2.43) = …
=0.0075

TS = -0.54 H1: µ >


p = P(Z > -0.54)
= 1 – P(Z < -0.54) = …
= 0.7054

Bereken van ‘n tipe II fout (β),


onderskeidingsvermoë en steekproefgrote (n)
Calculation of a type II error (β), power and
sample size (n)

β Waarskynlikheid ons verwerp nie H0 nie; indien H1 geld.


Probability we do not reject H0; given that H1 is true.

Power = 1- β
Waarskynliheid ons verwerp H0, indien dit verkeerd
is, dus H1 waar is.
Probability we reject H0 if it is false, thus H1 is true.

n = f(α, β, …)

Stappe in ‘n hipotesetoets
Steps in a hypothesis test

1) Stel H0 en H1 (= , ≠) (≤ , >) (≥ , <)


State H0 and H1

2) Besluit op die betekenisvlak van die toets ( )


Select the significance level of the test

3) Kies die toetsstatistiek wat gebruik gaan word.


Select the test statistc that will be used.

4) Kry die verwerpings en nie-verwerpingsgebiede.


Get the rejection and nonrejection regions.

12
5) Kry die data en bereken die toetsstatistiek.
Get the data and calculate the test statistic.

6) Kyk of die berekende toetsstatistiek in die verwerpings


of nie-verwerpingsgebied val of bereken p-waarde.
See if the calculated test statistics falls into the rejection
or the nonrejection regions or calculate p-value.

7) Maak ‘n statistiese besluit: Verwerp of Nie-Verwerp H0.


Make a statistical decision: Reject or Don’t Reject H0.

8) Interpreteer die statistiese besluit in die konteks van die


probleem.
Interpret the statistical decision in die context of the
problem.

Hipotesetoets vir μ, σ bekend


Hypothesis Test for μ, σ known

Indien ons ‘n hipotese toets oor µ wil doen en ons weet


wat σ is, is die toetsstatistiek wat ons gaan gebruik:
If we want to do a hypothesis test about µ and σ is known,
out test statistic is:
x
TS  Z   0 ~ Normaal
n
Hierdie toetsstatistiek volg ‘n normaal verdeling en dus
gaan ons ons kritieke waarde in die Z-tabel opsoek. Ons
kan ook ‘n p-waarde vir hierdie toets bereken.
This test statistic has a normal distribution and we will look
up our critical value in the Z-table. We can also calculate
a p-value for this test.

EXAMPLE 9-2 (p. 303)


 kritieke waarde of p-waarde
critical value or p-value

EXAMPLE 9-3 (p. 305)


 Klas manier om tipe II fout te bereken
Class example to calculate the type II error

 Los steekproefgrote
Leave sample size

13
Hipotesetoets vir μ, σ onbekend
Hypothesis Test for μ, σ unknown

Indien ons ‘n hipotese toets oor µ wil doen en ons weet nie
wat σ is, is die toetsstatistiek wat ons gaan gebruik:
If we want to do a hypothesis test about µ and σ is
unknown, out test statistic is:
x  0
TS  t  ~ tn1
S
n
Hierdie toetsstatistiek volg ‘n t-verdeling en dus gaan ons
ons kritieke waarde in die t-tabel opsoek.
This test statistic has a t-distribution and we will look up
our critical value in the t-table.

EXAMPLE 9-6 (p. 313)


 kritieke waarde
critical value

Hipotesetoets vir σ2
Hypothesis Test for σ2

Indien ons ‘n hipotese toets oor σ2, die populasie variansie,


is die toetsstatistiek wat ons gaan gebruik:
If we want to do a hypothesis test about σ2, the population
variance, our test statistic is:

TS   2  (n 1)s ~ n21
2
2 0
Hierdie toetsstatistiek volg ‘n chikwadraat verdeling en dus
gaan ons ons kritieke waarde in die χ2-tabel opsoek.
This test statistic has a chi-squared distribution and we will
look up our critical value in the χ2-table.

EXAMPLE 9-8 (p. 321)

14
Hipotesetoets vir p
Hypothesis Test for p

Indien ons ‘n hipotese toets oor p, die populasie proporsie,


is die toetsstatistiek wat ons gaan gebruik:
If we want to do a hypothesis test about p, the population
proportion, our test statistic is:
p  0
TS  Z  ~ Normal
 0(1 0)
n
Hierdie toetsstatistiek volg ‘n normaal verdeling en dus
gaan ons ons kritieke waarde in die Z-tabel opsoek. Ons
kan ook ‘n p-waarde vir hierdie toets bereken.
This test statistic has a normal distribution and we will look
up our critical value in the Z-table. We can also calculate
a p-value for this test.

EXAMPLE 9-10 (p. 325)


 kritieke waarde of p-waarde
critical value or p-value

 gebruik (9-41) as TS
use (9-41) as TS

15
16

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy