You are on page 1of 7

One proportion Z-test

Difference between;

Z-test - it is about the mean

One proportion Z-test – from the word it self “proportion” that has to do with parameter proportion

*If we are testing a claim about a population proportion, then the appropriate random variable will be
“p̂ ” (p hat), the sample proportion

*If we select random samples of size 𝑛, where both 𝑛𝑝 > 10 and 𝑛(1 – 𝑝) > 10, then the sampling
distribution for p̂ will be approximately normal.

 𝑛𝑝 > 10 and 𝑛(1 – 𝑝) > 10 – the criteria we need to look for before we start using the one
proportion Z-test because if the given is not true if “np” is not greater than 10 also, “n” 1 minus
“p” is not greater than 10, then we cannot proceed with ne proportion Z-test.
 𝑛𝑝 > 10 and 𝑛(1 – 𝑝) > 10 – if this is true then, the sampling distribution for “p̂” (p hat) will be
“approximately normal” also, if we have a normal distrbution we can use the Z table.

Assumptions
*Data comes from a random sample σ

*The sample size an hypothesized proportion satisfy: 𝑛𝑝 > 10 and 𝑛(1 – 𝑝) > 10

 The “p” in “np” is classify as hypothesized proportion


 The “n” is classfy as Sample size

Test Statistic

̂−𝑝𝑜
p
We can also be simplied it as,
σp
Mean and Standard Deviation
*The mean for this sampling distribution is 𝜇𝑝 = p

* The standard deviation for this sampling distribution is

Example
*A consumer group asserts that the proportion of households that have three cell phones is 30%. A cell
phone company has reason to believe that the proportion more than 30% (the 30% is the claim). Before
they start a big advertising campaign, they conduct a hypothesis test.

* Their marketing people survey 150 (the n = 150) households with the result that 60 of the households
have three cell phones; this data is then used to test the claim that more than 30% of households have
three cell phones. At the 5% significance level, is there sufficient evidence to support the company’s
claim?

- The 30% is the claim so it means it must be the hypothesized proportion;

P0 = 0.3 or the 30%

n = 150
60
p̂ = = 0.4 or 40%
150

× = 0.05

Set up 𝐻o and 𝐻a
The phone company’s claim is that 𝑝 > 0.30, which will be represented by the alternative hypothesis 𝐻a .
The null hypothesis will be contradictory statement, giving us the hypotheses:

𝐻o: 𝑝 ≤ 0.30

𝐻a: 𝑝 > 0.30

Statistical Test
*Since the claim involves a population proportion, the appropriate sample statistic will be a sample
proportion, p̂ .

*Also, note that 𝑛𝑝 = 150(.30) = 45 and 𝑛(1 – 𝑝) = 150(.70) = 105, both are greater than 10, so the
sampling distribution for 𝑝Ƹwill be approximately normal.

*Hence, we will use a one-proportion z-test.

P-value
P –value

*If the table that you are using is giving you the area to the right of the –score but if it’s not suppose
that it the table gives values to the left then you have to subtract from 1.

Decision

*This p-value (0.0038) is less than the significance level (0.05), so our decision is to reject 𝐻o.

- (Decision rule: Reject Ho if the p –value < ×, otherwise fail to reject Ho)

*Moreover, this p-value tells us that if the null hypothesis is true, there is only about a 0.38% chance
that the sample proportion p̂ for a randomly selected sample will either be at least as far off from the
expected proportion of 0.30.

Conclusion
*Thus, at the 0.05 level of significance, the sample data provides sufficient evidence that the percentage
of households with three cell phones is more than 30%.

*In other words, we have significant statistical evidence that the company’s claim is correct.
Second Example
*In a study of 420,019 (the n = 420,019) cell phone users, 172 (the x =172) of the subjects developed
brain cancer. Test the claim that cell phone users developed brain cancer at a greater rate than that for
non-cell phone users (the rate of brain cancer for non-cell phone users is 0.0340%).
𝑥 172 𝑃= 0.0340%
p̂ 𝑛
= 420,019 = 𝑃=0.00034

× = 0.005
*Since this is a critical issue, use a 0.005 significance level.

Set up 𝐻o and 𝐻a
*Let 𝑝 be the proportion of cell phone users that develop brain cancer; then the claim is that 𝑝 >
0.00034. This claim will be represented by the alternative hypothesis, so we get:

𝐻o: 𝑝 ≤ 0.00034

𝐻a: 𝑝 > 0.00034

Statistical Test
*Since the claim involves a population proportion, the appropriate sample statistic will be a sample
proportion, p̂ .

*Also, note that 𝑛𝑝 = 420019(0.00034) = 142.8 and 𝑛(1 – 𝑝) = 420019(0.99966) = 419876.2, both are
greater than 10, so the sampling distribution for p̂ will be approximately normal.

*Hence, we will use a one-proportion z-test.

P –value
Decision
*This p-value (0.0073) is greater than the significance level (0.005), we do not reject 𝐻o.

- (Decision rule: Reject Ho if the p –value is less than ×, otherwise fail to reject Ho)

Conclusion
*We conclude that at 0.005 level of significance there is not enough evidence to support the claim of
higher brain cancer rates for cell phone users.

Analysis of Variance
ANOVA (Analysis of Variance)
Completely Randomized Design

- CRD is best suited for experiment with a small number of ttreatments. One can use CRD when
experimental units are homogenous and if there is effective local control
- Analysis of Variance (ANOVA) – is a statistical test that is used to determined if there are
differnces between population means when there are more than 2 groups being compared.

ANOVA Table:
Source of df SS MS Fc Fx (df1, df2)
Variation
Treatment T–1 TrSS MSTr 𝑀𝑆𝑇𝑟 *from the F-
Experimental N-t ESS MSE 𝑀𝑆𝐸 Table
error
total n-1 TSS

Decision Rule: Reject if Ho if p –value is less than x , o.f.t.r.Ho alternatively, since we are using Fc and Fx ,
the decision can be stated as:

Reject if Ho if Fc and Fx , o.f.t.r.Ho

Problem
*An aquaculture operator wishes to determine if the effect on fat evels of his tiger prawns differ among
four randomly selected types of feed (A,B,C,D). a type of feed was applied to several enclosed cages
containing the prawns. From each cage, a single prawn was obtained and the fat content (in mg) was
determined.

The data illustrates in the table

Set up 𝐻o and 𝐻a

There are no differences among the effect of the different type of fertilizer

There are differences among the effects of different type of fertilizer

Test Procedure: ANOVA


𝑀𝑆𝑇𝑟
Test Statistic or Fcomputed : Fc = 𝑀𝑆𝐸

Ftabular or Fx (df1, df2) = *can be found using F-table

Computations:
(𝑡𝑜𝑡𝑎𝑙 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛)2
Correction factor = 𝑛
= 6447.468

Total Sum of Squares (TTS) = (11.22 + 11.62 + … + 9.3)2 – CF = 481.092


140.22 109.72 140.22 49.72
TrSS = ( 9
+ 7
+ 8
+ 6
) – CF = 324.379

ESS = TSS – TrSS = 156.713


𝑑𝑓1=3
ANOVA Table ( use x(alpha) = 0.05
𝑑𝑓2=26

Source of df SS MS Fc Fx (df1, df2)


Variation
Treatment 3 324.379 108.13 17.9 2.98
Experimental 26 186.713 6.03
error
total 29 481.092

Decision
Since 17.939 >2.98, reject Ho

Conclusion
At x = 0.05, we have sufficient evidence to say that there are differences among the effects of feed
types.

You might also like