µ = 67.

4
σ = 3.9

Numerical variables from a €

€
single sample Y is normally distributed
whenever:
Y is normally distributed
Chapter 11 or
µ = 67.4
µY = € n is large
σ 3.9
σY = = = 1.7
n 5
€

Inference about means But... We don’t know s . . .

Because Y is normally distributed, we can convert
its distribution to a standard normal distribution:
However, we do know s, the standard deviation
of our sample. We can use that as an estimate
€ Y −µ Y −µ of s.
Z= =
σY σ n

This would give a probability distribution of the

difference between a sample mean and the population
mean.
€
 In most cases, we
A good approximation to the
N
don’t know the real
µ = 67.4 population
distribution.
standard normal is then:
σ = 3.9

€
Y −µ Y −µ
€ We only have a
sample.
t= =
SE Y s/ n
Y = 67.1 s = 3.1
s 3.1
SE Y = = = 1.4
n 5
€ €
We use this as an
estimate of σ Y
€
€

t has a Student’s t €
distribution Degrees of freedom
Discovered by William Gossett, of
the Guinness Brewing Company

df = n - 1
}

Y −µ
Z=
σY
Y −µ
t= Z t9
SE Y
€
 We use the t-distribution to calculate a 95% confidence interval for a
confidence interval of the mean mean

Example:
Y −µ Paradise flying snakes
−tα ( 2),df < < tα ( 2),df
SE Y
We rearrange the above to generate:

Y − tα ( 2),df SE Y < µ < Y + tα ( 2),df SE Y

€ Undulation rates (in Hz)
Another way to express this is: Y ± SE Y tα ( 2),df 0.9, 1.4, 1.2, 1.2, 1.3, 2.0, 1.4, 1.6

€
Estimate the mean and standard
deviation

Y = 1.375
https://www.youtube.com/watch?v=16aGSx9gFO4
s = 0.324
n=8

€
 Find the critical value of t
Find the standard error

df = n − 1 = 7
Y ± SE Y tα ( 2),df
tα ( 2),df = t 0.05( 2),7
= 2.36
s 0.324
SE Y = = = 0.115
n 8

€ Table C: Student's t distribution

Finding the critical value of t in R

qt(0.025, df = 7, lower.tail = FALSE)

[1] 2.364624
Putting it all together... 99% confidence interval
tα ( 2),df = t 0.01( 2),7 = 3.50
Y ± SE Y tα ( 2),df = 1.375 ± 0.115 (2.36)
= 1.375 ± 0.271 Y ± SE Y tα ( 2),df = 1.375 ± 0.115 ( 3.50)
€ = 1.375 ± 0.403
1.10 < µ < 1.65
(95% confidence interval) 0.97 < µ < 1.78

Confidence interval for mean€

in R Confidence interval for the variance

t.test(undulationRate)$conf.int
[1] 1.104098 1.645902 df s2 2 df s2
attr(,"conf.level") 2
≤σ ≤ 2
[1] 0.95 χα χ α
,df 1− ,df
2 2

t.test(undulationRate,
conf.level = 0.99)$conf.int
[1] 0.9740838 1.7759162
attr(,"conf.level") €
[1] 0.99
 2.5%
α = 0.05 95% confidence interval for the variance
of flying snake undulation rate
2.5%
Frequency
df s2 2 df s2
2
≤σ ≤ 2
χα χ α
,df 1− ,df
2 2

df = n - 1 = 7
χ2
χ2 1- α/2 χ2 α/2 € s2 = (0.324)2 = 0.105

χ α2 2
= χ 0.025,7 = 16.01
,df
2

χ2 α 2
= χ 0.975,7 = 1.69
1− ,df > qchisq(0.025,df=7,lower.tail=FALSE)
2
[1] 16.01276
Table A
df > qchisq(0.975,df=7,lower.tail=FALSE)
X 0.999 0.995 0.99 0.975 0.95 0.05 0.025 0.01 0.005 0.001 [1] 1.689869
1 1.6 3.9E-5 0.00016 0.00098 0.00393 3.84 5.02 6.63 7.88 10.83
E-6
€
2 0 0.01 0.02 0.05 0.1 5.99 7.38 9.21 10.6 13.82
3 0.02 0.07 0.11 0.22 0.35 7.81 9.35 11.34 12.84 16.27
4 0.09 0.21 0.3 0.48 0.71 9.49 11.14 13.28 14.86 18.47
5 0.21 0.41 0.55 0.83 1.15 11.07 12.83 15.09 16.75 20.52
6 0.38 0.68 0.87 1.24 1.64 12.59 14.45 16.81 18.55 22.46
7 0.6 0.99 1.24 1.69 2.17 14.07 16.01 18.48 20.28 24.32
8 0.86 1.34 1.65 2.18 2.73 15.51 17.53 20.09 21.95 26.12
95% confidence interval for the variance One-sample t-test
of flying snake undulation rate

df s2 2 df s2 The one-sample t-test compares the mean of a random

≤σ ≤ 2 sample from a normal population with the population mean
χ α2 χ α
,df 1− ,df proposed in a null hypothesis.
2 2

2 2
7 (0.324 ) 7 (0.324 )
≤σ2 ≤
16.01 1.69

0.0459 ≤ σ 2 ≤ 0.435

Test statistic for one-sample t-test Hypotheses for one-sample t-tests

Y − µ0 H0 : The mean of the population is µ0.

t=
s/ n
HA: The mean of the population is not µ0.

µ0 is the mean value proposed by H0

€
 Human body temperature
Example: Human body temperature
n = 24
Y = 98.28
H0 : Mean healthy human body
temperature is 98.6ºF. s = 0.940
HA: Mean healthy human body
temperature is not 98.6ºF.
Y − µ0 98.28 − 98.6
t=€ = = −1.67
s/ n 0.940 / 24

€ Comparing t to its distribution

Degrees of freedom to find the P-value

df = n – 1 = 23
A portion of the t table

df α(1) α(1) α(1) α(1) α(1)

=0.1 =0.05 =0.025 =0.01 =0.005
α(2)=0.2 α(2)=0.10 α(2)=0.05 α(2)=0.02 α(2)=0.01
... ... ... ... ... ...
20 1.33 1.72 2.09 2.53 2.85
21 1.32 1.72 2.08 2.52 2.83
22 1.32 1.72 2.07 2.51 2.82
23 1.32 1.71 2.07 2.5 2.81
24 1.32 1.71 2.06 2.49 2.8
25 1.32 1.71 2.06 2.49 2.79

t.test(bodyTempSmallData$temperature, mu =
98.6)

One Sample t-test

–1.67 is closer to 0 than –2.07, so P > 0.05.
data: bodyTempSmallData$temperature
With these data, we cannot reject the null t = -0.56065, df = 24, p-value = 0.5802
alternative hypothesis: true mean is not equal
hypothesis that the mean human body to 98.6
temperature is 98.6. 95 percent confidence interval:
98.24422 98.80378
sample estimates:
mean of x
98.524
 Body temperature revisited:
Body temperature revisited: n = 130 n = 130
n = 130
Y = 98.25 t = −5.44
s = 0.733 t 0.05(2),129 = ±1.98

Y − µ0 98.25 − 98.6 €
t is further out in the tail than the critical value, so we
t=€ = = −5.44 could reject the null hypothesis. Human body
s / n 0.733/ 130 temperature is not 98.6ºF. P = 2.4 x 10–7
€

One-sample t-test: Assumptions

1. The variable is normally distributed.

2. The sample is a random sample.