Professional Documents
Culture Documents
7) M A1 Hypothesis Testing Notes
7) M A1 Hypothesis Testing Notes
b) Calculate:
Population parameter
A population parameter is a statistical Test statistic
measure relating to a population A test statistic is the result of the experiment we
are using, which we use to test !!
7A
Hypothesis Testing
A Hypothesis is a statement made What is the test statistic for this situation?
about the value of a population
parameter. You can test a
hypothesis to see if there is enough
evidence to change it
Let us think about a practical Write a sensible null hypothesis for this situation
example:
à We roll the dice 20 times and get Write a sensible alternative hypothesis for this
a 6 on 8 occasions situation
7A
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
Null Hypothesis A researcher wants to test,
A Hypothesis is a statement made à The dice is at the 5% significance level,
about the value of a population unbiased whether the dice is biased
parameter. You can test a
hypothesis to see if there is enough 1 Under what conditions would
!!: # =
evidence to change it 6 we reject the null hypothesis?
Hypothesis Testing
Null Hypothesis A researcher wants to test,
A Hypothesis is a statement made à The dice is at the 5% significance level,
about the value of a population unbiased whether the dice is biased
parameter. You can test a
hypothesis to see if there is enough 1 Under what conditions would
!!: # =
evidence to change it 6 we reject the null hypothesis?
Imagine we believe that a dice is Think about why we are rejecting the null hypothesis
biased towards landing on 6s.
7A
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
Null Hypothesis A researcher wants to test,
A Hypothesis is a statement made à The dice is at the 5% significance level,
about the value of a population unbiased whether the dice is biased
parameter. You can test a
hypothesis to see if there is enough 1 Under what conditions would
!!: # =
evidence to change it 6 we reject the null hypothesis?
7A
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
Null Hypothesis A researcher wants to test,
A Hypothesis is a statement made à The dice is at the 5% significance level,
about the value of a population unbiased whether the dice is biased
parameter. You can test a
hypothesis to see if there is enough 1 Under what conditions would
!!: # =
evidence to change it 6 we reject the null hypothesis?
7A
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A Hypothesis is a statement made
about the value of a population
parameter. You can test a
hypothesis to see if there is enough If we do not specify which way we
evidence to change it believe the coin to be biased, then our
answer to part c changes…
7A
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A Hypothesis is a statement made
about the value of a population
parameter. You can test a hypothesis
to see if there is enough evidence to
change it
Hypothesis Testing
% (the number of sixes from
Probability
A critical region is one which, if the 20 throws)
test statistic falls within it, would 0 0.026
cause you to reject the null 1
hypothesis 2 0.198
3
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null
hypothesis
% (the number of 5%
sixes from 20 Probability
throws)
0 0.026
1 0.104
2 0.198
3 0.238 Originally, we were considering whether the dice was
4 0.202 biased towards 6s, at the 5% significance level. This can
5 0.129
be represented on the diagram…
6 0.064
7 0.025
8 0.0084
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null
hypothesis
% (the number of 5%
sixes from 20 Probability
throws)
0 0.026
1 0.104
2 0.198
3 0.238 Note that we were only considering whether the dice was biased
4 0.202 towards sixes
5 0.129
à If we were just considering whether the dice was biased, we
6 0.064 could also include the value of 0 in the critical region
7 0.025
8 0.0084
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null
hypothesis
% (the number of
sixes from 20 Probability
throws)
0 0.026
1 0.104
2 0.198
3 0.238
4 0.202
5 0.129
6 0.064
7 0.025
8 0.0084
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
!! : # = 0.35
A critical region is one which, if the
test statistic falls within it, would Assuming !! is true, then '~)(6, 0.35)
cause you to reject the null
hypothesis à You can then use your calculator or the
statistical tables to find the value for which the
probability would be less than 5%
A single observation is taken from a
Binomial distribution )(6, #). The
observation is then used to test
!!: # = 0.35 against !&: # > 0.35.
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
!! : # = 0.35
A critical region is one which, if the
test statistic falls within it, would Assuming !! is true, then '~)(6, 0.35)
cause you to reject the null
hypothesis à You can then use your calculator or the
statistical tables to find the value for which the
probability would be less than 5%
A single observation is taken from a
Binomial distribution )(6, #). The
observation is then used to test
!!: # = 0.35 against !&: # > 0.35.
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null
hypothesis
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the The ‘actual significance level’ is the probability
test statistic falls within it, would of the test statistic falling within the critical
cause you to reject the null region
hypothesis
à It can also be thought of as ‘the probability
of incorrectly rejecting the null hypothesis’
A single observation is taken from a
Binomial distribution )(6, #). The
observation is then used to test
!!: # = 0.35 against !&: # > 0.35.
). )*%
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null hypothesis
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
It can help to summarise the information…
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null hypothesis
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
It can help to summarise the information…
A critical region is one which, if the !!: # = 0.25 + = 40 # = 0.25 2%
test statistic falls within it, would !&: # ≠ 0.25
cause you to reject the null hypothesis
, - ≤ 3 = 0.47%
Hypothesis Testing
A critical region is one which, if the So the critical regions are as follows:
test statistic falls within it, would
cause you to reject the null hypothesis
, - ≤ 3 = 0.47%
, - ≥ 17 = 1.16%
7B
If: '~)(+, #) + " #$"
. '=/ = # 1−#
/
Hypothesis Testing
A critical region is one which, if the
test statistic falls within it, would
cause you to reject the null hypothesis
, - ≤ 3 = 0.47%
, - ≥ 17 = 1.16%
b) State the actual significance level
of the test
7B
Hypothesis Testing
You need to be able to carry out a one-
tailed hypothesis test
Reminder:
This might be given to you as numbers
only or through a context A one-tailed test is where we are only
considering only one ‘end’ of the binomial
distribution.
For example:
7C
Hypothesis Testing
You need to be able to carry out a one-
tailed hypothesis test
7C
Hypothesis Testing
It can help to summarise the information…
You need to be able to carry out a one- !!: # = 0.4 + = 12 # = 0.4 5%
tailed hypothesis test
!&: # > 0.4 4=8
7C
Hypothesis Testing
You need to be able to carry out a one-
tailed hypothesis test
7C
Hypothesis Testing
Summarise the information…
You need to be able to carry out a one- '~)(20, #)
tailed hypothesis test + = 20 # = 0.25
!!: # = 0.25
!&: # > 0.25 4 = 10 5%
The standard treatment for a
&
particular disease has a probability
'
of success. A certain doctor has
undertaken research in this area and
has produced a new drug which has
been successful with 10 out of 20
patients. The doctor claims that the
new drug represents an improvement
on the standard treatment. Test, at
the 5% significance level, the claim Write out the calculation we need to do…
made by the doctor.
7C
Hypothesis Testing
You also need to be able to carry
out a two-tailed hypothesis test
7D
Hypothesis Testing
Summarise the information…
You also need to be able to carry out a
two-tailed hypothesis test
7D
Hypothesis Testing
Summarise the information…
You also need to be able to carry out a
two-tailed hypothesis test '~)(10, #)
!!: # = 0.45 + = 10 # = 0.45
A single observation, !, is taken from
a Binomial distribution "~$ 10, ( , and !&: # ≠ 0.45 4=1 2.5%
a value of 1 is obtained. Use this
observation to test )! : ( = 0.45
against )" : ( ≠ 0.45 using a 5%
significance level.
7D
Hypothesis Testing
You also need to be able to carry
out a two-tailed hypothesis test
7D
Hypothesis Testing
You also need to be able to carry In the previous question, we needed to find
out a two-tailed hypothesis test .(' ≤ 1) when + = 10 and # = 0.45
7D
Hypothesis Testing
Start by pressing the ‘mode’ button
You also need to be able to carry
out a two-tailed hypothesis test
On this screen, press 7
1. A test statistic has a distribution X ~ B(9, p). Given that H0: p = 0.25, H1: p > 0.25, find the critical
region for the test using a 5% significance level. [4]
2. A random variable has distribution X ~ B(20, p). A single observation of x = 2 is taken from this
distribution. Test, at the 10% significance level, H0: p = 0.15 against H1: p 0.15 . [4]
3. An article states that 45% of drivers in town X drive a black car. A researcher wants to test, at the
10% significance level, whether the article is overestimating the number of black-car drivers. The
researcher asks 40 drivers what colour their car is. Thirteen people say black.
a) Write down a suitable test statistic. [1]
b) Write down a suitable null hypothesis and a suitable alternative hypothesis. [2]
c) Explain the condition under which the null hypothesis would be rejected. [1]
4. On average, a machine fails 4 times out of 10. An engineer designs a new machine that he believes
has a reduced failure rate. He uses his new machine 15 times in order to test his belief.
a) Describe the test statistic. [1]
b) State suitable null and alternative hypotheses. [2]
c) Using a 10% level of significance, find the critical region for a test to check the engineer’s
belief, ensuring the probability is as close to 0.1 as possible. [4]
d) Write down the actual significance level of the test. [1]
5. It is claimed that 10% of women use a particular perfume called ‘Daisy’. In a random survey of 50
women, 41 said they do not use this perfume.
Test, at the 5% significance level, whether or not there is evidence that the proportion of women
using the ‘Daisy’ perfume is 0.1. State your hypotheses carefully. [7]
6. A doctor claims that 80% of patients suffering from a certain illness recover when they are treated
with a new medicine.
A random sample of 25 patients with this illness is taken from hospital records.
a) Write down a suitable distribution to model the number of patients in this sample who recover
when given the new medicine. [1]
b) Assuming that the claim is correct, find the probability that the medicine will be successful for
exactly 19 patients. [2]
The hospital believes that the doctor’s claim is incorrect and the percentage who will recover is
lower. A random sample of 40 patients with the illness who had been prescribed the medicine is
taken from the hospital records. It is found that of these 40 patients, 26 had recovered.
c) Stating your hypotheses clearly, test, at the 2% level of significance, the hospital’s belief. [6]
TOTAL 36 MARKS
1A
! = #$ ! %&'! = %&'# + )%&'$ ! = #* " %&'! = %&'# + $%&'*
Regression, correlation and
hypothesis testing
You need to be able to use ! = #* "
logarithms and coding to analyse
trends in non-linear data
%&'! = %&'#* "
àYou have seen this in the Pure %&'! = %&'# + %&'* "
Year 1 course, chapter 14
%&'! = %&'# + $%&'*
àLet’s have a reminder of how
the relationships should be
written…
1A
! = #$ ! %&'! = %&'# + )%&'$ ! = #* " %&'! = %&'# + $%&'*
Regression, correlation and
hypothesis testing
You need to be able to use ! 3 5 6 8 9 11
logarithms and coding to analyse
" 1.04 1.49 1.79 2.58 3.1 4.46
trends in non-linear data
The table to the right shows some data à Using the coding, when % = 0, # = 0 as well…
collected on the temperature, in °C, of a
colony of bacteria (t), and its growth à Substitute # = 0 into the equation…
rate (g).
1A
Regression, correlation and
hypothesis testing
You need to be able to calculate and
use the product moment correlation
coefficient (PMCC)
1B
Regression, correlation and
hypothesis testing
Day of
You need to be able to calculate and month
1 2 3 4 5 6 7 8 9 10
use the product moment correlation
coefficient (PMCC) ! 4 4 8 7 12 12 3 4 7 10
g 13 12 19 23 33 37 10 n/a n/a 23
From the large data set, the daily mean
windspeed, ! knots, and the daily maximum gust, à The n/a in the table indicates that no data is
" knots, were recorded for the first 10 days in available on those days
September in Hurn in 1987.
1B
Regression, correlation and
hypothesis testing
Day of
You need to be able to calculate and month
1 2 3 4 5 6 7 8 9 10
use the product moment correlation
coefficient (PMCC) ! 4 4 8 7 12 12 3 4 7 10
g 13 12 19 23 33 37 10 n/a n/a 23
From the large data set, the daily mean
windspeed, ! knots, and the daily maximum gust, à Now press 2, since we want a linear regression
" knots, were recorded for the first 10 days in
September in Hurn in 1987.
b) Calculate the product moment correlation à Now enter the data (remember to ignore the
coefficient for the remaining 8 days n/a values from this question…)
à On your casio classwiz, press menu
à Now press 6
à Now press 4
1B
Regression, correlation and
hypothesis testing
Day of
You need to be able to calculate and month
1 2 3 4 5 6 7 8 9 10
use the product moment correlation
coefficient (PMCC) ! 4 4 8 7 12 12 3 4 7 10
g 13 12 19 23 33 37 10 n/a n/a 23
From the large data set, the daily mean
windspeed, ! knots, and the daily maximum gust, à Now press 4
" knots, were recorded for the first 10 days in
September in Hurn in 1987.
1B
Regression, correlation and
hypothesis testing
Day of
You need to be able to calculate and month
1 2 3 4 5 6 7 8 9 10
use the product moment correlation
coefficient (PMCC) ! 4 4 8 7 12 12 3 4 7 10
g 13 12 19 23 33 37 10 n/a n/a 23
From the large data set, the daily mean
windspeed, ! knots, and the daily maximum gust,
" knots, were recorded for the first 10 days in
September in Hurn in 1987.
1B
Regression, correlation and
hypothesis testing
You need to be able to perform a à When testing if the population PMCC, p, is either
hypothesis test to determine whether a greater than or below zero, you need to use a one-
data set has no correlation tailed test
An important note is that in the à When testing whether it is not equal to 0, you should
previous section, we used # to denote use a two-tailed test…
the PMCC
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a +! : $ = 0 +" : $ ≠ 0 Sample size = 30
hypothesis test to determine whether a Significance level in each tail: 0.05
data set has no correlation
Finding the critical region
A scientist takes 30 observations of the
masses of two reactants in an
experiment. She calculates a PMCC of
# = −0.45.
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a
hypothesis test to determine whether a
data set has no correlation
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a
hypothesis test to determine whether a
data set has no correlation
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a +! : $ = 0 +" : $ ≠ 0 Sample size = 30
hypothesis test to determine whether a Significance level in each tail: 0.05
data set has no correlation
Finding the critical region
A scientist takes 30 observations of the
masses of two reactants in an
experiment. She calculates a PMCC of
# = −0.45.
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a . 31 28 38 37 18 17 21 29
hypothesis test to determine whether a
data set has no correlation / 99 94 87 80 80 89 84 86
1C
Regression, correlation and
hypothesis testing
You need to be able to perform a . 31 28 38 37 18 17 21 29
hypothesis test to determine whether a
data set has no correlation / 99 94 87 80 80 89 84 86
1C