nhypa 1

A hypothesis is a claim (assumption) about a population parameter

 Example: The mean monthly cell phone bill of NSU student is μ = Tk 2000.00
 Example: The proportion of adults in this city with cell phones is p = 0.68.

Null Hypothesis H : A maintained hypothesis that is held to be true unless sufficient evidence to the contrary is
0

obtained. Null hypothesis: Always contains “=” , “≤” or “” sign, it may or may not be rejected

Alternative Hypothesis H : A hypothesis against which the null hypothesis is tested and which will be held to
a

be true if the null is held false.

 Alternative hypothesis: is the opposite of the null hypothesis, e.g., The average number of TV sets in BD
homes is not equal to 1 ( H1 or H a : μ ≠ 1 ), it never contains the “=” , “≤” or “” sign, is generally the
hypothesis that the researcher is trying to support

p-value: Probability of obtaining a test statistic more extreme ( ≤ or  ) than the observed sample value
given H 0 is true

Simple Hypothesis: A hypothesis that specifies a single value for a population parameter of interest.
Composite Hypothesis: A hypothesis that specifies a range of values for a population parameter.
nhypa 2

One-Sided Alternative: An alternative hypothesis involving all possible values of a population parameter on
either one side or the other of (that is, either greater than or less than) the value specified by a simple null
hypothesis.
Two-Sided Alternative: An alternative hypothesis involving all possible values of a population parameter other
than the value specified by a simple null hypothesis.

Hypothesis Test Decisions: A decision rule is formulated, leading the investigator to either reject or fail to reject
the null hypothesis on the basis of sample evidence.

Type I Error: The rejection of a true null hypothesis.

Type II Error: The failure to reject a false null hypothesis.

Significance Level: The probability of rejecting a null hypothesis that is true. This probability is sometimes
expressed as a percentage, so a test of significance level a is referred to as a 100% level test.

Power: The probability of rejecting a null hypothesis that is false.

Critical value: The boundary point(s) between the acceptance region and critical region.
nhypa 3

Summarized Steps in Hypothesis Testing

1. Specify the Null Hypothesis and the Alternative Hypothesis.
2. Set the Significance Level ( )
3. Specify the Test Statistic.
4. Use the level of significance and alternative hypothesis to develop the rejection rule
5. Calculate the Test Statistic and the Corresponding p-value.
6. Draw Conclusion.
nhypa 4

Testing single mean

Null hypothesis: H :   
0 0

Case 1.  unknown, n  30, and population normal

n  x  0 
Test statistic: t   t (n  1)
s

Case 2.  is known and population is normal,

n  x  0 
Test statistic: z ~ N (0,1).

Case 3.  unknown and n  30,
n  x  0 
Test statistic: z ~ N (0,1).
s
nhypa 5

Testing about proportion:

Null hypothesis: H 0 : p  p0
pˆ  p0
z ~ N (0,1).
Test statistic p0 (1  p0 ) (for large n).
n
nhypa 6

CONFIDENCE INTERVAL

(1   )100% t-Confidence interval for population mean  is

 s s 
 x  t /2 ( n  1) , x  t /2 ( n  1) 
 n n
95% t-Confidence interval for population mean  is
 s s 
 x  t 0.025 ( n  1) , x  t 0.025 ( n  1) 
 n n 
-------------------------------------------------------------------------------
(1   )100% z-Confidence interval for population mean  is
   
 x  z /2 , x  z /2 
 n n 
95% z-Confidence interval for population mean  is
   
 x  z 0.025 , x  z 0.025 
 n n
 Chapter 8: Inferences on Population mean

8.1.1 A sample of 31 data observations has a sample mean   x = 53.42 and a sample standard
deviation s = 3.05. Construct a 95% two-sided t-interval for the population mean. (This problem is
continued in Problem 8.1.9.)

Solution:
(1 − α )100% = α 0.95;=
95% , that is 1 −= α 0.05;
= α
2 0.025 .

Thus, t=
α ,( n −1) = 2.042
t0.025,30 (using student’s t table),
2

the 95% confidence interval for population mean, µ is

(x −t ,x +t
s
0.025,30 n ) s
0.025,30 n

= ( 53.42 − 2.042 × ,53.42 + 2.042 × ) = (52.30, 54.54).

3.05 3.05
31 31

AjM_Summer_2020 1
8.1.2 A random sample of 41 glass sheets is obtained and their thicknesses are measured. The sample
mean is x = 3.04 mm and the sample standard deviation is s =0.124 mm. Construct a 99% two-sided
t-interval for the mean glass thickness. Do you think it is plausible that the mean glass thickness is
2.90 mm? (This problem is continued in Problems 8.1.10 and 8.6.12.)

Solution: (1 − α )100% = α 0.99;=

99% , that is 1 −= α 0.01;
= α
2 0.005 .
Thus, t=
α ,( n −1)
2
= 2.704 (Student’s t table), the
t0.005,40
99% confidence interval for mean glass thickness is

(3.04 − 2.704 × 0.124

41 )
,3.04 + 2.704 × 0.124
41 = (2.99, 3.09).

The confidence interval does not contain the value 2.90. Thus, .90 is not a plausible value
for the mean glass thickness.

AjM_Summer_2020 2
8.1.3 The breaking strengths of a random sample of 20 bundles of wool fibers have a sample mean x = 436.5 and
a sample standard deviation s = 11.90 .
Construct 90%, 95%, and 99% two-sided t-intervals for the average breaking strength μ.
Compare the lengths of the confidence intervals.
Do you think it is plausible that the average breaking strength is equal to 450.0? (This problem is continued in
Problems 8.1.11 and 8.6.13.)
Solution: At 90% confidence the critical point is t0.05,19 = 1.729 and the confidence interval is

( 436.5 − 1.729 × 11.90

20
, 436.5 + 1.729 × 11.90
20 ) = (431.9, 441.1).
At 95% confidence the critical point is t0.025,19 = 2.093 and the confidence interval is

( 436.5 − 2.093 × 11.90

20
, 436.5 + 2.093 × 11.90
20 ) = (430.9, 442.1).
At 99% confidence the critical point is t0.005,19 = 2.861 and the confidence interval is

( 436.5 − 2.861× 11.90

20
, 436.5 + 2.861× 11.90
20 ) = (428.9, 444.1).
Even the 99% confidence level confidence interval does not contain the value 450.0, and so 450.0 is not
a plausible value for the average breaking strength.

AjM_Summer_2020 3
8.1.4 A random sample of 16 one-kilogram sugar packets is obtained and the actual weights of the
packets are measured. The sample mean is x = 1.053 kg and the sample standard deviation is s =
0.058 kg.
Construct a 99% two-sided t-interval for the average sugar packet weight.
Do you think it is plausible that the average weight is 1.025 kg? (This problem is continued in Problem
8.6.14.)

Solution: (1 − α )100% = α 0.99;=

99% , that is 1 −= α 0.01;
= α
2 0.005 .
Thus, t=
α ,( n −1)
2
= 2.947 and the 99% confidence interval is
t0.005,15

(1.053 − 2.947 × 0.058

16 )
,1.053 + 2.947 × 0.058
16 = (1.010,1.096)

The confidence interval contains the value 1.025, and so 1.025 is a plausible value for the
average weight.

AjM_Summer_2020 4
8.1.5 A sample of 28 data observations has a sample mean x = 0.0328. If an experimenter wishes to

use a “known” value σ = 0.015 for the population standard deviation, construct an appropriate 95%
two-sided confidence interval for the population mean μ.

Solution: With z0.025 = 1.960 the confidence interval is

( x − z0.025,30 σ
n
, x + z0.025,30 σ
n ) = (0.0272, 0.0384).

AjM_Summer_2020 5
8.2.1 A sample of n = 18 observations has a sample mean of x = 57.74 and a sample standard

deviation of s = 11.20. Consider the hypothesis testing problems:

(a) H 0 : µ = 55.0 versus H A : µ ≠ 55.0
(b) H 0 : µ ≥ 65.0 versus H A : µ < 65.0
In each case, write down an expression for the p-value. What do the critical points in Table III tell you
about the p-values? Use a computer package to evaluate the p-values exactly.
n (x − µ)
Solution: Under H the test statistic
0 = is t ~ t (n − 1) .
s
18 × ( 57.74 − 55.0 )
(a) t = 1.04
11.2
The p-value is 2 × P (t17 > 104) = 0.313
18 × ( 57.74 − 65.0 )
(b) t = = −2.75
11.2
The p-value is P (t17 ≤ −2.75) =0.0068

AjM_Summer_2020 6
8.2.2 A sample of n = 39 observations has a sample mean of x = 5532 and a sample standard deviation

of s = 287.8. Consider the hypothesis testing problems:

(a) H 0 : µ = 5680 versus H A : µ = 5680
(b) H 0 : µ ≤ 5450 versus H A : µ > 5450
In each case, write down an expression for the p-value.
What do the critical points in Table III tell you about the p-values? Use a computer package to evaluate
the p-values exactly.
n (x − µ)
=
Solution: Under H 0 the test statistic is t ~ t (n − 1) .
s
39 × ( 5532 − 5680 )
(a) t= = −3.21
287.8
The p-value is 2 × P (t38 ≤ −3.21) = 2 × P (t38 > 3.21) = 0.003
39 × ( 5532 − 55450 )
(b) t = 1.78
287.8
The p-value is P (t38 ≥ 1.78) =
0.042

AjM_Summer_2020 7
8.2.3 A sample of n = 13 observations has a sample mean of x = 2.879 . If an assumed known

standard deviation of σ = 0.325 is used, calculate the p-values for the hypothesis testing problems:

(a) H0 : µ = 3.0 versus HA : µ = 3.0

(b) H 0 : µ ≥ 3.1 versus H A : µ < 3.1
n (x − µ)
Solution: Under H0 the test statistic is z = ~ N (0,1)
σ
13 ( 2.879 − 3)
(a) z= = −1.34
0.325
The p-value is P ( z ≤ −1.34) = Φ (−1.34) =
0.180 .
13 ( 2.879 − 3.1)
(b) z= = −2.45
0.325
The p-value is P ( z ≤ −2.45) = Φ (−2.45) =
0.007 .

AjM_Summer_2020 8
 8.2.4 A sample of n = 44 observations has a sample mean of x = 87.90 . If an assumed known standard

deviation of σ = 5.90 is used, calculate the p-values for the hypothesis testing problems:

(a) H 0 : µ = 90.0 versus H A : µ = 90.0

(b) H 0 : µ ≤ 86.0 versus H A : µ > 86.0

n (x − µ)
Solution: Under H0 the test statistic is z = ~ N (0,1)
σ
44 ( 87.90 − 90.0 )
(a) z= = −2.36
5.90
The p-value is 2 × P ( z ≤ −2.36) = Φ (−2.36) =
0.018 .
44 ( 87.90 − 86.0 )
=
(b) z = 2.14
5.90
The p-value is P ( z > 2.14) =
1 − P ( z ≤ 2.14) =
1 − Φ (2.14) =
0.016 .

AjM_Summer_2020 9
8.2.5 An experimenter is interested in the hypothesis testing problem
H 0 : µ = 3.0 versus H A : µ = 3.0
where μ is the average thickness of a set of glass sheets in mm. Suppose that a sample of n = 41 glass
sheets is obtained and their thicknesses are measured.
(a) For what values of the t-statistic does the experimenter accept the null hypothesis with a size α = 0.10?

(b) For what values of the t-statistic does the experimenter reject the null hypothesis with a size α = 0.01?
Solution: (a) The critical point is t0.05,40 = 1.684 and
the null hypothesis is accepted when | t |≤ 1.684.
(b) The critical point is t0.005,40 = 2.704 and
the null hypothesis is rejected when | t |> 2.704

AjM_Summer_2020 10
Suppose that the sample mean is x = 3.04 mm and the sample standard deviation is s = 0.124 mm.

(c) Is the null hypothesis accepted or rejected with α = 0.10? With α = 0.01?
(d) Write down an expression for the p-value and evaluate it using a computer package.
Solution:

n (x − µ)
(c) Under H0 the =
test statistic is t ~ t (n − 1) .
s
41 × (3.04 − 3.00)
t= =2.006
0.124
The null hypothesis is rejected at size α = 0.10 and
accepted at size α = 0.01

(d) The p-value is 2 × P (t40 ≥ 2.066) =

0.045.

AjM_Summer_2020 11
8.2.7 An experimenter is interested in the hypothesis testing problem
H 0 : µ = 1.025 kg versus H A : µ = 1.025 kg

where μ is the average weight of a 1-kilogram sugar packet. Suppose that a sample of n = 16 sugar
packets is obtained and their weights are measured.
(a) For what values of the t-statistic does the experimenter accept the null hypothesis with a size α =
0.10?
(b) For what values of the t-statistic does the experimenter reject the null hypothesis with a size α =
0.01?

(a) The critical point is t0.05,15 = 1.753

and the null hypothesis is accepted when | t |≤ 1.753 .
(b) The critical point is t0.005,15 = 2.947
and the null hypothesis is rejected when |t| > 2.947.

AjM_Summer_2020 12
Suppose that the sample mean is x = 1.053 kg and the sample standard deviation is s = 0.058 kg.

(c) Is the null hypothesis accepted or rejected with α = 0.10? With α = 0.01?
(d) Write down an expression for the p-value and evaluate it using a computer package.
n (x − µ)
=
(c) Under H 0 the test statistic is t ~ t (n − 1) .
s
16 × (1.053 − 1.931)
t= =1.931
0.058
The null hypothesis is rejected at size α = 0.10 and
accepted at size α = 0.01.

(d) The p-value is 2 × P (t15 ≥ 1.931) =

0.073

AjM_Summer_2020 13
8.2.9 An experimenter is interested in the hypothesis testing problem
H 0 : µ ≤ 0.065 versus H A : µ > 0.065
where μ is the average density of a chemical solution. Suppose that a sample of n = 61 bottles of the
chemical solution is obtained and their densities are measured.
(a) For what values of the t-statistic does the experimenter accept the null hypothesis with a size α = 0.10?

(b) For what values of the t-statistic does the experimenter reject the null hypothesis with a size α = 0.01?
Solution:
(a) The critical point is t0.10,60 = 1.296
and the null hypothesis is accepted when t ≤ 1.296.
(b) The critical point is t0.01,60 = 2.390
and the null hypothesis is rejected when t > 2.390.

AjM_Summer_2020 14
8.2.9 (cont.) Suppose that the sample mean is x = 0.0768 and the sample standard deviation is s =
0.0231.
(c) Is the null hypothesis accepted or rejected with α = 0.10? With α = 0.01?
(d) Write down an expression for the p-value and evaluate it using a computer package.

Solution:
n (x − µ)
(c) Under H0 =
the test statistic is t ~ t (n − 1) .
s
61 × (0.0768 − 0.065)
t = 3.990
0.0231

The null hypothesis is rejected at size α = 0.01

and consequently also at size α = 0.10.

(d) The p-value is P (t60 ≥ 3.990) =

0.0001 .

AjM_Summer_2020 15
8.2.11 A machine is set to cut metal plates to a length of 44.350 mm. The lengths of a random sample
of 24 metal plates have a sample mean of x = 44.364 mm and a sample standard deviation of s =
0.019 mm. Is there any evidence that the machine is miscalibrated?

Soilution:
Consider the hypotheses H 0 : µ = 44.350 versus
H A : µ 6 ≠ 44.350.
n (x − µ)
Under H0 =
the test statistic is t ~ t (n − 1) .
s
24 × (44.364 − 44.350)
t = 3.61
0.019

The p-value is 2 × P (t23 ≥ 3.61) =

0.0014 .

There is sufficient evidence to conclude that the machine is miscalibrated.

AjM_Summer_2020 16
8.2.12 A food manufacturer claims that at the time of purchase by a consumer the average age of its
product is no more than 120 days. In an experiment to test this claim a random sample of 36 items are
found to have ages at the time of purchase with a sample mean of x = 122.5 days and a sample
standard deviation of s = 13.4 days. With this information how do you feel about the manufacturer’s
claim?

Solution:

Consider the hypotheses H 0 : µ ≤ 120 versus

H A : µ > 120
n (x − µ)
Under H0 =
the test statistic is t ~ t (n − 1) .
s
36 × (122.5 − 120)
=t = 1.12
13.4
The p-value is P (t35 ≥ 1.12) =0.135 .
There is not sufficient evidence to conclude that the manufacturer’s claim is incorrect.

AjM_Summer_2020 17
AjM_Summer_2020 18
8.2.13 A chemical plant is required to maintain ambient sulfur levels in the working environment
atmosphere at an average level of no more than 12.50. The results of 15 randomly timed measurements
of the sulfur level produced a sample mean of x = 14.82 and a sample standard deviation of s = 2.91.
What is the evidence that the chemical plant is in violation of the working code?
Solution:
Consider the hypotheses H 0 : µ ≤ 12.50 versus
H A : µ > 12.50
n (x − µ)
Under H0 =
the test statistic is t ~ t (n − 1) .
s
15 × (14.82 − 12.50)
t = 3.09
2.91
The p-value is P (t14 ≥ 3.09) =
0.004 .

There is sufficient evidence to conclude that the chemical plant is in violation of the working
code.

AjM_Summer_2020 19
 9.3 ANALYSIS OF INDEPENDENT SAMPLES 409

Two-Sample t-Procedure (Equal Variances)

Consider a sample of size n from population A with a sample mean x̄ and a sample
standard deviation sx , and a sample of size m from population B with a sample mean ȳ
and a sample standard deviation s y . If an experimenter assumes that the population
variances σ A2 and σ B2 are equal, then the common variance can be estimated by
(n − 1)sx2 + (m − 1)s y2
s 2p =
n+m−2
which is known as the pooled variance estimate.
A two-sided 1 − α level confidence interval for the difference in population means
μ A − μ B is
  
1 1 1 1
μ A − μ B ∈ x̄ − ȳ − tα/2,n+m−2 s p + , x̄ − ȳ + tα/2,n+m−2 s p +
n m n m
One-sided confidence intervals are
 
1 1
μ A − μB ∈ −∞, x̄ − ȳ + tα,n+m−2 s p +
n m
and
 
1 1
μ A − μB ∈ x̄ − ȳ − tα,n+m−2 s p + ,∞
n m
The appropriate t-statistic for the null hypothesis H0 : μ A − μ B = δ is
x̄ − ȳ − δ
t=
s p n1 + m1
A two-sided p-value is calculated as 2 × P(X > |t|), where the random variable X
has a t-distribution with n + m − 2 degrees of freedom, and one-sided p-values
are P(X > t) and P(X < t). A size α two-sided hypothesis test accepts the null
hypothesis if
|t| ≤ tα/2,n+m−2
and rejects the null hypothesis when
|t| > tα/2,n+m−2
and size α one-sided hypothesis tests have rejection regions t > tα,n+m−2 or
t < −tα,n+m−2 .
These procedures are known as two-sample t-tests with a pooled variance
estimate.

These results are seen to match well with the corresponding analyses conducted previously
using the general procedure, which does not require the equality of the population variances.
When should an experimenter use this method with a pooled variance estimate, and when
should the general procedure be employed that does not require the equality of the population
variances? The safest answer is to always use the general procedure since it provides a valid
analysis even when the population variances are equal. However, if the population variances are
410 CHAPTER 9 COMPARING TWO POPULATION MEANS

equal or quite similar, then the pooled variance procedure generally provides a slightly more
powerful analysis with slightly shorter confidence intervals and slightly smaller p-values.
An experimenter’s assessment of whether or not the population variances can be taken to
be equal may be based on prior experience with the kind of data under consideration, or may
be based on a comparison of the sample standard deviations sx and s y . A formal test for the
equality of the population variances σ A2 and σ B2 is described in the Supplementary Problems
section at the end of this chapter. However, this test has some drawbacks and its employment
is not widely recommended.
In general, analyses performed with and without the use of a pooled variance estimate will
often provide similar results. In fact, if the results are quite different, then this will be because
the sample standard deviations sx and s y are quite different, in which case it is proper to use
the general procedure that does not require the equality of the population variances.

9.3.3 z-Procedure

Two-Sample z-Procedure
Consider a sample of size n from population A with a sample mean x̄, and a sample of
size m from population B with a sample mean ȳ. Suppose that the population
variances are assumed to take known values σ A2 and σ B2 .
A two-sided 1 − α level confidence interval for the difference in population means
μ A − μ B is
⎛   ⎞
σ 2
σ 2
σ 2
σ 2
μ A − μ B ∈ ⎝x̄ − ȳ − z α/2 A
+ B , x̄ − ȳ + z α/2 A
+ B⎠
n m n m

One-sided confidence intervals are

⎛  ⎞
σ A2 σ B2 ⎠
μ A − μ B ∈ ⎝−∞, x̄ − ȳ + z α +
n m

and
⎛  ⎞
σ A2 σ B2
μ A − μ B ∈ ⎝x̄ − ȳ − z α + , ∞⎠
n m

The appropriate z-statistic for the null hypothesis H0 : μ A − μ B = δ is

x̄ − ȳ − δ
z=
σ A2 σ2
n
+ mB
A two-sided p-value is calculated as 2 × (−|z|), and one-sided p-values are
1 − (z) and (z). A size α two-sided hypothesis test accepts the null hypothesis if
|z| ≤ z α/2
and rejects the null hypothesis when
|z| > z α/2
and size α one-sided hypothesis tests have rejection regions z > z α or z < −z α .
These procedures are known as two-sample z-tests.
 Paired t-tests

1 Introduction
A paired t-test is used to compare two population means where you have two samples in
which observations in one sample can be paired with observations in the other sample.
Examples of where this might occur are:

• Before-and-after observations on the same subjects (e.g. students’ diagnostic test

results before and after a particular module or course).

• A comparison of two different methods of measurement or two different treatments

where the measurements/treatments are applied to the same subjects (e.g. blood
pressure measurements using a stethoscope and a dynamap).

2 Procedure for carrying out a paired t-test

Suppose a sample of n students were given a diagnostic test before studying a particular
module and then again after completing the module. We want to find out if, in general,
our teaching leads to improvements in students’ knowledge/skills (i.e. test scores). We
can use the results from our sample of students to draw conclusions about the impact of
this module in general.
Let x = test score before the module, y = test score after the module
To test the null hypothesis that the true mean difference is zero, the procedure is as
follows:

1. Calculate the difference (di = yi − xi ) between the two observations on each pair,
making sure you distinguish between positive and negative differences.
¯
2. Calculate the mean difference, d.

3. Calculate the standard deviation of the differences, sd , and use this to calculate the
standard error of the mean difference, SE(d)¯ =√ sd
n

4. Calculate the t-statistic, which is given by T = d¯ . Under the null hypothesis,

¯
SE(d)
this statistic follows a t-distribution with n − 1 degrees of freedom.

5. Use tables of the t-distribution to compare your value for T to the tn−1 distribution.
This will give the p-value for the paired t-test.

1
NOTE:
For this test to be valid the differences only need to be approximately normally distributed.
Therefore, it would not be advisable to use a paired t-test where there were any extreme
outliers.

Example
Using the above example with n = 20 students, the following results were obtained:
Student Pre-module Post-module Difference
score score
1 18 22 +4
2 21 25 +4
3 16 17 +1
4 22 24 +2
5 19 16 -3
6 24 29 +5
7 17 20 +3
8 21 23 +2
9 23 19 -4
10 18 20 +2
11 14 15 +1
12 16 15 -1
13 16 18 +2
14 19 26 +7
15 18 18 0
16 20 24 +4
17 12 18 +6
18 22 25 +3
19 15 19 +4
20 17 16 -1

Calculating the mean and standard deviation of the differences gives:

d¯ = 2.05 and sd = 2.837. Therefore, SE(d)
¯ =√ sd = 2.837
√ = 0.634
n 20
So, we have:
2.05
t= = 3.231 on 19 df
0.634
Looking this up in tables gives p = 0.004. Therefore, there is strong evidence that, on
average, the module does lead to improvements.

2
3 Confidence interval for the true mean difference
The in above example the estimated average improvement is just over 2 points. Note that
although this is statistically significant, it is actually quite a small increase. It would be
useful to calculate a confidence interval for the mean difference to tell us within what limits
the true difference is likely to lie. A 95% confidence interval for the true mean difference is:
sd
d¯ ± t∗ √ or, equivalently d¯ ± (t∗ × SE(d))
¯
n
where t∗ is the 2.5% point of the t-distribution on n − 1 degrees of freedom.

Using our example:

We have a mean difference of 2.05. The 2.5% point of the t-distribution with 19 degrees
of freedom is 2.093. The 95% confidence interval for the true mean difference is therefore:

2.05 ± (2.093 × 0.634) = 2.05 ± 1.33 = (0.72, 3.38)

This confirms that, although the difference in scores is statistically significant, it is actually
relatively small. We can be 95% sure that the true mean increase lies somewhere between
just under one point and just over 3 points.

4 Carrying out a paired t-test in SPSS

The simplest way to carry out a paired t-test in SPSS is to compute the differences (using
Transform, Compute) and then carrying out a one-sample t-test as follows:
— Analyze
— Compare Means
— One-Sample T Test
— Choose the difference variable as the Test Variable and click OK

The output will look like this:

One-Sample Statistics
N Mean Std. Deviation Std. Error Mean
Difference 20 2.0500 2.83725 .63443

One-Sample Test
Test Value = 0
95% Confidence Interval
t df Sig. (2-tailed) Mean Difference Lower Upper
Difference 3.231 19 .004 2.0500 .7221 3.3779