Test for significance: z-tests and

one-sample t-tests
Chapter 8
Case Study: How do we compare with
national averages?
Grace is the manager of a small company that provides tutoring to primary school
students called Prescription for Success.
Her goal is to produce a pamphlet with ‘scientific evidence’ that her tutors are
superior to all others in the area. Part of her organization’s mission is to connect ‘kind,
compassionate and caring tutors’ with students in need. So she decides to assess her
employees’ kindness, compassion and care for children, with the goal of comparing
these scores to those in her area. After consulting with a research psychologist, she
identifies three surveys that can be used to assess the three traits she believes to be
strong in her tutors: the MacMillen Kindness Inventory, the Cincinnati Index of
Compassion and the Child Focus Survey. Each produces a score on a scale of 0 to 100,
with 100 indicating maximum kindness, compassion and child focus, respectively.
She plans to administer the three surveys to her 27 tutors. She wants to draw
conclusions about the kindness, compassion, and care for children of her tutors in
general, believing that they are better than those of her competitors.
RQ1/2: Are the tutors at Prescription for Success kinder/more compassionate than
the competition?
RQ3: Do the tutors at Prescription for Success care for children more than the
competition?
One-sample tests
 One-sample test: a hypothesis test comparing a sample
mean to a given population mean
 Example: z-tests and one-sample t-tests
 Research questions for one-sample tests follow the same
general format:
 Does this sample differ from the known population?
 Hypotheses for one-sample tests use the format:
 Two-tailed test where the population mean = 100
 H0: μ = 100
 H1: μ ≠ 100
 One-tailed test where the population mean = 100
 H0: μ ≤ 100
 H1: μ > 100
 RQ1: Are the tutors at Prescription for Success kinder than
the competition?
 H0: μ ≤ 45
 H1: μ > 45
 RQ2: Are the tutors at Prescription for Success more
compassionate than the competition?
 H0: μ ≤ 55
 H1: μ > 55
 RQ3: Do the tutors at Prescription for Success care for
children more than the competition?
 H0: μ ≤ 67
 H1: μ > 67
 Critical values and decision rules: To determine your decision
rule, you first need to identify the type of test
z-tests and one-sample t-tests
 z-test: a one-sample test where a sample mean is
compared to a given population mean as stated in the null
hypothesis, when the population standard deviation is
known
 One-sample t-test: a one-sample test where a sample
mean is compared to a given population mean as stated in
the null hypothesis, when the population standard
deviation is not known
How many standard errors away from the population
mean a sample mean needs to be for us to reject H0?

11-7
 z 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

1 0.1587 0.1562 0.1539 0.1515 0.1492 0.1469 0.1446 0.1423 0.1401 0.1379

1.1 0.1357 0.1335 0.1314 0.1292 0.1271 0.1251 0.1230 0.1210 0.1190 0.1170

1.2 0.1151 0.1131 0.1112 0.1093 0.1075 0.1056 0.1038 0.1020 0.1003 0.0985

1.3 0.0968 0.0951 0.0934 0.0918 0.0901 0.0885 0.0869 0.0853 0.0838 0.0823

1.4 0.0808 0.0793 0.0778 0.0764 0.0749 0.0735 0.0721 0.0708 0.0694 0.0681

1.5 0.0668 0.0655 0.0643 0.0630 0.0618 0.0606 0.0594 0.0582 0.0571 0.0559

1.6 0.0548 0.0537 0.0526 0.0516 0.0505 0.0495 0.0485 0.0475 0.0465 0.0455

1.7 0.0446 0.0436 0.0427 0.0418 0.0409 0.0401 0.0392 0.0384 0.0375 0.0367

1.8 0.0359 0.0351 0.0344 0.0336 0.0329 0.0322 0.0314 0.0307 0.0301 0.0294

1.9 0.0287 0.0281 0.0274 0.0268 0.0262 0.0256 0.0250 0.0244 0.0239 0.0233
 Assuming α = .05:
 The critical value for a one-tailed z-test is always + or - 1.645
 Example: zcrit = +1.645
z-tests: Critical values and decision rules
 Assuming α = .05:
 The critical value for a two-tailed z-test is always ± 1.96
 Example: zcrit = ± 1.96
Common critical values of the z-test

α = .05 α = .01
Two-tailed z-test ±1.96 ±2.58
Left-tailed z-test -1.645 -2.33
Right-tailed z-test +1.645 +2.33
One-sample t-tests: Critical values and
decision rules
 The critical value for any t-test depends on alpha, the
degrees of freedom for the test, and whether the test is
one-tailed or two tailed
 Example: For a two-tailed t-test where α = .05 and n = 25,
tcrit(24) = 2.064
 Confidence intervals
90% 95% 98% 99%
2-tailed α
.10 .05 .02 .01
d.f. One-tailed α
 .05 .025 .01 .005
21 1.721 2.080 2.518 2.831
22 1.717 2.074 2.508 2.819
23 1.714 2.069 2.500 2.807
24 1.711 2.064 2.492 2.797
25 1.708 2.060 2.485 2.787
26 1.706 2.056 2.479 2.779
27 1.703 2.052 2.473 2.771
28 1.701 2.048 2.467 2.763
29 1.699 2.045 2.462 2.756
30 1.697 2.042 2.457 2.750
 The critical value for any t-test depends on alpha, the
degrees of freedom for the test, and whether the test is
one-tailed or two tailed
 Example: For a one-tailed t-test where α = .05 and n = 25,
tcrit(24) = 1.711
One-sample hypothesis testing
 RQ1: Are the tutors at Prescription for Success kinder than the
competition?
 H0: μ ≤ 45
 H1: μ > 45
 α = .05, zcrit = +1.645

 RQ2: Are the tutors at Prescription for Success more

compassionate than the competition?
 H0: μ ≤ 55
 H1: μ > 55
 α = .05, tcrit(26) = +1.706

 RQ3: Do the tutors at Prescription for Success care for children

more than the competition?
 H0: μ ≤ 67
 H1: μ > 67
 α = .05, tcrit(26) = +1.706
z-tests: Conducting the statistical test
 Calculate a z-test using the formula:
 Raw score → z score: single raw score – sample mean
 Z statistic: sample mean – pop mean
 Xbar = 57.67
 μ = 45
 σ = 12
 n = 27
One-sample t-tests: Conducting the
statistical test
 Calculate a one-sample t-test using the formula:
 RQ2: ‘compassion’ variable
 X bar = 55.10
 μ = 55
 s = 10.86
 n = 27
 RQ3: ‘childcare’ variable
 Xbar = 56.31
 μ = 67
 S = 13.58
 n = 27
z-tests: Formally stating the results
t(26) = .05, p > .05
t(26) = -4.09, p > .05
One-sample tests: Conducting supplemental
analyses
 If we found statistical significance, compute a confidence
interval and an effect size.
 If we did not find statistical significance, compute a
confidence interval only.
 Effect size: the difference between two values; may be
unstandardized (for example, a difference score) or
standardized (for example, Cohen’s d)
 RQ1: statistical significance
 CI are always two-tailed.
 When conducting a two-tailed test, critical values will be the same.
 When conducting a one-tailed test, they will be different.
 CI
 Effect size represent the observed difference between
two values and represent the practical significance of a
finding.
 Practical significance: the ‘real world’ value of a particular
finding
 Larger sample size → easier to find statistical significance
 Large sample sizes sometimes hide effects that are not very
practically useful.
 Example: Employee satisfaction: 3.1, 3.0, n = 2000
 Unstandardized effect size
 Difference score: the difference between two numbers;
typically used as an unstandardized measure of effect size
 Calculate a difference score using the formula
z-Tests: Effect size
 For a z-test, calculate Cohen’s d using the formula:
One-sample t-tests: Effect size
 For a one-sample t-test, calculate Cohen’s d using the
formula:
Absolute value of d Size of effect
<0.2 Very small
effect
.2 – .5 Small effect
.5 – .8 Medium effect
>0.8 Large effect
Supplemental Analyses after Retaining the
Null
 RQ2 & RQ3: retain the null
One-sample tests: Drawing conclusions
 For both the z-test and the one-sample t-test, your
conclusions should include:
 A formal statement about retaining the null or rejecting the
null and accepting the alternative.
 A formal statement about the statistical significance of the
finding.
 A sentence interpreting the results in terms of the research
question.
 Interpretation of any supplemental analyses.
Hypothesis testing using a z-test: Example
 Last month, Acme Co. employees (n = 100) made an average
of $520.00 in sales. The industry average for monthly sales is
$550 with a standard deviation of 150. Are monthly sales for
Acme Co. different than monthly sales for the industry?

 RQ: Are monthly sales for Acme Co. different than monthly
sales for the industry?
 Hypotheses:
 H0: μ = 550
 H1: μ ≠ 550
 α = .05
 zcrit = ± 1.96
 520 − 550
z= = −2.00
150
100

 z = -2.00, p < .05

520 − 550
d= = −.20
150

 CI.95 = 520 ± 15*1.96

CI.95 [490.60, 549.40]
Hypothesis testing using a z-test: Example
 Reject the null and accept the alternative. The difference
is statistically significant. Monthly sales for Acme Co. are
different than monthly sales for the industry. If we assume
this sample to represent the population, we would expect
95% of sample means to fall between $490.60 and
$549.40. The difference between Acme Co.’s monthly
sales and monthly sales for the industry is a small effect.
Hypothesis testing using a one-sample t-
test: Example
 Acme Co. has been tracking the number of sick days taken per year
by its 100 employees. The employees take an average of 10 sick days
per year with a standard deviation of 3.4 days. The industry average
for sick days is 11. Do employees at Acme Co. take fewer sick days
than the industry in general?
 RQ: Do employees at Acme Co. take fewer sick days than the
industry in general?
 Hypotheses:
 H0: μ ≥ 11
 H1: μ < 11
 α = .05
 tcrit = -1.660
 10 − 11
t= = −2.941176
3.4
100

 t(99) = -2.94, p < .05

10 − 11
d= = −.294118
3.4

 CI.95 = 10 ± 0.34*1.984
CI.95 [9.33, 10.67]
Hypothesis testing using a one-sample t-
test: Example
 Reject the null and accept the alternative. The difference
is statistically significant. Employees at Acme Co. take
fewer sick days per year than the industry in general. If we
assume this sample to represent the population, we
would expect 95% of sample means to fall between 9.33
and 10.67. The difference between sick days taken by
employees at Acme Co. and the general industry is a small
effect.
Learning check
 Beta Inc. has collected data on customer satisfaction. A
recent report shows that on average, customers in this
industry report a 3.5 out of 5 on a customer service
survey. Industry-wide, customers also make an average of
15 complaints per month with a standard deviation of
4.75 complaints.
 Situation 1: Beta Inc. customers (n = 25) reported an
average 3.7 out of 5 with a standard deviation of .75 on
the customer service survey. Is reported customer
satisfaction for Beta Inc. customers different than for
industry customers in general?
Learning check
 Situation 2: Beta Inc. customers (n = 25) reported making
an average of 18 complaints last month. Do Beta Inc.
customers make more complaints than industry
customers in general?
Learning check: Situation 1
 RQ: Is reported customer satisfaction for Beta Inc.
customers different than for industry customers in
general?
 Hypotheses:
 H0: μ = 3.5
 H1: μ ≠ 3.5 3.7 − 3.5
t= = 1.333333
 α = .05 .75
 tcrit = ±2.064 25
 t(24) = 1.33, p > .05
 CI.95 = 3.7 ± .15*2.064
CI.95 [3.39, 4.01]
Learning check: Situation 1
 Retain the null. The difference was not statistically
significant. Reported customer satisfaction for Beta Inc.
customers is different than for industry customers in
general. If we assume this sample to represent the
population, we would expect 95% of sample means to fall
between 3.39 and 4.01.
Learning check: Situation 2
 RQ: Do Beta Inc. customers make more complaints than
industry customers in general?
 Hypotheses:
 H0: μ ≤ 15
 H1: μ > 15
18 − 15
 α = .05 z= = 3.157895
4.75
 zcrit = +1.645
 z = 3.16, p < .05 25
18 − 15
d= = .631579
4.75
 CI.95 = 18 ± .95*1.96
CI.95 [16.14, 19.86]
Learning check: Situation 2
 Reject the null and accept the alternative. The difference
is statistically significant. Beta Inc. customers make more
complaints than industry customers in general. If we
assume this sample to represent the population, we
would expect 95% of sample means to fall between 16.14
and 19.86. The difference between the number of
complaints made by Beta Inc. customers and industry
customers in general is a medium effect.