T TEST 1

T-tests
PSYC2009 Lecture 7
Welcome back!
2
Announcements
• End of semester exam (final exam)
– DRAFT: Wednesday 12/06/2024, 9:00am,
• 3 hours, + 15 minutes of reading time
• Writing Assessment
– Due 22/4
– Watch 18/3 lecture for explanations
– Key: Remember 4 parts, graded independently
• No tutorials this week; starts up again next
week
3
More announcements
• My research talk:
– Thursday, 18/4/24, 12-1PM
– Peter Baume Building 42A Level 2, Room 2.01,
University Avenue ANU and Zoom
– Zoom Link details:
https://anu.zoom.us/j/84043348603?pwd=M2dq
ZTVYcnE2MDdkeUZHREZKNUZ5dz09 |
Meeting ID: 840 4334 8603 | Password: 990845
4
Lastly
• “lengthen your time horizon”
5
https://pages.apa.org/2024-trends-in-psychology/
6
https://pages.apa.org/2024-trends-in-psychology/
7
In this week's lectures, we will cover..
• Introduction to the t-test
• One sample t-tests
• Next week:
– Independent t-tests
– Paired samples t-tests
8
Introduction to the t-test
• What does a t-test do?
– Compares two means
– Determines statistical significance of
difference
• Used when the population standard
deviation is not known (although the
population mean might be known)
9
t-distribution: refresher
• t-test: an inferential test
10
Calculating t
• The basics: How is t calculated?
11
Types of t-tests
• One sample t-test
– (or 'single sample t-test’). Tests whether an unknown
population mean differs from a known 'standard'
• Independent sample t-test
– (or 'unrelated t-test’). Compares two different groups of
participants or conditions, on a single variable
• Paired sample t-test
– (or 'related t-test'). Same group of participants at two
time-points (repeated measures), or 'matched-pairs'
12
Examples of different types of t-tests
• One sample t-test
– Do university workers enter retirement later
than most?
• Independent sample t-test
– Do students who attend lectures experience
better well-being than students who do not?
• Paired sample t-test
– Two history exams on the same persons
before and after taking a history course
13
Parametric techniques
• T-tests, ANOVAs, Pearson correlation
• Assumptions:
– Level of measurement (DV is interval or ratio)
– Random sampling
– Independence of Observations
– Normal distribution
– Homogeneity of variance
14
Parametric techniques
• Assumptions not met?
– Cross fingers (‘robustness’)
– Data manipulation
– Non-parametric alternative
15
One-sample
t-test
16
One-sample t-test: examples
• When might we use a one-sample t-test?
– When a population mean is known: e.g.,
height / weight / blood pressure
– When a population mean is hypothesised:
e.g., working hours
17
One-sample t-test
• What do we need?
– One test variable. Population mean.
– A hypothesis (or expectation) about how the sample
mean will differ from the population mean
• Our null hypothesis (H0)
– That our sample mean is equal to the population
mean
• Alterative hypothesis
– 2 ways: A > B or A ≠ B
18
One-sample t-test
• Assumptions
– DV should be measured at the interval or ratio
level
– Data are independent
– No significant outliers
– DV is approximately normally distributed
19
One sample t-test example
• Are quiz scores
due to chance?
• MCQ- 5 options
• 20% chance
•  20 is the
population mean
20
21
22
23
• Half-width of the confidence interval
w = tα/2sx̄
where sx̄ = standard error (1.164)
24
25
w = tα/2sx̄ = (2.2622)(1.164) = 2.633
Pr(25 – 2.633 < u < 25 + 2.633)

Pr(22.37 < u < 27.63)
26
Population mean = 20
Lower CI = 20+2.366
Upper CI = 20+7.634
 95% CI = [22.37, 27.63]
27
One-sample t-test: Effect size
• Eta-squared
• Cohen’s d
– Cohen’s d Effect size:
Very Small Small Medium Large

< .20 .20 - .49 .50 - .79 .80 and above
28
One sample t-test: Cohen’s d calculation
• x̄ is the sample mean

• s is the sample standard deviation with
n -1 degrees of freedom
• µ is the theoretical/known mean
29
One sample t-test: Effect size
d = (25.00 – 20.00) / 3.682
d = 1.358
30
Very Small Small Medium Large

< .20 .20 - .49 .50 - .79 .80 and above
31
Reporting a one-sample t-test
“A one-sample t-test was run to test whether quiz scores
were indicative of random responding. The test group
produced higher quiz scores (M = 25.00, SD = 3.68) than
the hypothesized mean for the population (M = 20.00),
t(9) = 4.29, p = .001. The difference between the sample
mean and the population mean (5.00) was large, d = 1.36;
95 CI [2.366, 7.634].”
32
Non-parametric alternatives
• Wilcoxen signed rank test (Wilcoxen’s T)
33
Non-parametric alternatives
• Wilcoxen signed rank test:
34
Next week
• Independent and paired-samples t-tests
35

T TEST 1

Uploaded by

Copyright:

Available Formats

T TEST 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

T TEST 1

Uploaded by

Copyright:

Available Formats

T-tests

• t-test: an inferential test

– Cross fingers (‘robustness’)

where sx̄ = standard error (1.164)

w = tα/2sx̄ = (2.2622)(1.164) = 2.633

Pr(25 – 2.633 < u < 25 + 2.633)

Very Small Small Medium Large

• x̄ is the sample mean

d = (25.00 – 20.00) / 3.682

Very Small Small Medium Large

• Wilcoxen signed rank test (Wilcoxen’s T)

• Wilcoxen signed rank test:

You might also like