The T-Test

- Independent Sample T-Test

- Paired Sample T-Test
- One Sample T-Test
- Test of Significance

The One-Way ANOVA

- Post Hoc Comparisons
- Contrasts
Descriptive Statistics
 What is a T-Test

T-Test is a procedure used for comparing Sample

Means to see if there is sufficient evidence to infer that
the means of the corresponding population
distributions also differ. The important things are;

1. Two (t-test always compare two different means)

2. Some variable of interest

Ref: Statistics- A first course by John E Freund and Benjamin M.Perler, 7 th Edition
 SPSS
SPSSPROVIDES
PROVIDESTHREE
THREE
DIFFERENT
DIFFERENTTYPES
TYPESOF
OFT-TESTS
T-TESTS

INDEPENDENT
INDEPENDENTSAMPLE
SAMPLET-TEST
T-TEST

PAIRED
PAIREDSAMPLE
SAMPLET-TEST
T-TEST

SPSS
SPSSPRODUCES
PRODUCESTWO-TAILED
TWO-TAILED
ONE
ONESAMPLE
SAMPLET-TEST
T-TEST SIGNIFICANCE
SIGNIFICANCEVALUE
VALUEBY
BY
DEFAULT
DEFAULT

ONE
ONETAIL
TAILTESTS
TESTS

TWO
TWOTAIL
TAILTESTS
TESTS
 Difference Between Independent Sample T-Test
and Paired Sample T-Test
Independent Sample T-Test Paired Sample T-Test
• The two samples share some • Usually based on the group of
variable of interest in individuals who experience
common, but there is no both conditions of the
overlap between variable of interest.
membership of two groups. • Also called a Repeated
Measure Design or a Paired
• Compare the running speeds Design.
of horses and zebra would be • Compare the running speed
an independent design as of horses for a week of eating
there is no sensible way to one type of feed with the
pair off each horse with each same horses for a week on a
zebra. different type of feed would
be a paired design as you can
pair off measurements from
the same horse
 Independent Sample T-Test

MALE
MALE EXAM SCORE FEMALE
FEMALE

Paired Sample T-Test

QUIZ
”1”
RM
RM RM
RM
Students
Students Students
Students
QUIZ
”2”

QUIZ RM
RM QUIZ
”1” Students
Students ”2”
 One Sample T-Test

It is designed to test whether the mean of a

distribution differs significantly from some present
value e.g.

Professor
Professor
Performance
Performance goal
goalisis82%
82%
of
ofRM
RMClass
Class average
average

Differ Significantly
 Can You identify a T-Test

Textile
Textile Steel
Steel
PRODUCTIVITY
Mills
Mills Mills
Mills

LIFE
Married
Married SATISFACTION
SCORE Unmarried
Unmarried

SET
SETOF
OF
SAT GRE
STUDENTS
STUDENTS
Independent Sample T-Test

A Typical T-Table
 Test of Significance

Test of Significance can be one-tailed or two tailed test;

Two tailed test examines whether the mean of one

distribution differs significantly from the mean of other
distribution. (Regardless of the direction +ve or –ve)

The one tailed test measures only whether the second

distribution differs in a particular direction from the
first.

Ref: Statistics- A first course by John E Freund and Benjamin M.Perler, 7 th Edition
 One Tailed OR Two Tailed
If you have stated your experimental hypothesis with care, it will tell
you which type of effect you are looking for.

For example, the hypothesis that "Coffee improves memory" is

_______tailed test.

The hypothesis, "Men weigh a different amount from women"

suggests a _________tailed test.

So remember, don't be vague with your hypothesis if you are looking

for a specific effect! Be careful with the null hypothesis too - avoid "A
does not effect B" if you really mean "A does not improve B".

Ref: http://www.socialresearchmethods.net/kb/stat_t.php
 Independent Sample T-Test

How close a sample mean might be to

the mean of the population from which
the sample came.

Ref: Student Edition Statistics (Tutorial and Softare) for Behavioral Sciences by Joseph D.Allen
and David .J. Pittenger, 2nd Edition
 Independent Sample T-Test

P-Value
P-Value

Levene’s
Levene’s Test
Test determines
determines whether
whether
the
the variability
variability from
from two
two groups
groups isis
significantly
significantly different.
different. IfIf this
this were
were
significant,
significant,one
onemight
mightconsider
considerusing
using
the
thet-test
t-testfor
forun-equal
un-equalvariances
variances

The degrees of freedom (df) of an

estimate
Ref: is theStatistics
Student Edition number of and Softare) for Behavioral Sciences by Joseph D.Allen
(Tutorial
independent
and pieces of2ndinformation
David .J. Pittenger, Edition
on which the estimate is based.
Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
 Independent Sample T-Test
So far we have learned the following things about a t-test;

1. The t-test produces a single value, t, which grows larger as the difference
between the means of two samples grows larger;
2. t does not cover a fixed range such as 0 to 1 like probabilities do;
3. You can convert a t-value into a probability, called a p-value;
4. The p-value is always between 0 and 1 and it tells you the probability of the
difference in your data being due to sampling error;
5. The p-value should be lower than a chosen significance level (0.05 for example)
before you can reject your null hypothesis.

This value is called the critical value. The final thing to do is compare this value
with your value of t;

If your t-value is greater than or equal to this value, then t is significant and you
have found a difference

If your t-value is less than this value, then t is not significant.

 Independent Sample T-Test

The
Thestandard
standarderror
errorof
ofthe
thedifference
difference
isis the
the function
function of of two
two groups
groups
individual
individualstandard
standarderrors;
errors;

The
The ‘Mean
‘Mean Difference’
Difference’ isis thethe
Ref: Student
difference Edition Statistics (Tutorial and Softare) for Behavioral Sciences by Joseph D.Allen
difference between
between the the two
two group
group
and David .J. Pittenger, 2nd Edition
means.
means.
Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
 T-Test

Ref: http://www.socialresearchmethods.net/kb/stat_t.php
 T-Test

Ref: http://www.socialresearchmethods.net/kb/stat_t.php
CASE OF EQUAL VARINACES CASE OF EQUAL VARINACES NOT
ASSUMED ASSUMED

n1+n2-2
n1+n2-2

Standard Error of the Standard Error of the

Difference Difference
 Paired Sample T-Test

Ref: Student Edition Statistics (Tutorial and Softare) for Behavioral Sciences by Joseph D.Allen
and David .J. Pittenger, 2nd Edition

Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
 Paired Sample T-Test

Ref: Student Edition Statistics (Tutorial and Softare) for Behavioral Sciences by Joseph D.Allen
and David .J. Pittenger, 2nd Edition

Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
 One Sample T-Test

The
The Mean
Mean Difference
Difference isis the
the difference
difference
Ref: Student Edition Statistics (Tutorial and Softare) for Behavioral
between
between the Sciences
the sample
samplebymean
Joseph
mean andD.Allen
and the
the
and David .J. Pittenger, 2 Edition
nd
user-specified
user-specifiedtest
testvalue.
value.

Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
ONE WAY ANALYSIS OF VARIANCE
(ONE WAY ANOVA)
 ANOVA
• Analysis of variance is a procedure used for comparing
sample means to see if there is sufficient evidence to infer
that the means of the corresponding population
distributions also differ.

• Where t-test compare only two distributions, analysis of

variance is able to compare many.

• What does the one-way part mean? It is one dependent

variable (always continuous) and exactly one independent
variable (always categorical). A single independent variable
can have many levels.
Ref: Business Research Methods by Alan Bryman $ Emma Bell
 ANOVA

MALE
MALE EXAM SCORE FEMALE
FEMALE

-- Asian
Asian

-- Black
Black
Ethnic
Ethnic
Group EXAM SCORE
Group
-- White
White

-- Hispanic
Hispanic

One-way
One-wayANOVA
ANOVAwill
willgenerate
generateaasignificance
significancevalue
valueindicating
indicatingwhether
whetherthere
thereare
aresignificant
significant
differences
differenceswithin
withinthe
thecomparisons
comparisonsbeing
beingmade.
made.This
Thissignificance
significancevalue
valuedoes
doesnot
notindicate
indicate
where
wherethe
thedifference
differenceisisor
orwhat
whatthe
thedifferences
differencesare;
are;but
butaa‘Test’
‘Test’can
canidentify
identifywhich
whichgroups
groups
differ
differsignificantly
significantlyfrom
fromeach
eachother.
other.
 One
One Way
Way ANOVA
ANOVA

Ref: SPSS Software

 One
One Way
Way ANOVA
ANOVA

The
The sum
sum ofof squares
squares isis the
the estimate
estimate of of the
the variance
variance
multiplied
multiplied byby the
the degrees
degrees of of freedom.
freedom. ItIt can can bebe
calculated
calculatedininthree
threeways;
ways;
-SST
-SST -SSW
-SSW -SSB -SSB
The
The advantage
advantage of of sum
sum ofof squares
squares isis that
that itit can
can deal
deal
Ref:
withStudent
withunequal
unequal Edition
subgroupStatistics
subgroup sizes (Tutorial
sizesand theyand
andthey addSoftare)
add up like; for Behavioral Sciences by Joseph D.Allen
uplike;
and
SSTDavid
SST SSB.J.
==SSB +SSWPittenger, 2nd Edition
+SSW

Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
One
One Way
Way ANOVA
ANOVA
 One
One Way
Way ANOVA
ANOVA

When
When“F”“F”ratio
ratioisisclose
closeto
to“1”,
“1”,the
theestimates
estimateswill
willbe
besaid
saidto
tobe
besimilar
similar(no
(noindication
indication
of
of detectable
detectable differences
differences between
between subgroup
subgroup means).
means). When
When F-ratio
F-ratio isis large,
large, the
the
estimates
estimatesare
aredissimilar
dissimilar(indication
(indicationof
ofdifferences
differencesbetween
betweensubgroup
subgroupmeans).
means).

Ref:
SPSSStudent
SPSS Edition Statistics
calculated
calculated that
that anan(Tutorial
with 3and
FF with Softare)
3 and
and 96 forthat
96 df
df Behavioral
that equalsSciences
equals 2.272by
2.272 Joseph
has
has D.Allen
aa two
two tailed
tailed
and David .J. Pittenger,
probability
probability of 0.0852 Edition
of0.085
nd

Ref: SPSS for psychologists by Nicola Brace, Richard Kemp and Rosemary Snelgar
 One
One Way
Way ANOVA
ANOVA –– Post
Post Hoc
Hoc Tests
Tests

Ref: SPSS Software