T Test For Independent Samples

T-TESTS
A t-test is used to determine if the scores of two groups

differ on a single variable. A t-test is designed to test for the ·
differences in mean scores. For instance, yo~ could use a
t-test t.o determif?-e whether performance differs among
students in two classrooms.
Note: A t-test is appropriate only ,when looking at paired

data. It is useful in analyzing scores of two groups of
participants on a particular variable or in analyzing scores of
a single group of participants on two -v ariables.
Two Sample Tests

This is a very common problem to compare the means of,
for example, the experimental group and the control group
on some independent variable.
Tl1 e t,,,o -sa mp le t-te sts we wil l con sid er are :
1 · Th e ind epe nd en t t-te st, wh ich is use d to com par e two
\
sru npl e me ans wh en the two sam ple s are ind epe nde nt of
one a.ii. oth er.
.
2 · Th e no n-i nd epe nd ent or dep en den t t-te st wh ,ich is use d
for ma tch ed sam ple s (wh ere the two sam ple s ·a re not
ind epe nd ent of one ano the r as the y are ma tch ed) and for
pre -te st/ po st- tes t com par iso ns wh ere the pre -te st and
po st- tes t are tak en on the sam e gro up of sub jec ts.
The formu la .. th .
ior e indep enden t t-test is:
Wher e:
X1 = Mean for group 1,

X2 - Mean for group 2,
SS1 = The sum of squar es for group 1,
•
SS2 = The sum of squar es for group 2,
n 1 - Numb er of subje cts in group 1, and
n 2 = Numb er of subje cts in group 2.
The sum of squar es is a new way of lookin g at varian ce.

It gives us an indica tion of how spread out the scores in a
sampl e is. The t-valu e we are findin g is the differe nce betwe en
the two mean s divide d by their sum of square s and taking
the degree s of freedo m into consid eratio n. ·
ss 1
- I x 1 2 - t I X1J 2
"1
and
ss :l - I x:i :i - , I xa-1 :i
"2
We can see that each sum of squa res is the sum of the
squa red scor es in the sam ple min us the sum of the scor es
quan tity squa red divid ed by the size of the sam ple (n).
So to calc ulat e the inde pen den t-t valu e we nee d to

kno w:
1. The mea n for sam ple o~ grou p 1

2 . The me ~ for sam ple or grou p 2
3. The sum mat ion X and sum mat ion X squa red
for grou p 1
4. The sum mat ion X and sum mat ion X squ ared
for grou p 2
S. The sam ple size for grou p 1 (n ) ·
1
6. The sam ple size for grou p 2 (n )
2
We also need to kno w the deg rees of free dom for the
inde pen den t t-tes t whi ch is:
d/ -n1 +~ -2
2. Determine Degrees of Freedom (df). Use the associated
df from the test you 1?-ave chosen for your business
problem. The basic rule is that for each sample there will
be n- l degrees of freedom . For instance, in a test of two
means where there are 2 samples used, the df becomes
nl + n2 - 2 (which is the same as (nl - 1) + (n2 - I)).
3. Locate df in Table. The ·f irst column of the t-table lists
df. Locate the df from step 2, if that number of df is not in
the table - round DOWN to the next nearest number.
4. Find t-Critical Value. Using the row associated with the

df identified in step 3, go over to the number under the
column representing your alpha level from step 1 (ie. - if
alpha= .05 then the critical value would be in the column
labeled "t.050") This number is then your critical val\le.
Research Problem:
Job satisfaction as a function of work schedule was investigated in two

different factories. In the first factory the employees are on a Night shift system
while in the second factory the workers have a Day shift system. Under the Night
shift system, a worker always works the same shift, while under the Day shift
system; a worker works in the whole shifts. Using the scores below determine if
there is a significant difference ,n job satisfaction between
the two groups of workers.
Night shift (X1) Day shift (X2) (X1) 2 (X2) 2

I 79 63 6241
(79x79) = (63x63)=3969
83 71 6889 5041
68 46 4624 2116
59 57 3481 3249
81 53 6561 2809
76 46 5776 2116
80 57 6400 3249
74 76 5476 5776
58 52 3364 2704
49 68 2401 4624
68 73 4624 5329
LX1= 775 LX2 = 662 L(X1) 2 = 55837 2
! L(X2) = 40982
ss1 - 55 83 7 - l 77 S J 2
11
- 55 8 37 - 60 062 5
11
- 5 58 37 - 54 60 2.2 7
- 12 34 . 73
ss2 = I x2 2- t I x~ 2
n2
ss2 - 40 98 2 - t 66 2 )~
11
- 40 98 2 - 43 82 44
11
- 40 98 2 - 39 84 0.3 6
- 11 41 .6 4
X 1 - 77 5 ::,: 70 .4 5
11
x2 _ 66 2 - 60 .1 s
11
t - 70 .4 5 - 60 .1 8
[1 23 4. 73 + 11 41 .6 4
11 + 11 -2
J[1 1J
11
+
11
- 70 .4 5 - 60 .1 8
✓[23726037 J [.182 J
- 10 .2 7
J (1 18 .82 ) (.1 82 )
- 10 .2 7
4. 65 0
- 2. 20 9
. IADL t t; .
Cri~ica-1Values of Student's ·I DistribuUota ·
1 .10 .05 ✓ -025 .01 .005 .0805

df .20 .10 .osc ✓· 1-tall
.02 .010 .0010. 2-talf
1
2
f'\
J
4
5
6
7
s ,,,.,
9
10
11
12
13 :
14- -
lS
16
·.17
.18
19". -
20 :✓
-2,1
-i2
23
24
,~
Problem:
Is ther e signi fican t diffe rence In lhfJ jot, s,1tls f'-1ctl on lw

t w t•t· n ttw two
grou ps of work ers?
Hypothesis (Ha):
('t'n two
There is no significant differenc e In th e job sa t lsfc1c::t io 11 bPlW
groups of workers.
Leve/of significance: a=.0 5
df (degrees of freedom) = n1+ n1 - 2

= 11+ 11 - 2
= 20
t.05 = -2.086 (see t he table for crit ica l value of

t distribut ion. 2-tail)
samples
Statistics: t-tes t for depe nden t samples or correlated
Decision rule:
al value
If the computed or calculated value is greater than the critic
(cv), reject null hypothesis (Ho) .
Conclusion:
of
Since the calculated value of 2.209 is grea ter than the t-cri tico/ valu e
2.086 at .05 level of sign ificance with 20 degrees of free dom the null
e
hypothesis is rejected, which means that there is significant differenc
in the job satisfaction between the night shift and day shift system.

T Test For Independent Samples

Uploaded by

Copyright:

Available Formats

T Test For Independent Samples

Uploaded by

Document Information

Original Title

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 For Independent Samples

Uploaded by

Copyright:

Available Formats

T-TESTS

A t-test is used to determine if the scores of two groups

Note: A t-test is appropriate only ,when looking at paired

Two Sample Tests

X1 = Mean for group 1,

The sum of squar es is a new way of lookin g at varian ce.

So to calc ulat e the inde pen den t-t valu e we nee d to

1. The mea n for sam ple o~ grou p 1

4. Find t-Critical Value. Using the row associated with the

Job satisfaction as a function of work schedule was investigated in two

Night shift (X1) Day shift (X2) (X1) 2 (X2) 2

1 .10 .05 ✓ -025 .01 .005 .0805

Is ther e signi fican t diffe rence In lhfJ jot, s,1tls f'-1ctl on lw

Leve/of significance: a=.0 5

df (degrees of freedom) = n1+ n1 - 2

t.05 = -2.086 (see t he table for crit ica l value of

You might also like