4. Effect Size Test
Effect size was a way to measure the size of the difference between two groups. It was important for quantifying the effectiveness of a particular
intervention. It facilitated us to find a more extended result in a research, like to find how well a strategy works in a research does.
In calculating effect size, the researcher used Cohen’s d with the manual
formula
2
:
To get the pooled standard deviation, the formula is
3
: √
Explanation: d
: Cohen’s d effect size
M1 : Mean score of post test experimental group
M2 : Mean score of post test control group
SD1 : Standard deviation score of experimental group
SD2 : Standard deviation score of control group
SDpooled : The average standard deviation score of the two scores
from two groups The standard criteria of
Cohen’s d effect size were
4
: 0.1
0.1 - 0.3 0.3 - 0.5
0.5 = trivial effect
= small effect = moderate effect
= large difference effect The other Cohen’s d effect size interpretation was5:
2
http:staff.bath.ac.ukpssiwstats2page2page14page14.html retrieved
on 1st
November 2016.
3
Ibid.
4
http:meera.snre.umich.edupower-analysis-statistical-significance-effect-size retrieved
on November 1
st
2016.
0.2 0.5
0.8 = Smalllow effect
= Mediummoderate effect = Largehigh effect
G. Statistical Hypothesis
In order to get the answer of the hypothesis above, the researcher proposed alternative hypothesis H
1
and null hypothesis H which was provided as
follows: H
= sig. 2-tailed of t-test 0.05 H
1
= sig. 2-tailed of t-test 0.05 Where:
H : There is no effect of using Question Generation Strategy in learning
reading of narrative text. H
1
: There is an effect of using Question Generation Strategy in learning reading of narrative text.
If sig. 2-tailed of t-test 0.05, H null hypothesis is accepted, and H
1
alternative hypothesis is rejected. If sig. 2-tailed of t-test 0.05, H
null hypothesis is rejected, and H
1
alternative hypothesis is accepted.
5
Lee A. Becker, http:web.uccs.edulbeckerPsy590es.htm
retrieved on November 1
st
2016.