Professional Documents
Culture Documents
By Suprihadi
English Education Department Universitas Muria Kudus
First of all you have to understand and should be your mindsets that:
1. Not all statistics or numbers which are physically different are always statistically different
or significantly different.
2. To know whether or not some numbers are significantly different from each other, we have
to do a hypothesis testing or significance testing.
3. The term significant in hypothesis testing, or in statistics in general, means meaningful or
important.
Whether or not two numbers are significantly different does not depend on the distance
between the two numbers.
Example: A distance between Rp. 1.000.000,- and Rp. 5.000.000,- for a rich person, is not
significant because having Rp. 1.000.000,- and Rp. 5.000.000,- is not different,
but
Your score of 85 is significantly different from 84 although the distance is only 1. This is
because those who get 85 will get A while those who get 84 will get AB.
4. A research can be conducted if there is a problem or there are problems. Without any
problem there will be no research, because the general purpose of a research is to solve a
problem.
Example:
If the reading comprehension of the students of SMA X is low, you can see it as:
a. a problem because, for example, the ability to read is very important to get knowledge
b. not a problem at all because, for example, the students can learn to read by themselves
Whether a certain phenomenon is seen as a problem or not depends, among others, on: a)
the significance or importance if the problem is solved, and b) the knowledge of the
researcher.
5. In a research, the problem is usually called research question, problem formulation, or
statement of the problem (although it is called statement but the real form of the sentence
is question).
6. There must be a problem in a research but there is not always a hypothesis in a research: a
research may not have a hypothesis.
Example 2 (2 variables):
Problem: A teacher noticed that the speaking skills of his students were not
satisfactory. He further noticed that they seemed to be ashamed when
they were asked to come forward to speak. He thought that his way of
teaching using Technique X might be boring and monotonous. Therefore,
he decided to change his way of teaching with Technique Y so as to
increase the speaking skills of hi students.
Hypothesis: 1. The speaking skills of the students of SMA 10 taught by using
Technique Y are better than the speaking skills of the students of
SMA 10 taught by using Technique X. (in the second statement the
speaking skills of the students of SMA 10 can be changed with those,
which is grammatically better)
or
2. The speaking skills of the students of SMA 10 taught by using
Technique X are different from the speaking skills of the students of
SMA 10 taught by using Technique Y. (the speaking skills of the
students of SMA 10 can be changed with those, which is
grammatically better)
Note:
Hypothesis no. 1 is called directional hypothesis while hypothesis no. 2 is
called non-directional hypothesis. In non-directional hypothesis a
researcher only says different, symbolized by the mathematical symbol ≠,
which may mean better (>) or worse (<). The choice between 1 and 2
D. Hypothesis testing
2. Types of hypothesis testing based on the researcher’s expectation about the population
from which the sample was selected. Learn PPT 8 slide 19-20 and pages 188 – etc. of the
book.
a. One-tailed hypothesis testing: use only one side/tail (right side or left side) to be the
critical region
b. Two-tailed hypothesis testing: use both sides/tails (right side and left side) to be the
critical regions
Note:
In X2 Test, there is only one-tailed test using the right tail
ESTABLISHING THE CRITICAL REGION, ONE-TAILED TESTS VS TWO-TAILED TESTS at α
= 0.05 USING Z DISTRIBUTION (PPT 8 slide 21 and page 191 of the book)
Two-Tailed Test
(Non-Directional
The blue-shaded areas are what so called critical regions and the point of the
intersection between the vertical lines and the horizontal lines on those areas are the
beginnings of the critical values (In Figure 8.5 above are +1.65 and -1.65).
Those types of hypothesis relate with the classifications based on the expected finding,
those are directional and non-directional hypotheses: one-tailed test is directional while
two-tailed test is non-directional.
3. How to write hypotheses in hypothesis testing
Examples:
1) Research hypothesis:
written in sentence
The listening ability of the students taught by using audio visual media is better
than that taught by using audio media.
written in symbol
Ha: 𝑋1 > 𝑋2
𝑋1 = Mean of the listening ability of the students taught by using audio visual
media
𝑋2 = Mean of the listening ability of the students taught by using audio media
2) Research hypothesis:
written in sentence
The mortality rate this year is less than that of last year.
written in symbol
Ha: 𝑋1 < 𝑋2
𝑋1 = Mean the mortality rate this year
𝑋2 = Mean of the mortality rate last year
3) Research hypothesis:
written in sentence
The writing skills of Class A are different from the writing skills of Class B.
written in symbol
Ha: 𝑋1 ≠ 𝑋2
𝑋1 = Mean of the writing skills of Class A
𝑋2 = Mean of the writing skills of Class B
Step 3. Selecting the sampling distribution and establishing the critical region
Suplementary Material Page 9
1. With Z distribution:
Sampling distribution = z distribution
α = 0.05, two-tailed test
Z (critical) = ± 1.96
With Z distribution, the Z (critical) has been determined and standardized depending
on the level of the alpha (α). Z (critical) is also called Z (table) to compare with
Z(obtained)
FINDING CRITICAL Z SCORES
For more detail, learn PPT 8 slides 19-20 and page 190 of your statistics book.
2. With t distribution:
Sampling distribution = t distribution
α = 0.05, one-tailed test
Degrees of freedom = (N1+N2) – 2 = …
t (critical) = ….
The degree of freedom (df) depends on the number of the sample (N1 and N2), and
the values of t(critical) depends on the α level and the df. In the example in your
statistics book (data on page 214), the α level is 0.05, the df is (42 + 37)-2 = 77. T
(critical) is also called t (table), because it is taken from table of t distribution.
3. With X2 distribution:
Sampling distribution = X2 distribution
α = 0.05
Degrees of freedom = (r – 1)(c – 1) = … (1)
X2 (critical) = …. (3.841)
The degree of freedom (df) in chi square (X2) test of significance is determined by
the formula (r – 1)(c – 1), for which r refers to number of the row while c refers to
the number of the column (Learn Hypothesis Testing IV). The value of X2 (critical) is
the intersection between the alpha level (α) and the degree of freedom (df).
The example on page 264 Table 11.3 indicates that the number of the row is 2 and
the number of the column is also 2. Therefore, the df is (2 – 1)(2 – 1) = 1, so that the
critical value of α 0.05 and df 1 is 3.841.
b. In Hypothesis Testing II
X1 X2
Formula 9.2 Z (obtained )
X X
Formula 9.4
2. With t distribution:
The formula to compute the value of t(obtained): refer to PPT 9C slides 15-18 and page
215 Formula 9.5 and 9.6.
Before we can use the formula above, we have to calculate the standard error of the
difference between two means or the de-numerator (number below the line in a
fraction) first using the following formula:
SS SS 1 1
= n n 2 n n
1 2
1 2 1 2
3. With X2 distribution:
The formula to calculate the X2(obtained) is as follows (refer to PPT 10 to use this
formula)
fo fe 2
X 2
fe
4. With r distribution:
r = relationship
a. The formula to calculate Pearson Product Moment coefficient correlation rxy for
interval-ratio data is as follows (refer to PPT 11 slides 9-16 to use this formula).
We can also use t distribution for testing Pearson’s r for significance (see pp.380-
381).
b. The formula to calculate Spearman rho (rank) Correlation Coefficient for ordinal data
(refer to PPT 11 slides 17-26 to use this formula).
6 D 2
1
N ( N 2 1)
EXPLANATION OF HYPOTHESTING 1
FINAL REMARKS
an idea or explanation for something that is based on known facts but has not yet been proved
(Cambridge Dictionary: https://dictionary.cambridge.org/dictionary/english/hypothesis)
A tentative assumption made in order to draw out and test its logical or empirical
consequences