Hypothesis Testing:

Basic Principles
The main purpose of statistics is to test a hypothesis.
 What is hypothesis?
• A tentative answer to the research question.
• A prediction of what the researcher expects to find based on past
studies in the literature that suggest certain outcomes.
• A prediction or testable statement that a researcher aims to test to
see if it is supported or rejected.
• A tentative explanation that accounts for a set of facts and can be
tested by further investigation.
• A statement in quantitative research in which the researcher
makes a prediction about the outcome of a relationship among
attributes or variables.
Research question and hypothesis
• RQ: Is there a relationship between pupils’ self-esteem and the
amount of time they spend watching television?
• H: “There is a relationship between pupils’ self-esteem and
the amount of time they spend watching television”

• RQ: Is there a relationship between parental income and

students’ achievement at school?
• H: There is a relationship between parental income and
students’ achievement at school.
Types of hypothesis
• the research hypothesis;
• the alternative hypothesis (the experimental hypothesis);
• the null hypothesis.
The Research Hypothesis
• The research hypothesis is a general prediction made at the
beginning of the study as to what a researcher expects will
happen.
• It is usually written at the end of the introduction.
• An example might be “People in rural areas are more likely to
help someone than people in large cities”
• “Students with high ICT literacy will have higher achievement at
school than students with low ICT literacy”
Alternative (or experimental) Hypothesis
• The alternative (or experimental) hypothesis is the
operational statement of the research hypothesis. That means that
it actually states the precise behaviors or responses that are going
to be used to measure the variable under investigation.
• Alternative hypothesis (Ha or H1) shows that there will be a change,
a difference, or a relationship for variables in the population.
• H1 : µA ≠ µB (is not equal to) = non directional
• H1 : µA > µB (greater than) = directional
• H1 : µA < µB (smaller than) = directional
• µ = population mean
• An example might be “A young woman who drops her shopping
outside a village shop will receive help more quickly than a
young woman who drops her shopping outside a large city
superstore”
• Students with high ICT literacy will learn English as a foreign
language better than students with low ICT literacy.
Null Hypothesis
• The null hypothesis is the hypothesis that a statistical test
actually tests. Siegel (1956) calls it ‘a hypothesis of no
differences’.
• Null hypothesis (H0) shows that there is no relationship between
independent and dependent variables or no difference between
groups of an independent variable or a dependent variable.
• H0 : µA - µB = 0 or µA = µB
Null Hypothesis
• An example might be “There will be no difference in how
quickly a young woman who drops her shopping outside a
village shop or a large city superstore receives help. ”
• Students with high ICT literacy will not learn English as a
foreign language better than students with low ICT literacy.
Other Examples
• A research wanted to know whether adding moving pictures to
a presentation will improve pupils’ memory of the key content
of the presentation.
• Alternative hypothesis (H1): adding moving pictures will improve
pupils’ retention of the content.
• Null hypothesis (H0): adding moving pictures will not improve
pupils’ retention of the content.
 Directional vs Non-directional hypotheses

• If the hypothesis predicts the expected direction of the result then

it is called a directional (or one-tailed) hypothesis.
• For example:
• School children who eat breakfast do better at school than
those who do not.
• There is a positive correlation between height and weight.
• If the hypothesis just says that there will be a difference or a
correlation—but does not predict the direction—then it is called a
non-directional (or two-tailed) hypothesis.
• For example:
• Female and male students differ in how quickly they learn a
foreign language (note that the hypothesis doesn’t state which
will learn fastest), or
• There is a correlation between exam results and hours spent
watching TV (note that it doesn’t say if watching a lot of TV will
be related to better or worse exam results).
• There is a difference between students who attend traditional
course and those who attend online course in terms of
academic achievement (Non directional)
• Students who attend traditional course will have better
academic achievement than and students who attend online
course (Directional)
• Directional or non-directional?
• Students whose parents and teachers communicate through the
Internet will have higher grades than students whose parents
and teachers do not communicate through the Internet.
Hypothesis
• Alternative hypothesis:
• Hasil belajar listening siswa yang menggunakan lab bahasa berbeda
dengan hasil belajar siswa yang menggunakan kelas biasa. (Non
directional)
• Alternative hypothesis:
• Hasil belajar listening siswa yang menggunakan lab bahasa lebih baik
dari hasil belajar siswa yang menggunakan kelas biasa. (Directional)
• Null hypothesis:
• Hasil belajar listening siswa yang menggunakan lab bahasa tidak
berbeda dengan hasil belajar siswa yang menggunakan kelas biasa.
(hypothesis for testing)
Hypothesis testing
• It is a procedure for making decisions about results by
comparing an observed value of a sample with a population
value to determine if no difference or relationship exists
between the values.
• Are the sample scores you receive a good estimate of the
population scores?
• How confident are you that your sample score is right?
 Steps in hypothesis testing
a) identify a null and alternative hypothesis;
b) set the level of significance, or alpha level (α);
c) collect data;
d) compute the sample statistic; and
e) make a decision about rejecting or failing to reject the null
hypothesis.
Level of Significance
• indicates the probability that the results that you obtained in
your research occurred on a chance basis (significance level).
• If the level of significance indicated is 0.05, this means there is
a 5% or 1 in 20 probability that your results are a chance
effect.
• It is the level of the risk that we are prepared to take that our
conclusions are wrong.
• In social science the 0.05 or 5% level has become the accepted
standard (in medical science, 0.01)
• If your statistical results reach this level, it is the convention
that you can reject the null hypothesis.
• <0.05: the null hypothesis is rejected/the alternative
hypothesis is accepted
• >0.05: the null hypothesis is not rejected/the alternative
hypothesis is rejected
Statistical test of hypothesis
• Z-test if the samples are big and independent (for example, the
samples are group A dan group B/two different groups of
scores).
• T-test for independent samples if the samples are small and
independent.
• T-test for correlated samples if the samples are big or small and
they are matched or correlated (not independent, i.e. two sets
of scores from one group only  tests are given twice to the
same group).
• Calculate the t-value or z-value.
 Correlations
  LA M C SA
Spearman's rho Listening Correlation 1.000 .475 .518**
**
.323*
Achievement Coefficient
(LA)
Sig. (2-tailed) . .002 .001 .042
N 40 40 40 40
Metacognitive Correlation .475** 1.000 .353* .334*
(M) Coefficient
Sig. (2-tailed) .002 . .026 .035
N 40 40 40 40
Cognitive Correlation .518** .353* 1.000 .319*
(C) Coefficient
Sig. (2-tailed) .001 .026 . .045
N 40 40 40 40
Socioaffective Correlation .323* .334* .319* 1.000
(SA) Coefficient
Sig. (2-tailed) .042 .035 .045 .
N 40 40 40 40
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).