Practicum 8

Lecture 8.2

Themes
• Probability theory (Taking a chance)
• The normal curve
• The standard normal distribution
• Inferential statistics
• Hypothesis testing
• T-tests
Sources:

Salkind – Chapter 9
Gravetter et al. – Chapter 8
In QUANTITATIVE methodology we try to
transform all our research ideas to look at:
▪ A difference between groups
▪ Or a correlation between variables

Idea RQ Hypothesis
DIFFERENCES BETWEEN CORRELATION BETWEEN
GROUPS 2 VARIABLES
E.g. Does the resilience of E.g. Is there a relationship
students exposed to trauma between resilience and
differ from the resilience of happiness?
students not exposed?
▪ States in question form what you would like to know

▪ Is there a difference in attitudes towards success

between male and female students?
▪ Is there a relationship between mindfulness and
well-being?

▪ Come up with your own research question

▪ Your educated guess about what the answer to your
research question is
▪ Must be a statement
▪ There is a difference between male and female
students in terms of their attitude toward success.
▪ There is a relationship between mindfulness and
well-being
Evening class test average = 65%

Morning class test average = 70%

Mathematical Apparent and what you can

difference see

Due to factor being investigated –

Statistical difference Independent variable
Test to see if it is by CHANCE or not
 Representativeness

Population Sample
Sampling
Hypothesis Error Use to test
pertains to hypothesis

Generalizability of
Research results
Null Hypothesis: H0

Research/
Alternative Hypothesis: H1
▪ H0
Assume that there is NO difference or relationship
Or that differences/relationship are due to chance
UNTIL it can be proved otherwise
▪ H0: µ1 = µ2

▪ Null Hypothesis serves as a starting point

▪ Assumed to be true, until can be proved otherwise
▪ Results are described in terms of the H0
There is no difference in the moral behaviour of women and the
moral behaviour of men.
OR
There is no relationship between mindfulness and well-being
amongst first year university students

Take a look at the following research questions and formulate a

null hypothesis
• Is there a relationship between the length of time a person spends
working on a community service project and their attitudes
towards volunteer work?
• Is there a difference in satisfaction with life between full-time
working mothers and stay-at-home mothers?
Question Null Hypothesis
Are teens better at math than adults? Age has no effect on mathematical ability.

Does taking aspirin every day reduce the Taking aspirin daily does not affect heart
chance of having a heart attack? attack risk.
Do teens use cell phones to access the Age has no effect on how cell phones are
internet more than adults? used for internet access.
Do cats care about the color of their Cats express no food preference based on
food? color.
Does chewing willow bark relieve pain? There is no difference in pain relief after
chewing willow bark versus taking a
placebo.
Corresponds to the null about the same thing
Can have more than one form of research hypothesis for any given
null
▪ H1
States that there IS a difference/relationship
Statement of inequality
Differences observed not due to chance = due to IV

Two types of research hypothesis, namely:

▪ Non-directional ( ≠) nature of difference not specified
▪ Directional ( > 𝐎𝐑 < ) nature of difference specified
Non-directional ( ≠) Directional ( > 𝐎𝐑 < )

• There is a relationship • There is a positive

between the length of relationship between length
time a person spends of time spent on community
working on a community service and attitude towards
service project and their volunteer work. >
attitudes towards
volunteer work.
• Full-time working mothers
• There is a difference in
are less satisfied with life
satisfaction with life
than stay-at-home mothers.
between full-time working
<
mothers and stay-at-home
mothers.
Take a look at the following research questions
and formulate a research hypothesis H1

▪ Is there a relationship between coping and

happiness? (non-directional)

▪ Is there a difference in coping between

adolescents and students? (directional)
Null Hypothesis Research Hypothesis
Represents equality: Represents inequality:
= ≠ OR > 𝐎𝐑 <
No relationship between Relationship between
variables variables
Refer to population Refer to the sample

Tested indirectly: through Tested directly: sample

the sample
Has corresponding Nondirectional:
research hypothesis two-tailed : ≠ or
Directional:
one-tailed test: > 𝐎𝐑 <
Is there a difference in academic performance between
students who meditate and students who do not
meditate?
▪ Formulate in words the null hypothesis
▪ Formulate in words a directional research hypothesis
▪ Formulate in words a non-directional research hypothesis

▪ However, hypothesis can also be represented

symbolically
▪ Refers to population parameters
▪ There will be NO DIFFERENCE in academic
performance between students who meditate and those
who don’t meditate
▪ H0 : µ1 = µ2
▪ Where
▪ µ1 = the mean academic performance score of the students
who meditate and
▪ µ2 = the mean academic performance score of the
students who do not meditate
▪ The academic performance of students who meditate will
be better than the academic performance of students
who don’t meditate
▪ H1: µ 1 > µ2
Where:
µ1 = the average academic performance mark of the
students who meditate and
µ2 = the average academic performance mark of the
students who do not meditate
▪ A directional alternative hypothesis
▪ The academic performance of students who meditate
will be lower than the academic performance of
students who don’t meditate
▪ H1: µ1 < µ 2
Where:
µ1 = the average academic performance mark of the
students who meditate and
µ2 = the average academic performance mark of the
students who do not meditate
▪ A directional alternative hypothesis
▪ There will be a difference in the academic
performance between students who meditate and those
who don’t meditate
▪ H1: µ1 ≠ µ2
Where:
µ1 = the average academic performance mark of the
students who meditate and
µ2 = the average academic performance mark of the
students who do not meditate
▪ A non-directional alternative hypothesis
DIRECTIONAL NON-DIRECTIONAL

BETTER than…BIGGER than…OR There is a DIFFERENCE or there

WEAKER than…SMALLER than is a RELATIONSHIP

H1: µ1 > µ2 OR H1: µ1 ≠ µ2

H1: µ1 < µ2
µ1 = the mean score in group 1 µ1 = the mean score in group 1
µ2 = the mean score in group 2 µ2 = the mean score in group 2
The academic performance of There will be a difference in the
students who meditate will be academic performance between
better than the academic students who meditate and those
performance of students who who don’t meditate
don’t meditate
Extension of your Your educated Stated in Posits an expected
original research guess about what declarative form relationship
question the answer to your between variables
research question OR a difference
is between groups

Reflects a theory or Brief and to the Testable

body of literature point
on which it is
based
▪ H1 is tested directly by means of statistical analysis
H0: µ1 = µ2 There is NO difference in coping of adolescents' students
H1: µ1 ≠ µ2 There is a difference in coping of adolescents' students
Do a statistical test on H1:
If H1 false, then H0 is true
If H1 is true, then H0 is false
▪ Based on the results for the sample, inferences/conclusions are made
about whether H0 is true and should be retained (there is NO
difference/relationship) or
▪ Whether H0 is false and should be rejected (there IS a
difference/relationship)
▪ Test our hypothesis with data

▪ Inferential statistics is the set of procedures used to

determine whether the research results support your
hypothesis

▪ However, there is always a probability that your results are BY

CHANCE and not because there truly is a relationship
between the variables
▪ Inferential statistics reaches conclusions that extend beyond the
immediate data.
▪ For instance, to infer from the sample data what the population
might think.
▪ Make judgments whether an observed difference between groups
is a dependable one or if it might have happened by chance in the
study.
The goal of the study is to evaluate the effect of a treatment.
A sample is selected from the population, and the treatment is administered to
the sample.
If, after treatment, the individuals in the sample are noticeably different from
individuals in the original population, then we have evidence that the treatment
does have an effect.
▪ In order to answer the research question….
▪ Statistical tests are conducted to:
▪ Either RETAIN or REJECT H0

▪ We REJECT H0 only if our research result is so unusual that it

is HIGHLY UNLIKELY that it occurred due to chance
▪ We then conclude that the differences are due to the factor
under investigation – the independent variable
▪ A null hypothesis can be true or false
▪ H0 is true →
 There is not a difference/relationship
 If the null is true THEN the research hypothesis is false.
 RETAIN THE NULL HYPOTHESIS
 Result is NOT statistically significant

▪ H0 is false →
 There is a difference/relationship
 If the null is false then the research hypothesis is true.
 REJECT THE NULL HYPOTHESIS
 Result IS statistically significant
 STATISTICAL SIGNIFICANCE

H0 is true H0 is false
If the null is true THEN the If the null is false then the
H1 is false H1 is true.

RETAIN THE H0 REJECT THE H0

Result is NOT Result IS

statistically significant statistically significant

There is NOT a There IS a

difference/relationship difference/relationship
If H0 is false
• The H0 is rejected
• There is a statistically significant result
• and thus, the research hypothesis IS
supported
• Differences/relationships in data are NOT
due to chance, but can be attributed to
the factor being studied (IV)

If H1 is false
• Then H0 is retained
• There is a NOT a statistically significant
result
• And thus, the research hypothesis is NOT
supported
• Differences/relationships in data are not
due to the IV but due to chance
 NO
 There is always a degree of error that cannot be controlled.
1. You can never be 100% sure that the sample is an
identical reflection of the population, i.e. perfectly
representative
2. There are always other influences on your sample.

▪ There is always a level of risk (LEVEL OF SIGNIFICANCE)

▪ Level of Significance = the risk you are willing to take
▪ Also known as alpha
▪ Symbol = α
▪ Standard of evidence required to make inferences
▪ Based on the principles of probability (ρ-values)
▪ Generally 0.05 (5%) → 5% chance that the change/difference is
not due to the intervention
▪ Or 0.01 (1%) → 1% chance that the change/difference is not due to
the intervention
▪ Or 0.001 (0.1%) chance that the change/difference is not due to the
intervention
TWO EXAMPLES
▪ The academic performance of
students who meditate will be better
than the academic performance of
students who don’t meditate
OR
▪ There will be a difference in the
academic performance between
students who meditate and those
who don’t meditate
▪ THEY NOW BECOME ➣
▪ The academic performance of students who
meditate will be statistically significant better
than the academic performance of students who
don’t meditate
▪ H1: µ1 > µ2

where
µ1 = the average academic performance mark of the
students who meditate and
µ2 = the average academic performance mark of the
students who do not meditate
▪ A directional H1
▪ There will be a statistical significant difference in
the academic performance between students who
meditate and those who don’t meditate
H1: µ1 ≠ µ2

where
µ1 = the average academic performance mark of the
students who meditate and
µ2 = the average academic performance mark of the
students who do not meditate
▪ A non-directional H1
▪ If my research results show I must REJECT H0
▪ It means that H1 is true
▪ And there IS a difference between the two groups I am
studying
▪ BUT then I always report my alpha level
▪ So that others can understand there is still a small risk that I
got these results by chance
▪ And that chances are 5% or 1% that H1 is not true
 STATISTICAL SIGNIFICANCE
VS. MEANINGFULNESS
Statistical significance is:

Not end-all goal of research

Null results can be very important

It still does not indicate what the practical value is

You RETAIN the H0 You REJECT the H0

☺ 
and the null hypothesis was true. and the null hypothesis was true.
In other words you made the right In other words you made the wrong
decision. decision. This is called a Type 1
error () - Level of significance

 ☺
and the null hypothesis was false. and the null hypothesis was false.
In other words you made the wrong In other words you made the right
decision. This is known as a Type II decision. This is known as the Power
error (β). of your test.
Type 1 errors
- H0 is true: there is NO difference between groups,
BUT the researcher rejects it, saying there is a
difference between the groups
- The probability of a Type 1 error occurring is
equal to 
Type 2 errors
- H0 is false: there IS a difference between groups,
BUT the researcher accepts it, saying there is no
difference between the groups
- The probability a Type 2 error occurring is equal to β
Re: Type 1 errors

▪ If the level of significance  = 0.05,

▪ there is a 5% chance that the H0 will be rejected
while it was actually true.

▪ If  = 0.01, there is a 1% chance that you will reject

the H0 when it is actually true.
 A conclusion is made about the population on the basis
of the results that have been obtained for the sample.

1. The researcher collects a representative sample

2. The data is collected (intervention and/or
questionnaires)
3. A statistical formula is used to test the data to decide if
the result is significant or not
4. A conclusion is made about the population
1. State the H0 and the H1
2. Set the level of risk: α= 0.05
3. Select the appropriate test statistic
4. Calculate the obtained value (test statistic)
5. Determine the critical (probability - ρ) value from table
NB – you must download this from BB for the next lecture
6. Compare the obtained and the critical values
7. Make your decision: REJECT or RETAIN the H0
8. Interpret briefly: