All Rights Reserved.
Some of this work has been summarised from the textbook Research in Practice:
Applied Methods for the Social Sciences by Martin Terre Blanche, Kevin
Durrheim, and Desmond Painter.
ISBN 9781485102687
It includes notes taken based on lecture slides from relevant topic lecturers, all of
whom are named in the appropriate section. These notes are elaborations based on
the author’s own understanding.
This document, including all content and materials contained within, is the
intellectual property of Arièlle Denis, with thanks to Kayla Paul and Ruby Webb
for their contributions on the textbook notes and further summaries.
No part of these notes may be reproduced, distributed, or transmitted in any form
or by any means, including photocopying, recording, or other electronic or
mechanical methods, without the prior written permission of the author, except in
the case of brief quotations embodied in critical reviews and certain other
non-commercial uses permitted by the copyright law.
For permission requests, please contact the author at
© 2026 Arièlle Denis,
University of Cape Town. All rights reserved.
, RESEARCH IN PSYCHOLOGY I
PSY2015F | SEMESTER 1
Quantitative Methods | Chapters 1-4 and 7-8 of Terre Blanche, Durrheim and
Painter (2006) textbook
by Cathy Ward
In the social sciences, we use a hypothetico-deductive model.
→ Deductive: Start from the theory, then hypothesise, observe and interpret.
→ Inductive: Start from the observation, then interpret, theorise and hypothesise.
Paradigms in Social Science
→ Systems of thinking and practice.
● Logical positivism paradigm
○ Assumes that there is a regular, external and stable reality
○ Assumes laws from physical sciences, because we can define these
○ Assumes that research is observed and detached (but we know this is not the case)
● Interpretive paradigm
○ Subjective reality is internal
○ Usually empathetic, attached
○ Does not assume neutrality
● Constructivist paradigm
○ Reality is socially constructed
○ Usually discourse (talking) analysis and power dynamics
○ Is often a bit suspicious and critical
○ Makes use of advanced quantitative techniques
,→ Hypotheses: Falsifiable predictions which give direction to the researcher to answer their
research question and outline the anticipated relationship between two or more variables. They
are derived from theory and must be clearly stated and testable.
→ Theories: A set of contents and propositions intended to describe and explain some aspect of
experience; A substantiated explanation of some aspect of the natural world.
● Summarises hypotheses that have been tested
● Direct future research
Types of Research
3 Typologies:
1. Phenomenon
● Exploratory: Preliminary investigations into unknown areas
● Explanatory: Casual phenomena
● Descriptive: Describes phenomena
2. Application
● Basic: Research advances worldly knowledge more directly
● Applied: Research has a practical application
3. Method
● Quantitative: Using numbers and statistics to make broad, generalisable comparisons
● Qualitative: Using words and themes to describe a phenomenon in depth and detail
● Mixed: Both quantitative and qualitative methods, which are used increasingly in modern
science
Validity in Research Designs
Integrity matters – both in the results reported on and speculated on. It is necessary to discuss
validity as if research results are unexpected when initially obtained or when replicated, we may
need to expand our research methods or replicate the study or experiment thereafter.
4 Types of Validity:
1. Internal validity
Basically: Did X cause Y? (Cause-and-effect relationship)
→ Causality: Correlations may be genuine, but not necessarily imply a cause-and-effect
relationship. Correlation does NOT mean causation, and can be “spurious” (unrelated or due to
another third/confounding variable). As we may measure many variables, it is very likely to
come to statistically significant conclusions by chance.
, Conditions for Causality:
● Covariation: If one variables varies, so does the other. These are usually correlated.
○ The experiment measures whether a change in one variable is associated with a
change in another variable.
● Temporal precedence: One thing comes before the other. This strengthens our ability to
say that A causes B.
○ The experiment manipulates variables, as the researcher decide which one comes
first.
● Elimination of 3rd variables: Anything that could change the cause or effect.
○ The researcher controls extraneous factors that might influence the outcome of the
experiment.
Experimental control:
● The only difference between the experimental and control groups is what the
experimentor/researcher does to them.
● This is partly achieved through random assignment of participants to experimental and
control groups AND through carefully thinking through possible confounding variables.
→ Experimental method: Manipulates the independent variable
● Experimental studies HIGH in internal validity because they identify covariation, are set
up with temporal precedence, and the design inherently controls for many possible
confounds
→ Correlational method: Measures two or more characteristics of the same
individual/phenomenon and computes the correlation
● Correlational studies LOW in internal validity because they identify covariation, but
typically cannot identify temporal precedence and struggle to eliminate confounding
variables
Quasi-experimental studies:
e.g. How head injured in young children affect educational outcomes
1. Think about the variables associated covarying with educational outcomes, e.g. age,
gender or grade
2. Find children with head injuries
3. For each child with a head injury, find a child of identical age, grade, gender and with
parents of the same socioeconomic status, etc.
The child with the head injury will be the “experimental” group and the matched child will be
the “control”. But because there is no randomisation, we cannot be sure the experimental and
control groups are equal.
● This kind of study is often done in neuropsychology – there is no set standard for the
times replicated, and we still cannot establish a causal relationship.
Some of this work has been summarised from the textbook Research in Practice:
Applied Methods for the Social Sciences by Martin Terre Blanche, Kevin
Durrheim, and Desmond Painter.
ISBN 9781485102687
It includes notes taken based on lecture slides from relevant topic lecturers, all of
whom are named in the appropriate section. These notes are elaborations based on
the author’s own understanding.
This document, including all content and materials contained within, is the
intellectual property of Arièlle Denis, with thanks to Kayla Paul and Ruby Webb
for their contributions on the textbook notes and further summaries.
No part of these notes may be reproduced, distributed, or transmitted in any form
or by any means, including photocopying, recording, or other electronic or
mechanical methods, without the prior written permission of the author, except in
the case of brief quotations embodied in critical reviews and certain other
non-commercial uses permitted by the copyright law.
For permission requests, please contact the author at
© 2026 Arièlle Denis,
University of Cape Town. All rights reserved.
, RESEARCH IN PSYCHOLOGY I
PSY2015F | SEMESTER 1
Quantitative Methods | Chapters 1-4 and 7-8 of Terre Blanche, Durrheim and
Painter (2006) textbook
by Cathy Ward
In the social sciences, we use a hypothetico-deductive model.
→ Deductive: Start from the theory, then hypothesise, observe and interpret.
→ Inductive: Start from the observation, then interpret, theorise and hypothesise.
Paradigms in Social Science
→ Systems of thinking and practice.
● Logical positivism paradigm
○ Assumes that there is a regular, external and stable reality
○ Assumes laws from physical sciences, because we can define these
○ Assumes that research is observed and detached (but we know this is not the case)
● Interpretive paradigm
○ Subjective reality is internal
○ Usually empathetic, attached
○ Does not assume neutrality
● Constructivist paradigm
○ Reality is socially constructed
○ Usually discourse (talking) analysis and power dynamics
○ Is often a bit suspicious and critical
○ Makes use of advanced quantitative techniques
,→ Hypotheses: Falsifiable predictions which give direction to the researcher to answer their
research question and outline the anticipated relationship between two or more variables. They
are derived from theory and must be clearly stated and testable.
→ Theories: A set of contents and propositions intended to describe and explain some aspect of
experience; A substantiated explanation of some aspect of the natural world.
● Summarises hypotheses that have been tested
● Direct future research
Types of Research
3 Typologies:
1. Phenomenon
● Exploratory: Preliminary investigations into unknown areas
● Explanatory: Casual phenomena
● Descriptive: Describes phenomena
2. Application
● Basic: Research advances worldly knowledge more directly
● Applied: Research has a practical application
3. Method
● Quantitative: Using numbers and statistics to make broad, generalisable comparisons
● Qualitative: Using words and themes to describe a phenomenon in depth and detail
● Mixed: Both quantitative and qualitative methods, which are used increasingly in modern
science
Validity in Research Designs
Integrity matters – both in the results reported on and speculated on. It is necessary to discuss
validity as if research results are unexpected when initially obtained or when replicated, we may
need to expand our research methods or replicate the study or experiment thereafter.
4 Types of Validity:
1. Internal validity
Basically: Did X cause Y? (Cause-and-effect relationship)
→ Causality: Correlations may be genuine, but not necessarily imply a cause-and-effect
relationship. Correlation does NOT mean causation, and can be “spurious” (unrelated or due to
another third/confounding variable). As we may measure many variables, it is very likely to
come to statistically significant conclusions by chance.
, Conditions for Causality:
● Covariation: If one variables varies, so does the other. These are usually correlated.
○ The experiment measures whether a change in one variable is associated with a
change in another variable.
● Temporal precedence: One thing comes before the other. This strengthens our ability to
say that A causes B.
○ The experiment manipulates variables, as the researcher decide which one comes
first.
● Elimination of 3rd variables: Anything that could change the cause or effect.
○ The researcher controls extraneous factors that might influence the outcome of the
experiment.
Experimental control:
● The only difference between the experimental and control groups is what the
experimentor/researcher does to them.
● This is partly achieved through random assignment of participants to experimental and
control groups AND through carefully thinking through possible confounding variables.
→ Experimental method: Manipulates the independent variable
● Experimental studies HIGH in internal validity because they identify covariation, are set
up with temporal precedence, and the design inherently controls for many possible
confounds
→ Correlational method: Measures two or more characteristics of the same
individual/phenomenon and computes the correlation
● Correlational studies LOW in internal validity because they identify covariation, but
typically cannot identify temporal precedence and struggle to eliminate confounding
variables
Quasi-experimental studies:
e.g. How head injured in young children affect educational outcomes
1. Think about the variables associated covarying with educational outcomes, e.g. age,
gender or grade
2. Find children with head injuries
3. For each child with a head injury, find a child of identical age, grade, gender and with
parents of the same socioeconomic status, etc.
The child with the head injury will be the “experimental” group and the matched child will be
the “control”. But because there is no randomisation, we cannot be sure the experimental and
control groups are equal.
● This kind of study is often done in neuropsychology – there is no set standard for the
times replicated, and we still cannot establish a causal relationship.
