Key concepts Social Research Methodology
Week 1
Building blocks in theory:
1. Concept: abstract term that describes a phenomenon
2. Law (considered outdated): regularity of facts (observation
3. Postulate: basic assumption in a theory
4. Proposition / hypothesis: explanatory relationship between concepts
Grand theories: very abstract therefore difficult to use in social sciences,
large scope.
Middle range theories: attempt to understand and explain a particular aspect
of the social world. Very useful for empirical research and has a limited scope.
Background literature: focussed search for theories, empirical research and
has a limited scope.
Fact collection: hardly any theory, ad-hoc selection (when facts are needed
or when there was a particular purpose)
Induction: from the empirical to the theoretical (observations/findings
theory) and the inductive approach uses verification logic.
Verification logic: A1, A2, A3, A4 and A5 all have property P = all A have
property P
Deduction: from theoretical to the empirical (theory observations/findings)
and the deduction approach uses falsification logic.
Falsification logic: all A have property P, K is an A = K has property P
Inductive empirical cycle: problem observation theory evaluation
repeat.
Problem of induction: how many observations are needed to arrive at a
general statement.
Deductive empirical cycle: problem theory hypothesis observation
test evaluation repeat.
The hypothesis is the part of the theory that can be tested!
Complete empirical cycle: problem induction (observation theory
evaluation) hypothesis deduction (theory observation) test
evaluation repeat.
Pragmatic empirical cycle: complete empirical circle, but how it happens in
real life.
Causality: A causes B
Correlation: A relates to B, but does not necessary causes it.
Field experiment: experiment that takes place in a real life setting.
Laboratory experiment: experiment that takes place in a laboratory or
contrived setting.
(True) classical experiment: to conduct a true experiment, you need to
manipulate the independent variable in order to determine whether it has an
influence on the dependent variable. This design is often referred to as the
randomized experiment or randomized control trial (RTC).
Quasi experiment: experiments that have some experimental characteristics.
, Pre-test/post-test design: a design in which the same assessment measures
are given to participants both before and after the experiment/treatment.
Reliability: are measurements consistent? Concerned with whether we would
get the same result if we repeated the experiment under the same conditions.
o The data cannot be unreliable data can be invalid. The
measurement is reliable or not.
Replicability: is the procedure in the whole research explained clearly
enough? Can this research be replicated?
o With sufficient replicability, the experiment can be replicated elsewhere,
but not in the same group.
Validity (quantitative research): how accurate are the findings?
o Measurement/construct validity
Are you measuring what you think you are measuring?
Does a measure for assessing a concept really reflect that
concept?
EXAMPLE: does IQ measure intelligence?
o Internal validity
Is a conclusion which proposes a causal relationship between
two variables – does x cause y – convincing.
The impact is the independent variable and the effect is the
dependent variable.
Possible problems due to external factors:
Selection (not random): people’s willingness, ability or
financial skills
History effect (reason control group): something that
could happen, such as a war breaking out or gas prices
rising
Test effect (reason control group): simply doing the test
might have an effect on how they later act upon this
Instrumentation: application of different tests
Mortality: people might pass away or move out of the
area
Development
Direction of causality
o External validity
Can the results of a study be generalised to a larger population?
This is one of the main reasons why quantitative researchers
want a representative sample.
Possible external factors:
Selection (not random)
History effect
Specificity of locality
Reactivity
Test effect
o Ecological validity
, Are findings meaningful to everyday lives (natural social
settings)?
EXAMPLE: the unnaturalness of a laboratory experiment could
indicate that the findings have low ecological validity.
Possible problems:
Contamination through enormous intrusion of people’s
life’s.
o Inferential validity
Concerns whether the inferences that researchers have made,
and the conclusions that they have drawn are warranted by the
research and its findings.
Often neglected!
Cross-sectional design (often referred to as survey-design)
o A cross-sectional design involves collecting data on a sample of cases
at a single point in time. When a person completes a questionnaire,
which may contain more than fifty variables, the answers are supplied
at the same time.
o Usually a large number of cases is selected, because this causes for
more opportunities to find variations in variables.
o The researcher can only identify relationships between variables (the
variables cannot be manipulated unambiguity of causal direction).
So, if a causal relation is uncertain, you can only say that the variables
are related.
o Reliability and measurement validity within the cross-sectional design:
relates to the quality of the measures used to get at the concepts.
o Replicability within the cross-sectional design: likely to be present.
o Internal validity within the cross-sectional design: produces association
rather than causal inferences.
o External validity within the cross-sectional design: when random
sampling methods are used, the external validity tends to be strong.
When non-random sampling methods are used, the external validity
tends to be weak.
o Ecological validity within the cross-sectional design: likely to be weak
because of the use of research instruments (self-completion surveys
and structured observation schedule for example).
Longitudinal design
o Collecting data from participants more than once.
o Not very widely used in social research because of time and costs.
o Panel study longitudinal design: a sample (often randomly selected
national one), is the focus of data collection on at least two ore more
occasions.
o Cohort study longitudinal design: either an entire cohort of people or
a random sample of them is selected as the focus of data collection.
The cohort is made up of people who share certain
characteristics, this is how it differs from a panel study. This
could be: being born in the same week, being unemployed.
Week 1
Building blocks in theory:
1. Concept: abstract term that describes a phenomenon
2. Law (considered outdated): regularity of facts (observation
3. Postulate: basic assumption in a theory
4. Proposition / hypothesis: explanatory relationship between concepts
Grand theories: very abstract therefore difficult to use in social sciences,
large scope.
Middle range theories: attempt to understand and explain a particular aspect
of the social world. Very useful for empirical research and has a limited scope.
Background literature: focussed search for theories, empirical research and
has a limited scope.
Fact collection: hardly any theory, ad-hoc selection (when facts are needed
or when there was a particular purpose)
Induction: from the empirical to the theoretical (observations/findings
theory) and the inductive approach uses verification logic.
Verification logic: A1, A2, A3, A4 and A5 all have property P = all A have
property P
Deduction: from theoretical to the empirical (theory observations/findings)
and the deduction approach uses falsification logic.
Falsification logic: all A have property P, K is an A = K has property P
Inductive empirical cycle: problem observation theory evaluation
repeat.
Problem of induction: how many observations are needed to arrive at a
general statement.
Deductive empirical cycle: problem theory hypothesis observation
test evaluation repeat.
The hypothesis is the part of the theory that can be tested!
Complete empirical cycle: problem induction (observation theory
evaluation) hypothesis deduction (theory observation) test
evaluation repeat.
Pragmatic empirical cycle: complete empirical circle, but how it happens in
real life.
Causality: A causes B
Correlation: A relates to B, but does not necessary causes it.
Field experiment: experiment that takes place in a real life setting.
Laboratory experiment: experiment that takes place in a laboratory or
contrived setting.
(True) classical experiment: to conduct a true experiment, you need to
manipulate the independent variable in order to determine whether it has an
influence on the dependent variable. This design is often referred to as the
randomized experiment or randomized control trial (RTC).
Quasi experiment: experiments that have some experimental characteristics.
, Pre-test/post-test design: a design in which the same assessment measures
are given to participants both before and after the experiment/treatment.
Reliability: are measurements consistent? Concerned with whether we would
get the same result if we repeated the experiment under the same conditions.
o The data cannot be unreliable data can be invalid. The
measurement is reliable or not.
Replicability: is the procedure in the whole research explained clearly
enough? Can this research be replicated?
o With sufficient replicability, the experiment can be replicated elsewhere,
but not in the same group.
Validity (quantitative research): how accurate are the findings?
o Measurement/construct validity
Are you measuring what you think you are measuring?
Does a measure for assessing a concept really reflect that
concept?
EXAMPLE: does IQ measure intelligence?
o Internal validity
Is a conclusion which proposes a causal relationship between
two variables – does x cause y – convincing.
The impact is the independent variable and the effect is the
dependent variable.
Possible problems due to external factors:
Selection (not random): people’s willingness, ability or
financial skills
History effect (reason control group): something that
could happen, such as a war breaking out or gas prices
rising
Test effect (reason control group): simply doing the test
might have an effect on how they later act upon this
Instrumentation: application of different tests
Mortality: people might pass away or move out of the
area
Development
Direction of causality
o External validity
Can the results of a study be generalised to a larger population?
This is one of the main reasons why quantitative researchers
want a representative sample.
Possible external factors:
Selection (not random)
History effect
Specificity of locality
Reactivity
Test effect
o Ecological validity
, Are findings meaningful to everyday lives (natural social
settings)?
EXAMPLE: the unnaturalness of a laboratory experiment could
indicate that the findings have low ecological validity.
Possible problems:
Contamination through enormous intrusion of people’s
life’s.
o Inferential validity
Concerns whether the inferences that researchers have made,
and the conclusions that they have drawn are warranted by the
research and its findings.
Often neglected!
Cross-sectional design (often referred to as survey-design)
o A cross-sectional design involves collecting data on a sample of cases
at a single point in time. When a person completes a questionnaire,
which may contain more than fifty variables, the answers are supplied
at the same time.
o Usually a large number of cases is selected, because this causes for
more opportunities to find variations in variables.
o The researcher can only identify relationships between variables (the
variables cannot be manipulated unambiguity of causal direction).
So, if a causal relation is uncertain, you can only say that the variables
are related.
o Reliability and measurement validity within the cross-sectional design:
relates to the quality of the measures used to get at the concepts.
o Replicability within the cross-sectional design: likely to be present.
o Internal validity within the cross-sectional design: produces association
rather than causal inferences.
o External validity within the cross-sectional design: when random
sampling methods are used, the external validity tends to be strong.
When non-random sampling methods are used, the external validity
tends to be weak.
o Ecological validity within the cross-sectional design: likely to be weak
because of the use of research instruments (self-completion surveys
and structured observation schedule for example).
Longitudinal design
o Collecting data from participants more than once.
o Not very widely used in social research because of time and costs.
o Panel study longitudinal design: a sample (often randomly selected
national one), is the focus of data collection on at least two ore more
occasions.
o Cohort study longitudinal design: either an entire cohort of people or
a random sample of them is selected as the focus of data collection.
The cohort is made up of people who share certain
characteristics, this is how it differs from a panel study. This
could be: being born in the same week, being unemployed.
