Academic Skills 2 – knowledge clips slides
Introduction to Social Research Design
Systematically support everything you say through data and evidence and be able to account for how
you come to your conclusions is key! Social research design: choose evidence you want to use choose method formulate research
question. Selecting data/evidence: what kind of project do you want to do? What kind of evidence do you
want to use?
Secondary data: existing data, library sources.
Primary data: new data.
Data analysis: what kind of method do you want to use?
Quantitative methods ‘numbers as data’.
o Statistical analysis: averages, outliers…
o Numerical relationships
o Numbers = broad – counting words, code feelings as numbers
o E.g. survey
Qualitative methods ‘words as data’
o Words, images…
o More naturalistic understanding of what participants express
o More open/no predefined categories
o E.g. interview
Mixed methods ‘both quantitative as qualitative’
o Can be a way of overcoming problems/limitations
o Can be a mess
o More is not always better
o Philosophical issues are key
A research can be inductive or deductive.
o Open research questions
o Usually qualitative
Deduction ‘driven by theory’
o Tests theory through the use of data
o Based on clear hypotheses/propositions
o Usually quantitative
Ontology: our understanding of the world/reality.
Epistemology: our understanding of how we can study the world/reality and generate good
Method: the tools you use and deem appropriate to actually study (a chunk of) reality.
Research questions provide sharp focus, which is essential to research design. Quantitative research
questions concerns selecting variables and makes predictions about expected relationships between
variables and they use hypotheses.
Null hypothesis: assumption no relationship exists.
Alternative/directional hypothesis: assumption a relationship exists in a specific direction and
is based on previous theory or literature.
Alternative non-directional hypothesis: assumption a relationship exists, but unclear in which
direction. -> when there is less known about the topic.
Large scale, less depth
Qualitative research questions are more open, often starting from a general, central question, after
which, with the use of key concepts, a more specific question is being formulated. E.g.:
How do people construct the body in medical interactions? ->
People: chronic pain patients
The body: the chronically ill body
Medical interaction: doctor-patient interaction in a pain clinic ->
How do chronic pain patients talk about their chronically ill bodies in doctor-patient
interactions in a pain clinic?
No predefined answers like with a survey
Small scale, more depth
It doesn’t make assumptions about the concepts/topic of the research.
While formulating research questions, looking at language can help. ‘What’ is descriptive and can be
both qualitative and quantitative. ‘How’ implicates complex answers, which is qualitative, but ‘how
many’ is quantitative. ‘Why’ can be both quantitative and qualitative and in case of quantitative
research it is often about a topic about which we already know a lot or we want to confirm theories. Week 2
Surveys Stratified sampling: dividing people in categories based on characteristics of which a random sample
is taken from each category. Reliability is about getting stable answers, e.g. weighing a cat on different scale getting the same
weight. Validity is about whether the right method is used and if it measures what it is supposed to
measure, e.g. wanting to know the health of the cat, only weighing is not enough.
Survey: research design with cross-sectional approach, seeking patterns in quantitative data.
Representation: sampling = taking a sample from the population representation (to which
extent does the sample reflects the population) of the population. If the sample is random
and of adequate size, then the external validity improves and the chances of generalisation
o Random sampling: 1. Set the population. 2. Choose a sampling frame. 3. Select units
from frame (with possibility of stratification). 4. Approach units.
Measurement: reliability = internal consistency and validity = construct validity.
A survey is focussing on a single point in time + a large sample to allow variety in population. You
cannot research the whole population (too large + takes too much time), so sampling occurs.
Questionnaire: asking subjects to respond to a range of questions.
Constructing a questionnaire:
1) Conceptualisation – literature review, focus group, interviews, experts.
2) Item pool generation.
3) Item selection with expert evaluation & literature.
4) Questionnaire pre-testing – cognitive interviews, field test.
5) Quantitative pre-test – item analysis, factor analysis, Cronbach’s alpha.
Item analysis: check for patterns in data to stand out.
Factor analysis: verify which items are measuring specific dimensions within a context.
Cronbach’s alpha: internal consistency of the scale.
Possible values: 0-10
Optimal: >0.7 and <0.9
Always first reverse score items with negative wording. Week 3
Causality & Experiments
An important focus in quantitative research is variables: characteristics of units studied that very. For
example: age, gender, education. Causality is a causal relation between two variables X=cause, Y=effect.
Survey (and other non-
There are some criteria for causality.
Correlation: X and Y are associated – constant conjunction: statistically significant causalation
between X and Y.
o Strength: the extent of linear pattern.
o Direction: negative or positive.
Time order: X precedes Y in time – in experiments X is manipulated.
o Non-experimental research: time series / longitudinal research = multiple
measurements overtime are taken to monitor the changes in the variables studied. It
allows us to detect pattern overtime which can indicate causal relation, but it is
possible that a 3rd variable influences.
No alternative explanation: in an experiment, the experiment and control group as equal as
possible random assignment to groups.
To improve external validity at a survey, random sampling is important, and at an experiment,
random assignment is important. Theory: a set of interrelated constructs (variables), definitions and propositions that presents a
systematic view of phenomena by specifying relations among variables, with the purpose of
explaining natural phenomena.