The Importance of Quantitative Methods in Social Science

QS101 – Introduction to Quantitative
Methods in Social Science
Week 2: The Importance of Quantitative Methods in Social
Sciences
Stella Chatzitheochari
Assistant Professor in Sociology
Course Director: BA Sociology and Quantitative Methods
[email protected]
Welcome to Warwick!
Intros & housekeeping
Questions for you:
What are quantitative methods?
Do you know any other methods?
So: why are QM important for you?
• Data are everywhere!
• Evidence-based policy
• You as a citizen (Lies, lies and
damned statistics!)
• Job market: huge demand
• Large-scale data and statistical
analysis are essential for the scientific study
of social groups and societies
(this is the focus of today’s lecture)
Research in the Social Sciences
“The natural sciences talk about
their results. The social sciences
talk about their methods”
“paradigm wars”,
qualitative/quantitative divide
The nature of foci of social
research is abstract rather than
concrete. Think about
democracy or poverty.
Generally, social research wants to:
 describe the structure and organization of
societies
 Identify regularities that are worth explaining
 Construct and test explanations for such
patterns and regularities
 Address social problems, implement changes
So we are interested in theories = stories on why
certain things happen. Causes and effects.
But remember: the term theory is often used
loosely by social scientists…(well, by everybody)
Theories and explanations
* Testable stories about relationship
between different things (hypotheses)
Theories are falsified, not verified & theories are
never final – “the game of science is without end”
Two types of reasoning
Deductive
Inductive
Theory
↓
Hypothesis
↓
Observations
↓
Confirmation
Theory
↑
Hypothesis
↑
Pattern
↑
Observations
So where do quantitative methods and statistical
inference fit it?
Induction in everyday life and in
the social sciences
 my friend’s love life  all men are cheaters
 Greek recession and the rise of Golden Dawn 
austerity provokes right-wing populism
BUT
• Is generalization (external validity) appropriate in
the above cases?
• Induction often relies on an inadequate number of
cases that do not represent the population of
interest (i.e. men, countries)
Sampling and statistical inference
 We cannot study everyone from the population of
interest. So we select a sample.
 Sampling is vital for inductive logic; it allows us to
make inferences from the sample to the
population
 Representative, large-scale samples obtained by
social surveys
 The key point is that they capture sufficient
variability in the characteristics that we are
interested in (variables)
 Falsification with a degree of certainty
UK Income inequality trends
And implications (?)
Examples of research questions
in quantitative social sciences
 What proportion of the UK population is poor?
 Has social mobility declined over time?
 Is quality of life higher in liberal welfare regimes?
 Does marriage make people happier?
 Does early political socialization predict political
identification in adulthood?
Relationships are everywhere (?)
Making sense of relationships…
Married people are
consistently found to
have better health
than unmarried
people in social
surveys. Does this
mean marriage
improves health?
What do you think?
Making sense of relationships…
Not, necessarily. Married people may have
been healthier than divorced people even
before they were married. And those with
worse health may have been less likely to
marry. But it may also be that married people
are more wealthy and this is the main driver
behind their better health. Hmm…oh, and…
Correlation is not causation!
We can only say that A causes B if:
1. There is a link between A and B
2. A took place before B (what happened first?)
3. The relationship is not driven by C (another
variable)
Causation and quantitative
methods
 detect spurious correlations (third variable
involved?)
 establish temporal order between causes and
effects
 Multiple causal factors can be distinguished
 Test potential mechanisms that bring about
an identified effect
…..but measurement limitations…
Small N-cases, case studies: in depth picture
but low generalizability
Methodological pluralism
 Each research question should be tackled on
the basis of its own particular features
 Mixed Methods Research (MMR): qual and
quant methodologies can be combined to
provide rich insights (vs methodological
purism focusing on irreconcilable differences)
 Examples: study of riots & social mobility of
different ethnic groups & even social survey
design
MMR in the social sciences
On intellectual craftsmanship
“Be a good craftsman: avoid any rigid set
of procedures. Above all, seek to use the
sociological imagination. Avoid the
fetishism of method and
technique…Let every man be his own
methodologist; let every man be his own
theorist; let theory and method again
become part of the practice of a craft.
Stand for the primacy of the individual
scholar; stand opposed to the
ascendancy of research teams of
technicians. Be one mind that is on its
own confronting the problems of man
and society” C.W.Mills