Organizational session /
Introduction to quantitative analysis
Lukáš Lehotský & Petr Ocelík
ESS401 Social Science Methodology / MEB431 Metodologie sociálních věd
Outline
• Course plan
• Background: quantitative research and statistics
• Data, variable types and levels of measurement
Methodology/methods courses
• Logic built around gradual expansion of method knowledge.
• Extension courses
– MEB433 – Introduction to quantitative data analysis (Fall 2017)
– MEB434 – Social network analysis (Fall 2017)
– HEN645 – Experiment v sociálních vědách (Fall 2017)
– MEB421 – Vybrané metody výzkumu v mezinárodních vztazích (Fall 2017)
• Support courses
– ESS412 – Academic skills review (Spring 2017)
• Methods not particularly entertaining but absolutely crucial.
Course requirements
• Read required readings
• Attend lectures
• 2 groups
• Write research proposal
• Write feedback report
• Deliver presentation of the research proposal
• Pass a written exam
Course requirements
• Deadlines
– April 20 – submission of proposal (group 1)
– April 22 – submission of feedback report (group 2)
– April 27 or 28 (TBC) – optional consultation with lecturers
(group 1)
– April 30 – submission of final version of proposal (group 1)
– May 11 – submission of proposal (group 2)
– May 13 – submission of feedback report (group 1)
– May 18 or 19 (TBC) – optional consultation with lecturers
(group 2)
– May 21 – submission of final version of proposal (group 2)
Grading
• Proposal: 20
• Final exam: 20
• Grading
– A 40 - 38 points
– B 37 - 36 points
– C 35 - 34 points
– D 33 - 32 points
– E 31 - 30 points
– F less than 30 points
(Social) research
• SR aims to advance knowledge in its field by using
scientific method.
• The procedures are public.
• The conclusions are uncertain.
• The content of science is its method.
Keohane, King, and Verba 1994
Two meanings of methodology
• 1) Methodology as a scientific discipline that studies
instruments and practices of scientific research.
• 2) Methodology as a system of rules of scientific conduct
which conditions specific research design.
• There is plurality of scientific methodologies.
• Methodology connects theory and investigated phenomena.
• Method: an instrument for collection and/or analysis of data
that follows specific theoretical-methodological trajectory.
Quantitative and qualitative methodology
• Just a didactic tool! (Borderline cases such as QCA)
• Outdated dichotomy
Quantitative Qualitative
Use of mathematical apparatus
Numerical data
Mathematical apparatus not necessary
Numerical data not necessary
Large(r) number of cases
Low(er) number of variables
Low(er) number of cases
Large(r) number of variables
Standardization and reduction of info
Fixed research designs
Idiosyncrasy and richness of info
Flexible research designs
Higher reliability
Lower validity
Higher generalizability
Lower reliability
Higher validity
Lower generalizability
Quantitative methodology
• QM is a systematic investigation of phenomena using
mathematical techniques.
• QM explains phenomena by collecting numerical
data that are analyzed using mathematically based
methods (Cresswell 1994).
• Always: numerical data, mathematical framework.
• Typically: empirical focus, large(r) N.
Ranganathan et al. 2014
Quantitative research = statistics?
• Not really.
• Formal modeling: e.g. game theory
• Generative methods: e.g. agent-based modeling.
• But: statistics most widespread and influential.
Box-Steffensmeier et al. 2008
What do we mean by statistics?
• Statistics consists of a body of methods for obtaining
and analyzing data (Agresti & Finlay 2009: 3).
• Statistics is a branch of mathematics dealing with the
collection, analysis, interpretation, and organization
of data (Wikipedia 2017).
• ...especially analysis of population characteristics by
inference from sampling (American Heritage Dictionary 2017).
What is statistics good for?
• Data description/exploration: quantifications,
relationships, patterns, trends, outliers...
• Theory testing: is the relationship (+ GDP  + democracy)
visible enough, or is it just random “noise”?
• Generalization to population: based on sample
estimations we can learn about a broader set of cases (a
population).
• Specific kind of an “open-source thinking”.
What is statistics good for?
• Data description/exploration:
– What is mean value of energy efficiency in the EU countries?
– Can we group the EU member states based on their energy
efficiency and liberalization of their energy market?
• Theory testing:
– Does increase in oil production decrease level of democracy?
(Resource curse theory)
• Generalization to population:
– What are attitudes of Czech adult population towards climate
change?
What is statistics good for?
• Not only academic indulgence, statistics are
everywhere – (near to) every public as well as
private organization uses numerical data.
• With rapidly progressing digitalization, there will be
(much) more...
• High demand for data analysts  increases your
chances on job market.
How to learn statistics?
• Gradually: build your skills step-by-step.
• Consecutively: more advanced concepts and
techniques are dependent on simpler ones 
do not skip!
• Via multiple sources: freely available textbooks,
lectures on YT, community forums, etc.
Data
• Concept of data usually taken for granted.
• Data is information that has been collected
and recorded.
• What we understand as data depends on our
philosophical position.
Types of data
• Qualitative vs. quantitative data.
• Typically described as words vs. numbers.
• Blaikie: all primary data start as words (Blaikie 2003).
Data
• What does data consist of?
• Case: is a unit of observation / analysis.
• Variable: is a concept that can have different
mutually exclusive values.
• Value / score: particular category or point on
a measurement scale.
Example of coding: variable stat.boredom
• Do you agree with following statement?
Statistics is boring.
• Strongly disagree = 1
• Disagree = 2
• Nor disagree nor agree = 3
• Agree = 4
• Strongly agree = 5
Case, variable, value
Kittel 2013
Ordinary usage Science Statistics
Object Unit of analysis
Unit of observation
Case
Property Attribute Variable
Specific property Attribute level Value
Case-by-variable matrix
Kittel 2013
Discrete and continuous variables
• Discrete variables: separate values without
intermediate values.
– Categories (single, married, divorced)
– Whole numbers (0, 1, 2, …)
• Continuous variables: any value over certain
interval.
– Real numbers (0, 1, 1/3, 0.333, 3.14, …)
Levels of measurement
• Categorical measurement:
– Assigns entity to a discrete category.
• Metric measurement:
– Assigns entity to a position at a given numerical
scale.
Levels of measurement
• Categorical measurement:
– Nominal
– Ordinal
• Metric measurement:
– Interval
– Ratio
Nominal measurement
• Construction of categories must be:
• Homogeneous
• Mutually exclusive
• Exhaustive
• We can arbitrarily assign numbers to
categories.
Ordinal measurement
• Defined by same conditions as nominal level
plus orders categories along some dimension.
• We can order categories, but distance
between them is not equal.
• Numbers indicate only order of categories.
Interval measurement
• Defined by same conditions as ordinal level
plus scores on a scale are at the same
distance apart.
• We can measure and compare intervals.
• No true zero: position of zero is arbitrary.
• Measures how many (counts),
not how much (ratios).
Ratio measurement
• Defined by same conditions as interval level
plus it has true zero.
• We can measure and compare ratios /
proportions.
• Degrees Celsius (interval) vs.
degrees Kelvin (ratio).
Levels of measurement: recap
Measurement level Comparison of characteristics Comparison of values Examples
Nominal same/different a = b, a ≠ b Nationality
Energy sources
Ordinal bigger/smaller a < b, a > b, a = b Level of democracy
Likert scale
Interval differences a – b = c – d Years
Temperature in °C
Ratio ratios a/b = c/d GDP
Energy consumption
Temperature in °K
Kittel 2013