Toolbox of a professional researcher Master Class on the Development of Analytical skills Žilvinas Martinaitis, 2009
Jan 12, 2016
Toolbox of a professional researcher
Master Class on the Development of Analytical skills
Žilvinas Martinaitis, 2009
How does a researcher look like?
Professional experience:
• 5 years at Public Policy and Management Institute;
• Over 30 applied research projects for the Government and the EU Commission;
Academic experience:
• 4 years of teaching experience at VU;
• PhD Student at VU.
Key questions:
• What is the difference between a researcher, fiction writer and Georg Wilhelm Friedrich Hegel?
• What is the meaning of life for a researcher?
• What does the toolbox contain:– Objectives, questions and problems; – The hypothesis;– Research methods.
How to recognize a researcher?
• They create value added.
Value added = asking important questions + providing generalizable answers + testing, if the answers are correct!
• You can actually read their papers!
Rules of thumb: KISS! Grandmother test! 5 tells; 1 and 8 rules.
Where are you?
Val
ue
add
ed/
evid
ence
bas
ed
Clear, sound style
1st year undergraduates
Georg Wilhelm Friedrich Hegel
Good fiction
Sound academic research paper
The meaning of life:
• Although the debate about THE TRUTH goes on, we attempt to get closer to it;
We raise the questions, develop and systematically test the answers (hypothesis, theories and laws).
What is the best indicator of your success?
The toolbox:
• Objectives, problems and questions;
• The solutions and the answers: hypothesis;
• Research methods.
The purpose statement of research paper
Highway to hell: • I want to write about..• I focus on Estonian
foreign policy;
• Foreign policy is a very important policy
Highway to paradise:• I seek to explain why..?• I seek to explain, why
Estonian – Russian relationships changed over the past 10 years.
• I seek to explain empirical and theoretical puzzles.
The problem/puzzle:
• Why bother: no problems, no solutions! “We fail more often not because we find the wrong solution, but because we solve the wrong problems” (Ackoff, 1974).
• A good problem: – Doesn’t lead to a question “so what?”;– Doesn’t provide obvious answers; – Looks like an interesting puzzle, which is worth
solving; – Involves empirical and/or theoretical contradictions.:
“Although ….., however……”.
The research question:• If you got the problem, the question is easy:
“Why…? How…?” • Link your question with wider theoretical debates
(the lit. review!).
High level of emigration from Lithuania
Emigration from other new EU MSs
Policy analysis: what is the role of government in regulating migration?
Academic discussion: why do people migrate?
Your paper
Developing a hypothesis• Once you have a question and a puzzle, you
need an answer/solution. • The answers should create value added,
because they are: – Based on the knowledge already available;– Empirically testable; – Generalizable to other cases.
• A hypothesis is theoretically driven statement about causal relationships between variables (the causes and effects).
• Good hypothesis could be written as follows:
A Bq
Key ingredients of a hypothesis:
Variables:• Dependent v.= phenomenon to
be explained; • Independent v.= factors
explaining the dep. v.;• Intervening v.= is caused by
indep. v. and causes dep. v.;
• Condition v. = frames antecedent conditions
Examples:• Sunshine causes
grass to grow;• Sunshine causes
grass to grow;• Sunshine causes
photosynthesis, which causes grass to grow;
• Sunshine causes grass to grow, but only when there is enough of rainfall
A hypothesis in its schematic form:
A q r Bx
C
Operationalizing your hypothesis
• Clearly define your variables
• Identify criteria for verification of the values of your variables:
• Identify observable implications of your hypothesis:
Example: left leaning coalitions promote employment security.
Research methods:
• In principle the data for testing the hypothesis could be collected in following ways:
• Experiment;
• Observation: – Large n analysis (quantitative); – One or two case studies (qualitative).
The overall logic of testing the hypothesis
Dependent variable
Variable C
Variable B
Variable A
Variable D
Comparative analysis: method of difference
The method of difference: • Choose two cases, which are similar in all
respects except for the phenomenon you want to explain (DV) and the factors explaining it (IV).
• Variables with the same values can not explain the difference in the results.
• Hence the variation in the outcomes is explained by the difference in the values of IV.
Examples: reforms in the pension system in the Czech republic and Poland (Muller 2001)
Variables Poland Czech rep.Demographic cond. ageing ageing
Sustainability of pension system
unsustainable unsustainable
Economic situation transition transition
Financial system Stable Stable
IV- government debt
large small
DV- pension system
Privately run State run
Comparative analysis: method of similarity
• It is exactly the opposite: find two cases, which are very different, except for the value of IV and DV;
Variables Case A Case B
Var. 1 yes no
Var. 2 high low
Var. 3 absent present
Var. I (IV) yes yes
Var. D (DV) yes yes
Issues and problems in comparative analysis
• Comparing
apples with
oranges
• Additional variation: for e.g. when using the method of difference you find an additional variation? use the shadow case study.
Shadow case in the method of differenceVariables Poland Czech rep. Hungary
Demographic cond.
ageing ageing ageing
Economic situation
transition transition transition
Financial system
Stable Stable Stable
Gov. coalition Right wing Left wing Left wing
IV- government debt
large small Large
DV- pension system
Privately run State run Privately run
Case studies:
• “For example…” is not a case study!
• The overall logic is the same as in comparative analysis: seek to explain variations!
• Two strongest types of case studies: – Process tracing; – Critical case studies.
Example of process tracing: electoral barrier and the number of effective parties in
Poland
1991 1993 1997
Electoral barrier % 0 5 5
Number of effective parties
13,5 9,8 4,6
Example of process tracing: electoral barrier and the number of effective parties in
Lithuania1992 1996 2000 2004
Electoral barrier % 4 5 5 (7) 5 (7)
No. of elected parties in PR tier
7 5 5 6
Total number of elected parties
11 14 15 7
Critical case studies
Key steps:1. Clearly defined hypothesis;2. Find a case, which “perfectly” matches the
conditions set out in the hypothesis;3. Show that despite the “perfect match” the
hypothesis is wrong. 4. Add additional antecedent conditions to the
theoryTwo variations: “the best suited cases”; and “the
worst cases”.
Summary: what is a good academic or policy paper?
• Solves puzzles, which are embedded in academic discussion and are relevant for the “real world people”;
• Raises questions and provides answers (the hypothesis);
• Creates value added on top of the existing knowledge;
• Performs systemic empirical tests to assess the validity of the answers (hypothesis);
• It is readable and understandable!
Further readings:
• Stephen Van Evera, Guide To Methods For Students Of Political Science, Ithaca and London: Cornell University Press, 1997;
• Gary King, Robert O. Koehane ir Sidney Verba, Designing Social Inquiry. Scientific Inference In Qualitative Research, Princeton, New Jersey: Princeton University Press, 1994.