Yale University Economics Dissertation

Description

This is a dissertation about Nduge.

11 attachmentsSlide 1 of 11attachment_1attachment_1attachment_2attachment_2attachment_3attachment_3attachment_4attachment_4attachment_5attachment_5attachment_6attachment_6attachment_7attachment_7attachment_8attachment_8attachment_9attachment_9attachment_10attachment_10attachment_11attachment_11.slider-slide > img { width: 100%; display: block; }
.slider-slide > img:focus { margin: auto; }

Unformatted Attachment Preview

Can nudging techniques be used for the consensus building over redistributive norms?
1.0 INTRODUCTION
Building consensus is a crucial task for policymakers since it may decide whether policies
succeed or fail. Consensus building, or the mediation of a disagreement involving many parties
prior to the adoption of a policy, has been chiefly utilized in the conflicts. Meanwhile, it may be
used to a variety of other controversial public policies at all levels of government.
The proposed research project employs both laboratory and field studies to investigate
alternate ways for gaining agreement on public policy initiatives based on behavioral
economics, including nudging. Nudging is a term that refers to a set of tactics that policymakers
may use to signal or attract attention to a certain decision that may enhance an agent’s
wellbeing without excluding any other options or materially affecting economic incentives.
Theoretically, nudging, like other behavioral economics techniques, assumes that agents are
not rational in the traditional economic sense but rather rationally limited and that; as a result,
they sometimes make biased judgments that are against their own best interests. Nudging is a
kind of libertarian paternalism in this sense since it is a strategy that respects people’s freedom
of choice while allowing both public and private organizations to nudge them in ways that will
benefit them. When such organizations establish arrangements that will persist until individuals
voluntarily choose differently, some kind of paternalism is probable.
2.0 LITERATURE REVIEW
This section aims to demonstrate the benefits of nudging and the techniques that may be
utilized to achieve agreement along with the questions. The following is a breakdown of how
this literature review is organized: Section 2.1 delves into the central idea of libertarian
paternalism and the main criticisms leveled at it; Section 2.2 delves into the behavioral
economics biases and heuristics studied by nudging techniques; Section 2.3 delves into the
1
public good, and bad games, as well as how more significant levels of cooperation and
contribution can be attained.
2.1 Libertarian Paternalism
Nudging is defined by three basic statements, which may be stated as follows:
1. Libertarian paternalism is not a contradiction in terms
2. Choice architecture, or how alternatives are presented to the decision-maker (DM), has a
significant influence on the decision-making process as well as the ultimate conclusion
3. Choice architecture is unavoidable: alternatives must be presented in some fashion, and
doing so in the best interests of the DM makes sense (Thaler and Sunstein, 2008). Theoretical
foundations of nudging can be found in the compound behavioral economics literature on
human fallibility, which includes heuristics, biases, and rules of thumb (Kahneman and Tversky,
2000; Kahneman, Knetsch, and Thaler, 1991), as well as in libertarian paternalism (Thaler and
Sunstein, 2003; Camerer et al., 2003). When a policy is developed to influence personal choices
to improve their status quo in an objective and quantifiable way, it is described as paternalistic
(Thaler and Sunstein, 2003). Generally, libertarian paternalism is an ideology that argues for
preserving individual’s freedom of choice while also allowing private and governmental
institutions to direct them in ways that will benefit them. Without any type of compulsion, the
goal is to eliminate random, arbitrary, or detrimental outcomes of sub-optimal judgments as
much as possible.
2.2 Behavioral economics insights
Grant aids used for fixed investment are regarded as a counterbalance to taxes as part of a
wider industrial policy programs; grant aids are grounded on redistribution, whereas taxes
follow contribution criteria. The related behavioral economics literature addresses concern of
fairness and social preferences from this viewpoint. People’s social preferences are described as
their concern for others’ payoffs and/or behaviors in addition to their own wellbeing.
2
Reciprocity altruism, inequity, and fairness aversion are examples of such preferences, among
the many social inclination’s theories (Bolton and Ockenfels, 2000; Charness and Rabin, 2002;
Fehr and Schmidt, 1999), attracting significant attention, particularly from researchers aiming
to test the models’ forecasts via lab experiments.
When investors’ grant assistance is reduced yearly, for example, they are less helpful in the
industrial policy programs, even if they have provided them with other advantages like
decreased taxes.
2.3 Social preferences, Public good games, and nudging
According to Andreoni’s (1995) study of public goods, social preferences may be motivated
differently depending on whether the supply of the public benefit is portrayed favorably or
negatively. The regular linear public good game was used to elicit this result under two
experimental circumstances: one with positive framing, in which participants were inspired by a
wish to evade the “cold prickle” of guilt, and the other with undesirable framing, in which
subjects were inspired by the desire to avoid a “cold prickle” of guilt. Consequently, participants
in the positive frame situation are more supportive than those in negative frame situation since
the positive frame encourages cooperation.
3.0 RESEARCH QUESTIONS
One of the unexplored potential applications of nudging is how modest pushes might assist
increase the effectiveness of consensus-building procedures when stakeholders have
competing interests. The proposed research project aims to figure out which behavioral
economics discoveries may be utilized to improve the amount of agreement on a redistributive
policy.
First question
3
How can nudging be interpreted in terms of economic techniques?
Theoretically, nudging, like other behavioral economics techniques, assumes that agents are
not rational in the traditional economic sense but rather rationally limited and that; as a result,
they sometimes make biased judgments that are against their own best interests.
Second question
What are the basic statements of nudging?
Nudging is defined by three basic statements, which may be stated as follows:
1. Libertarian paternalism is not a contradiction in terms
2. Choice architecture, or how alternatives are presented to the decision-maker (DM), has a
significant influence on the decision-making process as well as the ultimate conclusion
3. Choice architecture is unavoidable: alternatives must be presented in some fashion, and
doing so in the best interests of the DM makes sense (Thaler and Sunstein, 2008).
Third Question
What are the primary research objectives which follow the proposal dissertation?
The primary research objectives are as follows:
Examine and evaluate several representations of the redistributive norm that must be agreed
upon. Experiments in the lab are carried out in the field to assess the validity and statistical
significance of the lab findings. These study objectives will be developed in the interconnected
investigations, as shown in the following section.
4.0 RESEARCH DESIGN
Nudging strategies will be assessed in the lab and the field. The applied goal of this study is to
suggest to the Department of Trade and Industry of PAT how nudging approaches may be used
4
to teach stakeholders about redistributive norms included in Provincial Law 6/1999’s
applicability requirements. It is vital to remember that experimentation in the laboratory is
used to advise policymakers in areas where a theory has not yet been developed (Offerman,
2020). Stronger inferences about behavioral patterns may be formed in the field by adding the
strict controls developed in the lab. As a result, these predictions allow for more precise policy
guidance.
4.1 Lab test of nudging procedures for building consensus norms over redistributive norms
In the research paper, the hypothesis will be tested in the lab:
When the same redistributive norm is offered to people with particular biases and heuristics,
there is more agreement and support for the policymaker.
And the experiment design is as followed:
The participants in the experiment will be separated into two groups, A or B group, and will be
chosen at random. To illustrate competing interests between the two groups, each member will
be given a non-homogeneous beginning sum of virtual money. And participants must decide
whether and how much of their initial private endowment to contribute to a common pool or
whether and how much to withdraw from a common resource before it is shared among all
participants.
Control Group
Treatment 1 (T1)
Treatment 2 (T2)
Inequity Aversion
Established Practice Effect
Treatment (CG)
Heuristic/Bias
exploited
5
Game played
PG
PG
PG
CGPG
T1PG
T2PG
PG: public good
Treatment Code
T1 takes use of people’s inequality aversion: when it comes to other-related preferences,
people are prepared to pay money or give up a portion of their endowment to prevent uneven
reward distributions. T2 employs an ‘established practice effect’, in which participants are more
likely to follow established practice as transmitted by a competent or authoritative third person
and believe its reasonableness. The following is an example of the multiple formulations of the
same redistributive norm in the approaches that are instructive but not definitive:
Treatment
CG
The public good is shared as follows: 50% to Group A, and
50% to Group B.
T1
Inequity Aversion
T2
Established Practice Effect
The public good is shared as follows: 60% to Group A, and
40% to Group B.
The assignment was set by a famous professor of
economics and is still used today: 60% to Group A, and 40%
to Group B.
There are core commodities, private and public, and N people in a good public game. Each
person i = 1,.., N is given a certain quantity of the private good, zi. Based on a production
6
function Y =f(t), whereby t is the quantity of private good utilized to make Y, the i-th person’s
private good is used to create the public good. An experiment involving public good yields two
results: a quantity of public good Y, then reallocation of private good for every agent x1,.., xN.
Individuals assess the result using a utility function Ui (xi, Y). The public good game should
capture framing effects, which occur when a decision-making process based on similar
representations of a situation leads to consistently different results.
If this finding remains true in the suggested experimental setting, it will show that favorable
incentives for fixed capital investments promote consensus. Subsidies will be framed as
equitable redistributions rather than (mis)appropriation of contributions to a public benefit as
part of the nudging strategy.
Another point that will be evaluated is non-homogeneous beginning endowments since the job
of egalitarianism in a community is to reflect economic inequalities. In his key study, Olson
(1965) shows that greater degrees of inequality lead to larger levels of public good
contributions. Uler (2006) discovered that total contribution to the public good rise when
redistribution is added, and total contribution to the public good decrease when redistribution
is withdrawn.
The hypothesis testing will be structured as follows:
H0: CCGPG=CT1PG=CT2PG
H1: CCGPG?CT1PG?CT2PG1
Despite the meticulous investigation into the experimental design, certain possible effects
cannot be ruled out; hence additional controls will be included. Group effects, dependency
1
‘C’ refers to the amount of private contribution.
7
owing to communication, difference between real and virtual currency and other aspects are all
potential difficulties.
5.0 STRUCTURE
The paper is organized in 5 different sections where in the first section it provides an
introduction about the research question. In the section 2 gives a literature review to
demonstrate the benefits of nudging and the techniques that may be utilized to achieve
agreement. In section 3 it provides the proposed research project aiming to figure out which
behavioral economics discoveries may be utilized to improve the amount of agreement on a
redistributive policy. In section 4 it provides with the research experiment result, and the
section 5 presents with the conclusion of the findings to this study.
8
REFERENCES
Andreoni, J. (1995) Cooperation in Public-Goods Experiments: Kindness or Confusion? The
American economic review, 85(4), pp.891–904.
https://www.jstororg.ezproxy.nottingham.edu.cn/stable/2118238?sid=primo&seq=1#metadat
a_info_tab_contents
Andreoni, J. (1995) Warm-Glow versus Cold-Prickle: The Effects of Positive and Negative
Framing on Cooperation in Experiments. The Quarterly journal of economics, 110(1), pp.1–21.
https://www.jstororg.ezproxy.nottingham.edu.cn/stable/2118508?sid=primo&origin=crossref&
seq=1#metadata_info_tab_contents
Bolton, G.E. & Ockenfels, A. (2000) ERC: A Theory of Equity, Reciprocity, and Competition. The
American economic review, 90(1), pp.166–193
https://www.jstororg.ezproxy.nottingham.edu.cn/stable/117286?seq=1#metadata_info_tab_c
ontents
Charness, G. & Rabin, M. (2002) Understanding Social Preferences with Simple Tests. The
Quarterly journal of economics, 117(3), pp.817–869
https://www.jstororg.ezproxy.nottingham.edu.cn/stable/4132490?seq=1#metadata_info_tab_
contents
Chapman, G.B. and J. Liu (2009) “Numeracy, Frequency, and Bayesian Reasoning Judgment and
decision making, 4(1), pp.34–40.
https://www.proquest.com/docview/1675181887?pq-origsite=primo&accountid=16676
9
Charness, G. et al. (2020) Do measures of risk attitude in the laboratory predict behavior under
risk in and outside of the laboratory? Journal of risk and uncertainty, 60(2), pp.99–123.
https://www.proquest.com/docview/2435334208?OpenUrlRefId=info:xri/sid:primo&accountid
=16676
Humphrey, S.J., Lindsay, L. & Starmer, C. (2017). Consumption experience, choice experience
and the endowment effect
https://nottingham-repository.worktribe.com/output/897002
Kahneman, D. & Tversky, A. (2000). Choices, values, and frames American Psychologist, 39 (4),
341-350.
https://oce-ovid-com.ezproxy.nottingham.edu.cn/article/00000487-198404000-00002/HTML
Levin, I.P. et al. (2002). A New Look at Framing Effects: Distribution of Effect Sizes, Individual
Differences, and Independence of Types of Effects. Organizational behavior and human decision
processes, 88(1), pp.411–429.
https://www.sciencedirectcom.ezproxy.nottingham.edu.cn/science/article/pii/S074959780192
9838
Peters Ellen and Par Bjalkebring (2015), “Multiple Numeric Competencies: When a Number Is
Not Just a Number,” Journal of Personality and Social Psychology, 108 (5), 802–22.
Thaler, R.H. & Sunstein, C.R (2003) Libertarian paternalism is not an oxymoron
Thaler, R.H. & Sunstein, C.R (2003) Libertarian Paternalism. The American economic review,
93(2), pp.175–179.
https://nusearch.nottingham.edu.cn/permalink/f/s8ggqh/TN_cdi_gale_infotracacademiconefil
e_A120393995
Thaler, R.H. & Sunstein, C.R (2008) Nudge: Improving Decisions about Health, Wealth and
Happiness
10
https://nusearch.nottingham.edu.cn/primoexplore/fulldisplay?docid=44NOTCH_ALMA2130927
860005563&context=L&vid=44NOTCH&lang=en_US&search_scope=44NOTCH_COMPLETE&ada
ptor=Local%20Search%20Engine&isFrbr=true&tab=44notch_complete&query=any,contains,Nu
dge:%20Improving%20Decisions%20about%20Health,%20Wealth%20and%20Happiness&sortb
y=rank&facet=frbrgroupid,include,351219499&offset=0
Sugden, R. (2017) Do people really want to be nudged towards healthy lifestyles? International
review of economics, 64(2), pp.113–123.
https://www.proquest.com/docview/1915060322?pq-origsite=primo&accountid=16676
11
Economics Dissertation
Lecture 1: Conceiving your
dissertation
Why write a dissertation?
• It gives you insight into how knowledge is generated in economics
• It gives you a chance to apply economic theories and/or
econometric techniques to real world problems
• It gives you a chance to study a topic of interest in depth
• The transferable skills it assesses may be more useful for your
careers than those required to pass exams
• It gives more scope for creativity and provides a chance for the
best students to excel
• In recognition of the above, external examiners (and Masters
recruiters) tend to give it a high weight in their decisions
• You will produce a serious piece of work that may be a useful part
of your portfolio
• Your knowledge may be a useful talking point in job interviews or
Masters applications
Introductory dissertation lectures –
initial topics
1. Conceiving your dissertation? (today)
2. Planning your dissertation (next week)
3. Conducting your literature review
4. Conducting your data analysis
5. Writing up your dissertation
Lecture Outline
1. Teaching arrangements:
•
•
assessment
supervision
2. Conceiving your dissertation
•
•
Types of dissertation and examples
What makes a good topic?
1. Teaching arrangements
• A dissertation is primarily an exercise in
independent study
• But there will be:
– 5 hours of introductory lectures
– 3 individual supervision meetings
• It is a full year module
– 30 credits = 300 hours of study
– Work load often greater in Semester 2
Dissertation assessment
• 20% of marks will be for:
– Dissertation proposal; 2000 word limit
• Lecture 2 will discuss this in detail
– Includes brief literature review
• Lecture 3 will discuss this in detail
• 80% of marks will be dissertation
– 7500 word limit
• Lecture 5 will discuss format in detail
Module marks 2020/21
Dissertation
Number
First %
2(i)
2(ii)
3
Fail
Mean
SD
Maximum
Minimum
176
58
29
7
3
3
67.9
13.5
89
10
Year 4
average
27
34
21
12
5
60.3
14.5
Permitted fields
Microeconomics
Macroeconomics
Development Economics
Econometrics (Theory or
Applications)
Environmental Economics
Game Theory
Experimental/Behavioural
Economics
Health Economics
Industrial Economics
International Trade
International Macroeconomics
Labour Economics
Monetary & Financial
Economics
Public Economics
What to expect from your supervisor
• The dissertation is primarily an exercise in independent study:
– You are expected to identify a topic yourself
– And to carry out the research alone
• The supervisor will meet with you three times to give you feedback and
guidance on:
– Your dissertation ideas
– Your formal proposal
– A piece of work in progress that you submit
• The supervisor will not:
– Supply you with a dissertation topic
– Read entire drafts of your dissertation
• If your supervisor provides a lot of input, they may “share the marks”
2. Conceiving your dissertation
• Perhaps the hardest part of doing research is
finding a good research question
– You have a whole term to finalise your proposal
• Reading the literature helps map out a research
area & identify a question
• We will look at what is a dissertation, examples of
past dissertations and types of dissertation
– Conclude by considering what makes a good topic
An example: “It’s always harder to
play against ten men”
• Adam Greenberg, dissertation prize winner 2014
– Featured on BBC Radio 4’s More or less
– Research was published in Significance (official magazine
of both the Royal Statistical Society and the American
Statistical Association)
• What he did:
–
–
–
–
–
Start with a hypothesis
Find (or compile) a data-set to test it
Explain his results: do they make sense?
Quantify his results: big effects?
Look at extensions and variations: late red cards; home vs
away etc
What is a dissertation?
• a report on piece of research you have conducted
• an opportunity to describe and comment on
current research
• a chance for an in-depth study on a topic that
interests you
• a voyage of discovery
• something you can include in your portfolio of
work for employers or a Masters
• 30 credits = 300 hours work
– approx. 8 weeks full time study
Categories of Dissertation
In general, economics dissertations can fall
into 4 possible broad categories:
i. Empirical
ii. Theoretical
iii. Literature review
iv. Policy analysis
i. Empirical Study
Empirical dissertations usually adopt a model from an existing paper in
the literature and applying it to a new data set and/or extending it.
• Such a dissertation involves:
– A detailed and critical literature review of your chosen area and
draws on relevant economic theory
– finding and understanding your dataset
– Using appropriate (e.g. econometric) techniques (e.g. Eviews or
Stata)
– implementing your model
– understanding, criticising and checking the robustness of your
results.
– Interpretation and policy implications
Empirical DT examples
Examples of good empirical dissertations from
previous years are:
– Xu, H. (2016): An Estimation and Analysis on
China’s and Zhejiang Province’s International
Bilateral Trade Costs with Selected Trading Partners:
2001 – 2014
– Yiwei, Z (2014): To what extent does urban-torural
migration affect the left-behind elderly in
China
ii. Theoretical Analysis
Theoretical dissertations take model from an existing paper and extending it
in some interesting way.
This type of dissertation involves:
– a brief critical literature review of your chosen area
– acquiring a deep understanding of your model, in the context of the
core material you’ve covered
– Results often take the form of finding the equilibrium of the model
and comparative statics
Examples of good theoretical dissertation from previous years include:
• Ye, M. (2014): Asymmetric information in bank lending
• Wu, L. (2016): R&D Competition and Cooperation under Incomplete
Information: A Simple Model
iii. Literature Review
A literature review dissertation provides a clear outline of the intellectual
development of the a particular area.
Such a dissertation involves:
– a thorough understanding of the literature in your chosen area
– explaining the extent to which different contributors were
addressing similar of different questions and in what sense and
how far one contribution marks a significant improvement over
earlier contributions.
– Critically assessing the different contributions and of the field as a
whole
– A good literature review should add value and demonstrate critical
thinking and technical skills – it is not a simple summary of
existing literature
Literature review DT examples
Browse the Journal of Economic Literature for
examples
• Example: Helpman (2006), Fagerberg (1994)
An example of a good literature review dissertation
from the UK:
• Jones, S. (2011) : An Investigation of the Factors
Affecting House Prices in China
iv. Policy Analysis
• A “policy analysis” dissertation addresses a
specific policy question
– It reviews the literature on the topic
– It considers implications of relevant theory
– It reviews empirical evidence
– It provides policy recommendations: what should
the government do?
• A common (UK campus!) topic is China’s One
Child Policy
Types of UG dissertations
Type
Description
Pros and cons
Empirical
Centres around a piece of
original econometric analysis
?Often original
?Demonstrates technical skills
?Data availability?
?Field may be crowded
(see Lecture 4)
Theoretical
Either modifies a piece of theory ?Can have element of originality
or applies it to a new problem
?Demonstrates technical skills
?Technically demanding?
?May be hard to carve out a niche
Literature
review
Provides a synthesis and critical
analysis of the state of current
research in a given area,
( see Lecture 3)
?“Safe choice”
?But not the “easy choice” as can be
hard to add value & demonstrate
technical skills
Policy
analysis
Brings economic theory and
evidence to bear on a current
issue of economic policy
?Often well focussed question
?Lack of key information?
?Can be hard to demonstrate
technical skills
Examples of First Class dissertations
Some titles from 2010-11
Sub-discipline
The regional greenhouse gas initiative: refining permit
market design or repeating cap and trade errors
The relationship between population growth and India’s
economic growth
Impact of HIV/AIDS morbidity and mortality on household
income and consumption in Malawi
An empirical analysis of the costs of decriminalisation in the
drugs market
Population ageing and the implications for pension reform in
Cyprus
An analysis of the economic causes of divorce
Environmental
Empirical analysis of the contemporary fine art market
Monetary &
Financial
Industrial
The influence of hotel star rating and reputation ratings on
room prices in the London market
Development
Development
Health
Public
Microeconomics
More examples
Title – all these and those on previous slide are available Sub-discipline
to download from Moodle
Are smarter people more procedure invariant? An
investigation into the effects of cognitive ability on
preference reversal
Growth and inequality in India
Experimental
Consumption and growth in Asia and Europe
Macroeconomics
Hydro plant Nam Theun 2 in Laos
Environmental
Development
An investigation of the factors affecting house prices Monetary & Financial
in China
GP incentives
Health
Which consent system should a government
implement? An economic model of behaviour
Public
What makes a good topic?
• Important
– Ideally but note Larry Summers versus Karl Marx
• Interesting
– Ideally current issue & on the frontier of research
– In practice research often chips away at edges of knowledge
• Well defined
– Not too broad or vague: ideally pose a single question
• Feasible
– In terms of data availability and time needed.
• Grounded
– link to an existing module and to scholarly literature
– But note 2013/14 prize winner…
Conclusions
• Start thinking of your dissertation topic now
– Identify preferred field(s)
– Think what type of dissertation you want to write
• Next lecture will cover planning dissertation
? How to find literature
? How to find data
? How your dissertation will be marked
• Any questions?
Economics Dissertation
Lecture 2:
Planning your dissertation
Lecture Outline
1. The Dissertation Proposal
2. Literature Search
3. Where to find data
4. Marking criteria
1. Dissertation Proposal
2000 word limit. It should clearly identify:
(a) the research question(s)
(b) literature: briefly review key scholarly works on topic
(c) data & sources: identify what primary and secondary
material will be used
(d) methods: e.g. perhaps including an econometric
equation to be estimated
(e) structure: proposed section headings of the
dissertation
NOTE: The above is basically a write-up of the
dissertation ideas form you fill in before your
supervision.
Proposal marking criteria
(1) Formulation of research question
Does the dissertation identify an important research area and formulate a well-defined research
question?
(2) Knowledge of literature
Does the proposal reference academic studies that will inform the dissertation, e.g. on theory,
methods, issues etc and explain how the dissertation will relate to them?
(3) Identification of Data Sources
Does the proposal show knowledge of where to search for information for the dissertation e.g.
possible datasets?
(4) Research Strategy (Methods)
Does the proposal identify appropriate research methods to address its research question? How
detailed is the approach – e.g. does a proposed empirical dissertation write down the equation it
will estimate? How advanced are the methods to be used?
(5) Structure & organisation
Does the proposal outline a coherent structure for the dissertation? Are the proposed section
headings a good way to organise the presentation of the research?
(6) Feasibility and prospects
How likely is the proposal to succeed? Is the proposal feasible given the constraints of time,
sources, prior knowledge and training? Does it have the potential to provide original results?
[Originality is not required for UG dissertations, but is highly valued if present.]
2. The literature search
• Start with literature from associated module
• Google
– May pick up hot issues, newspapers, grey literature
• Google scholar: virtue of familiarity
– Can “chase” references: where did this paper lead? How did
other scholars use this paper?
• Bibliographic data bases:
– Econ Lit: restricts to leading economics journals
– Others: Web of Knowledge, JSTOR, ABI/Inform Global, EBSCO,
Science Direct, SCOPUS, Wiley Online Library
• Official sources:
– Reports and publications from World Bank, IMF, UN, countries
– Be wary of bias, e.g. from advocacy
Tips while searching
• Become a virtual stalker: Search key scholars as well as
keywords
• Don’t take keywords or titles too narrowly: study on X
may reveal Y
• Engage in “triage”: Compile a list of “possibles” then
decide whether & in what order to read
• Be both superficial & deep:
– Skim read – e.g. abstracts – until sure is useful
– Key articles may want to study forensically
• Be careful: Be wary of “cut and paste” leading to
inadvertent plagiarism
Plagiarism
• If you use someone else’s words, they must be in
quotation marks and attributed
– But there is seldom a reason to use quotations in
economics (unless pithy, ambiguous or erroneous?)
• Always completely re-write and re-phrase what you
have read
– Minor alterations, even to each sentence, will NOT pass
the software
– If cutting and pasting into notes use different font? Italics
or colour?
– Maybe best to write from scratch rather than edit notes:
will flow better as well as less risk of plagiarism
3. Where to find data?
• Data constrains all empirical work
• Data are often:
–
–
–
–
Not existing or not open access
Incomplete: for years, countries, key controls etc
Subject to measurement errors or inconsistencies
Non-experimental: (instrumental variables can help)
• A good empirical dissertation will discuss data
limitations
– What data would you have liked to have had?
– What would it have allowed you to do?
1st supervision: talk about data
• Supervisors will often have a good grasp of
– What data you would need for your topic
– If such data is unlikely to be available
• Supervisors are often knowledgeable about
particular data sources
– Or can put you in touch with another staff member
who is familiar with such sources
• Supervisors may have their “own” data that you
can use
– Or ideas for new things to do with public data
Data Sources
• When reading the literature, note where the
authors obtained their data.
– If data is “special”, could try e-mailing authors
• On Moodle, various possible data source links
have also been compiled for your easy
reference
• For econometrics, you need the “raw data”
(e.g. excel or stata files).
– Summary tables of means etc. are not sufficient
Macro vs Micro data
• Macro data is typically public
– originally collected by governments
– collated by international organisations (World Bank, IMF etc)
– For cross country analysis, Penn World Tables converts GDP to PPP
https://www.rug.nl/ggdc/productivity/pwt/
https://www.rug.nl/ggdc/docs/pwt91_user_guide_to_data_files.pdf
• Micro data sometimes restricted access
– collected by national governments who may control access
– or collected by researchers who want to use data first
– but journals & funders often require sharing of data
• Size of data
– macro data often small – viewable in Excel
– micro usually large & in multiple STATA files
– managing a micro dataset is non-trivial e.g. need to merge files by
household or individual ID variable
Using STATA
• Most of you will use STATA for your
econometrics
• Will need to import data if data is not in STATA
format yet
• STATA Video Tutorials
https://www.stata.com/links/video-tutorials/
• Getting started in STATA

Collecting your own data
• Types:
– Surveys: Household, school, firm, shoppers, students etc
• Postal, internet, visits etc
– Experiments
– Qualitative data: key informant interviews; focus groups
• Primary data collection
–
–
–
–
Demonstrates particular skills
Provides data that is designed for purpose
Has significant time and often direct costs
Raises many practical issues and requires ethical approval
• May be more appropriate for PG research but is worth
considering
– if only to discuss limitations of data you do use
4. What makes a good dissertation?
Area
Descriptor of first class work
Identification of
research area and
formulation of
research question
Introduction identifies an important research area and
formulates a well defined research question
Strategy for addressing Develops well-defined and imaginative research strateg