M2R “Development Economics” Research Methodology - Panth´eon Sorbonne University

Research Methodology 1: Defining a good research project Remi BAZILLIER [email protected]

Semester 1, Academic year 2017-2018

1 / 69

About the Master 2 thesis workshop

I

The Master 2 thesis is due on May 17, 2018 and all the defenses must be organized before May 25, 2018 (the thesis is defended in front of a jury that includes the supervisor and at least one other examiner).

I

Candidates to a doctoral fellowship must then submit their application before the end of June.

I

The Master 2 thesis workshop is therefore critical to allow you to meet these demanding deadlines.

2 / 69

About the Master 2 thesis workshop I

You are required to attend all the sessions of the workshop.

I

You will obtain a grade for both attendance at the sessions and participation/presentations.

I

These grades will also take into account attendance at the Paris Sorbonne Sustainable Development Research Seminars, since this will constitute a direct source of ideas for your own research. http://remi.bazillier.free.fr/PSSD_web.htm

I

Absence from more than 1/4 of the seminars will result in point deduction from the Master thesis seminar grade.

I

Webpage of the M2 thesis workshop: http://remi.bazillier.free.fr/masterthesis.htm (with all deadlines and calendar) 3 / 69

About the Master 2 thesis workshop

I

The two “Research methodology” classes are part of round 1 of the Master 2 thesis workshop (“Getting started”): I

September 14, 2017 (today): Research methodology 1: “Defining a good research project”

I

September 27, 2016: Research methodology 2: “Writing a good research paper”

I

October 12, 2016: Presentation of the potential supervisors and their research topics

I

October 26, 2016: Discussion with former students

4 / 69

About the Master 2 thesis workshop I

Round 2: “Choosing a research topic” I

November 9, 2017 and November 23, 2017: Brainstorming sessions (students present their research interest in front of the professors)

I

Be careful, only students who will have sent a one-page description of their research interest to Remi Bazillier by October 25, 2017 will be given the opportunity to elaborate on it during the brainstorming sessions. Administrative deadline: November 29, 2017, document with stated topic of the thesis and name of supervisor should be handed in.

I

I

January 11, 2018 - January 18, 2018 - January 25, 2018 February 1, 2018 : Presentation by the students of their benchmark paper. 5 / 69

About the Master 2 thesis workshop

I

I

Round 3: “Writing the Master thesis” I

February 3, 2018 (midnight): You are requested to send a short research proposal (no more than 5 pages) to Remi Bazillier

I

Later on: Three sessions of presentation by the students of their first results (March 8-15-22-29, 2018)

Students will then pursue the writing of their Master 2 thesis with the help of their supervisor.

6 / 69

About the Master 2 thesis workshop

I

Research methodology 1 (“Defining a good research project”) will be helpful to tackle round 2 as well as the writing of the research proposal that is due on February 3, 2018 (midnight).

I

Research methodology 2 (“Writing a good research paper”) will be useful for the writing of your Master 2 thesis.

7 / 69

About the Ph-D track

I

The goal: additional courses to prepare students to candidate for a Ph-D (common to all M2 Research)

I

Only a limited number of positions

I

You can apply for a Ph-D even if you did not follow the Ph-D track

8 / 69

About the Ph-D track I

First semester: I I

Microeconomic theory (15h) or Macroeconomics (15h) One course (9h) to be chosen in the following list: I I I I

I

Law and Economics Financial Economics International Economics Sustainable Development Economics

Second Semester: I I

I

Econometrics (15h) or experimental economics (15h) Behavioral Economics (9h) or Quantitative evaluation of public policies (9h) or Epistemology and History of economic thought (9h) Drafting a Ph-D proposal (5h)

9 / 69

About the Ph-D track - Selection Process

I

Send to Remi Bazillier a research proposal by October 2, 2016 (midnight) I I

Ph-D project (5 pages maximum) Your CV

I

A pre-selection will be made by the director of the Master

I

Second step: a committee with all Masters’ directors

I

Final decision made by the Dean of the Economics Department (UFR)

10 / 69

Introduction

I

What is a good research project?

11 / 69

Introduction

I

A good research project must be based on an interesting research question (not only for others but also for you!; not only for you, but also for others!)...

I

... that must lead to a new and feasible contribution.

12 / 69

Introduction

I

I

An interesting research question for both you and others implies that this is: I

a question you feel passion for;

I

a question whose answer is not obvious;

I

a question whose answer is consequential.

What is a “consequential” answer?

13 / 69

Introduction I

A “consequential” answer is an answer which: I

I

should help you and the reader better understand some microeconomic or macroeconomic mechanisms that matter for individuals’ well-being: I

parents’ decision to send their children to school in developing countries;

I

the distributional effects of trade liberalization... etc.

should help you devise policy recommendations that aim at improving individuals’ well-being; for instance, identifying ways to improve: I

parents’ incentives to send their children to school;

I

the situation of those who lose from trade liberalization... etc. 14 / 69

Introduction

I

Providing a new contribution on an interesting research question I

will not only increase your chance of getting a good grade for your Master 2 thesis,

I

but it will also increase your chance of getting published in a good journal (one that is ranked by CNRS) if you decide to spend time (at least one year in the context of a PhD) on improving your Master 2 thesis.

15 / 69

Introduction I

Economics journals are ranked according to their impact factor. For instance, the 2013 impact factor of journal x is computed as: A/B, where A is the number of times articles published in journal x in 2011 and 2012 were cited by a set of journals in economics during 2013, and B is the total number of citable items (articles, reviews, proceedings or notes) published by that journal in 2011 and 2012.

I

For instance, if journal x has an impact factor of 3 in 2013, this means that the papers it published in 2011 and 2012 received 3 citations each on average in 2013.

16 / 69

Introduction

I

Therefore, in order to improve their ranking or to keep their top rank, economics journal are eager to publish articles that show a high citation potential (i.e. that are interesting and whose contribution is really new).

I

As a result, editors are often more concerned with the risk of accepting low-quality articles than with the possibility of rejecting good articles.

I

Their risk-aversion makes them extremely cautious!

17 / 69

Introduction I

Only the unambiguously good articles are published (i.e. those which are supported by all the referees the editor sends the article to, or at least by a large majority of these referees).

I

A “large” majority means 2 out of 3, or 3 out of 4 referees... when the paper is sent to only 2 referees, both have to be supportive for the paper to be accepted for publication by the editor.

I

Publishing in good economics journal is therefore tough.

I

Invest a lot in your Master 2 thesis if you wish this thesis to become a working paper that you will be able to publish “well” in the coming years. 18 / 69

Introduction

I

In the following, we focus on the case of an empirical research project, that is a work that aims at: I

estimating the magnitude of an impact (e.g.: the impact of economic growth on civil war onset),

I

or testing a model (e.g.: the Harris-Todaro model that states that migration decision is based on expected income differentials between rural and urban areas),

based on empirical data.

19 / 69

Introduction I

I

Writing a theoretical rather than an empirical Master 2 thesis is very risky: I

The whole point of a model is to give a simplified representation of reality.

I

This is a very difficult art to practice.

I

Learning how to master this art is a long process (too long given the deadlines you are facing, unless you are particularly skilled – and supervised by an excellent theorist).

For those who are not discouraged, check Varian (2009): “How to build an economic model in your spare time.”

20 / 69

Introduction

I

Throughout the process of defining your research project, implementing your empirical strategy, and writing your Master 2 thesis, you must talk continuously with people about what you are doing: I

talk to economists: this master thesis workshop, other students, professors...

I

talk to non economists (your family, non economist friends): at each stage of the process, make sure that you can explain clearly to them what you are doing and why this is both interesting and new.

21 / 69

Introduction

I

Be a compulsive note-taker from now on: I

Right after this class, create a folder on your computer.

I

This folder should get a running stream of thoughts (yours) and suggestions (those coming from people you talk to): constantly add to it.

22 / 69

Introduction I

I

This folder will be extremely useful to tackle each step of the process consisting in: I

defining your research project (what is interesting in my research question? how does it contribute to the literature on the topic?... etc);

I

implementing your empirical strategy (what are the empirical concerns raised? what can I do to solve them?... etc).

It will considerably ease the writing of your research proposal (due on February 3, 2017) and more generally of your Master 2 thesis.

23 / 69

Introduction

I

Outline of the class: 1. Choosing an interesting research question 2. Defining a new and feasible contribution

24 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

Curiosity is a fundamental ingredient for research to begin.

I

Seek for sources of ideas everywhere!

I

Read the newspaper, listen to the radio, watch TV with an eye towards economics-related questions.

I

Watch TED (Technology, Entertainment, Design) conferences: http://www.ted.com/.

25 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas I

Read blogs that disseminate the results of new research in economics/social sciences, like: I

http://www.voxeu.org; http://rodrik.typepad.com/

I

http://chrisblattman.com/;

I

http://www.slate.fr/ and http://www.slateafrique.com/;

I

Twitter

I

Regularly visit the webpage of general-interest journals in economics (AER, QJE, ...etc) and of journals in your field (JDE, WBER, EDCC, WD, JDS, ...etc).

I

See rankings of journals on the master thesis website. 26 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

Subscribe to the World Bank newsletter (http://newsletters.worldbank.org/newsletters/list.htm) and to the OECD Development Centre newsletter (http://www.oecd.org/document/52 /0,3343,en_2649_33731_36562164_1_1_1_1,00.html).

I

Think about your classes: what are the unanswered questions?

I

Think about ways to develop your Master 1 thesis.

I

React to the list of topics provided by the professors teaching in the M2R “Development Economics”.

27 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

Go to seminars in the field(s) that is (are) of interest to you: the speakers are leading scholars who usually present their research when they are at the working paper stage: more questions are raised than answered, which offers room to brainstorm.

I

Regularly visit the webpages of researchers whose papers you enjoyed reading during your previous years of study: this will keep you informed about what is new in the fields you care about;

I

... etc

28 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

In your folder, create a file “journal of ideas”: whenever you have a research idea, write it down.

I

You must feel “passion” for these ideas.

I

It is good if these ideas are reactions to statements by economists, politicians, or journalists that puzzled you or made you angry.

29 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

For instance, Alexia Lochmann (who is now enrolled in the PhD program at Paris 1) dedicated her Master 2 thesis to the following research question: “What are the main causes of domestic violence?”

I

The hypothesis she wanted to test was whether income shocks may outweigh cultural beliefs in their weight on domestic violence. She uses Nepalese household data.

30 / 69

1. Choosing an interesting research question 1.1. Finding sources of ideas

I

Her research question was motivated by an anecdote reported by the third Millennium Development goal by the United Nations: the promotion of gender equality and the empowerment of women (UN, 2015): “Women’s lack of empowerment is widely recognized to be impeding development and reducing the well-being in the society. But what the main drivers of domestic violence (physical, emotional and sexual violence)? What about reverse causality?”

31 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

As already mentioned, the research question has first to be of genuine interest to you since you will have to dedicate lots of time and intellectual energy to writing your Master 2 thesis: working on this research question should give you fun!

I

Moreover, bear in mind that, for those of you who want to enroll in a PhD program, the Master 2 thesis usually becomes the first paper of your PhD dissertation (typically composed of 3 or 4 papers).

I

The research question should therefore be such that you will be willing to invest in it for at least a couple of years.

32 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting I

The research question also has to be interesting to others.

I

First, the answer to your research question should not be obvious: if you analyze the causal impact of A on B, it should be possible: I

that this causal impact doesn’t exist,

I

or, when there is little doubt that it exists, that this impact is either negative or positive.

In other words, if you happen to describe some of your empirical results as “good” and other results as “bad” you are mistaken: you shouldn’t find yourself hoping for one specific result. 33 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Second, the answer to your research question should be consequential (i.e. related to a real world problem).

I

Even non-experts should feel that addressing this question will lead to conclusions that matter for the world’s well-being.

34 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

For each of the research questions contained in the file “journal of ideas” of your folder, assume that the most skeptical persons among the economists and the non-economists you know, press you with the following questions: I

“Why isn’t the answer obvious?”

I

“Why should I care?”

35 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

For each of your research questions, elaborate a detailed answer to these two skeptical questions.

I

This will be very useful for the writing of: I

your research proposal;

I

the introduction of your Master 2 thesis.

36 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

About why the answer is not obvious: when you analyze the impact of A on B: I

you should identify the endogeneity problems that could lead to an absence of causal relationship between A and B;

I

or, when there is little doubt that this relationship exists, you should identify the theories that could justify a negative or a positive relationship.

37 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

About why you should care: I

For each possible answers to your research question, ask yourself: “Why are those answers consequential?”/“Which policy recommendations could they lead to?”.

I

Note that this process could help you fine tune your research question if you also ask yourself: “What would make my contribution more consequential?”.

38 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Going to weekly/biweekly/monthly seminars in the fields that are of interest to you won’t only work as a source of ideas.

I

It will also train you to answer the two skeptical questions when listening to others’ research.

I

Whenever you attend such seminars, ask yourself: “Is the speaker dealing with a question that is interesting?”

39 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Spend time on selecting the research question(s) within your journal of ideas that is (are) the most interesting: only keep the research question(s) for which you are able to convince the skeptic (and yourself!) that he/she/you should pay attention to your research.

I

As emphasized by Larry Summers, a former President of Harvard University: “It takes as much time to answer a minor question as an important one” (this is known as the Summer’s Law among economists).

I

So don’t waste your time. Choose the important ones!

40 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Let’s go back to Alexia Lochmann’s Master 2 thesis.

I

Why is her research question interesting? “Are the forms of domestic violence (physical, emotional and sexual violence) affected by changes in income?”

I

Think of the policy implications

41 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting I

First, this is a research question that is topical (Millenium goals)

I

Second but the answer to this research question is not obvious.

I

To be sure, many people (even those working in development) would take it for granted that attitude toward women is pre-determined by culture

I

and may downplay the role of the economic factors. They would realize after some thoughts that the link between income and domestic violence is two way...

I

... also that there are many conditioning factors (religion, alcohol consumption..) 42 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Put differently, the only way to provide an answer to Alexia Lochmann’s research question was to test it, by relying on empirical data.

I

This is exactly what we mean by “a research question whose answer is not obvious”.

43 / 69

1. Choosing an interesting research question 1.2. Keeping only the research questions which are interesting

I

Third, the answer to her research question is consequential.

I

There is a necessity to increase women’s empowerment.

I

However, how? if culture/religion rules, little hope

I

if policy targets a transmission channel (ban alcohol use) may somewhat ineffective

I

if income shocks matter then policies relating to insurance schemes appear straightforward

44 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Once you have identified one (or a set of) research question(s) that is (are) the most interesting, you have to define a contribution that is new.

I

This does not mean you have to revolutionize the field in economics you work on.

I

But it does mean that you have to extend the understanding of some microeconomic or macroeconomic mechanisms in a way that is conducive to more fine-tuned policy recommendations.

45 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

After reading your paper, researchers familiar with the literature related to you research question should see the world differently because you will have managed to: I

identify a clear causal (and maybe unsuspected) relationship between two variables;

I

cast doubt on some views of the world that people might otherwise have maintained (i.e.: by showing that data do not support the yet well established theory x, ...etc).

46 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

We obviously expect more from you than just take someone else’s regression model and analyze the impact of adding another control variable.

I

This is okay for an econometrics exercise but not for a Master 2 thesis, unless you manage to convince us that there is a strong theoretical motivation for the inclusion of this new variable.

I

Put differently, when we refer to a “new” contribution, we also refer to a contribution that is substantial.

47 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

To define a new contribution, the first step is to use the key words related to your research question and to search for papers which are closely related.

I

Of course, this requires that you focus on already published papers, through the electronic resources provided by the Centre d’Economie de la Sorbonne.

I

You will access these resources by clicking on the following link: http://cesdoc.univ-paris1.fr/ressources-electroniques/.

I

Presentation by France MARTIN of all electronic ressources (this afternoon)

48 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Note that: I

the Annual Review of Economics;

I

the Journal of Economic Literature;

I

the Journal of Economic Perspectives;

I

the Journal of Economic Surveys;

are particularly suitable for obtaining a broad, non-technical overview of a topic.

49 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

But you also have to be aware of the research frontier, which means of where the research in a particular area has been going in the last 12 or even 6 months.

I

If you focus too much on already published papers, you won’t find this frontier because material in the journals is inevitably dated.

50 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Indeed, an empirical project typically involves: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

conceptualizing the problem; applying to a grant and waiting for the approval; gathering and cleaning data; running your regressions; getting a set of results you find convincing; writing a paper; issuing it as a working paper; sending it to journals; getting rejections; doing revisions; submitting a final draft; and waiting for it to be finally published.

51 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new... I

The papers that are published today may therefore represent the state of thinking that prevailed many years ago.

I

So complement your literature review by a search on

I

I

Google Scholar: http://scholar.google.fr/;

I

IDEAS - REPEC: https://ideas.repec.org/;

I

NBER (through the CES databases)... etc

Also go to seminars in the fields related to your research question: researchers who present there are likely to be very much at the frontier. 52 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Identify the intellectual leaders in your research area and check out their webpage too.

I

In each research area, there is indeed a number of people who consistently push forward the frontiers of the research.

I

You have to become familiar with this community in order to answer the following questions: I

What are the problems they are struggling with currently?

I

How do they proceed?...etc

53 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Keep in mind that, if the most recent papers related to your research question are 5 years old, this is likely to be a dying issue.

I

Think twice before going on with this research question (ask yourself: “Why hasn’t this research question received more attention recently?”).

54 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

When reading the papers which seem to be related to your research question, proceed as follows. I

Read the title and abstract and then decide whether or not to delve deeper.

I

If you do decide to delve deeper, read the introduction: this section should allow you to get the data used, the empirical strategy followed, and the results obtained.

I

Save the paper in your folder in case it appears to be closely related to your research question.

I

Repeat this process.

55 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

Once you’ve identified the papers that are closely linked to your research question, read them again...

I

... but this time seriously and thoroughly, meaning that you have to check all the details and question everything.

I

Do not hold previous research in reverence... bear in mind that for you to define a contribution that is new, you have to admit that others’ papers are never perfect (and therefore to track down their flaws!).

56 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new... I

When you read these papers carefully, ask the following questions: I

Can the empirical strategy be applied to other countries or time periods?

I

How could this empirical strategy be improved? (i.e. could you find better instruments?; could you rely on more compelling impact evaluation techniques?...etc)

I

What else could be done with the dataset (look at the codebook to make ideas emerge)?

I

Are there any implications of the study which have not been fully exploited by the author and that you could develop?

I

... etc 57 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

This careful read will help you define contributions that are new.

I

In your folder, for each of the research questions you will have saved because they are the most interesting, write the list of the new contributions you could implement and explain clearly why they are new (i.e. how they improve upon the literature).

58 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

For instance, Alexia Lochmann wanted to complement the literature on the role of income stability for gender equality in primary education (Bjrkman-Nyqvist (JDE, 2013).

I

and look at the role on domestic violence. Studies on this topic are scarce because of the lack of data

I

Nepal is a notable exception since new survey following the 2009 “Domestic Violence Crime and Punishment Act”

I

Alexia soon realized that there was room for providing two improvements to the literature.

59 / 69

2. Defining a new and feasible contribution 2.1. Defining a contribution that is new...

I

The contribution of Alexia’s Master 2 thesis was therefore threefold: 1. Extract the causal link between income shocks and domestic violence (use of an IV); 2. Compare its contribution with that of more structural characteristics (culture, ethnicity, religion..) 3. Explore channels of effects (alcohol use)

60 / 69

2. Defining a new and feasible contribution 2.2. ... and feasible I

Feasibility involves data availability.

I

In the folder where you’ve listed the new contributions you could implement for each of your interesting research questions, highlight the contributions that are feasible (i.e. those for which data are available).

I

It is worth investing in finding out the optimal data.

I

Such data are those which allow you to best address your research question: you will be able to answer this question while implementing a neat identification strategy (one that solves endogeneity problems). 61 / 69

2. Defining a new and feasible contribution 2.2. ... and feasible

I

All in all, feasibility means availability of data that allow you to implement an identification strategy...

I

...that provides an answer to your research question that is purged of any endogeneity bias.

62 / 69

2. Defining a new and feasible contribution 2.2. ... and feasible

I

Note that you will have the opportunity to collect your own data if you do a PhD...

I

... provided the research question that motivates this data collection is considered interesting enough by institutions which typically fund such field work!

63 / 69

2. Defining a new and feasible contribution 2.2. ... and feasible

I

I

In your folder, select the research question: I

that is the most interesting of all;

I

that leads, meanwhile, to the newest and most feasible contribution.

This research question will be the one of your Master 2 thesis.

64 / 69

Conclusion I

At the end of this process of choosing an interesting research question and defining a new and feasible contribution, your folder will contain the research proposal that is due on February 3, 2018.

I

Notably, it will contain the answer to the following questions: I

What is your research question?

I

Why isn’t the answer obvious?

I

Why is the answer consequential?

I

To what extent does your contribution improve upon the literature on the topic?

I

Which identification strategy and dataset will you rely on? 65 / 69

Conclusion

I

The answer to each of these questions should be written with the greatest clarity.

I

Start working on your folder now because research takes time.

I

Research will take you even more time if you are not focussed.

66 / 69

Conclusion

I

“Research involves intense concentration over long periods. It is not possible to return to a research project casually at irregular intervals or just when there are no other pressing commitments. It is necessary to allocate regular times to research and always to keep a project moving forward. Indeed, concentration has to be such that [your research project] becomes something that is extremely hard to stop thinking about.” (Creedy, “Starting research‘”, 2001, Australian Economic Review).

67 / 69

Conclusion

I

Moreover, be prepared to throw stuff away. We want you to come up with one interesting research question that will lead to one new and feasible contribution.

I

This means that there will be a lot of things in your folder you won’t use in the end product, but that will have been critical in making you converge to the clear and clean research proposal you will send us.

68 / 69

Conclusion I

Finally, talk to people about your project...

I

Don’t wait until you think you’re done to get feedbacks. This will result in a considerable waste of time because people will inevitably bring up several important problems you hadn’t thought of in the process of writing your research proposal.

I

Talk to your advisor too: “the risks of saying nothing far outweigh the costs of occasionally saying something stupid (so long as you also occasionally say something interesting!)” (Davis, “PhD thesis research: where do I start?”, undated, Columbia University).

I

This means that you have to choose your supervisor ASAP!!! 69 / 69