Advanced Research Methods:

How to write it down

Markus Lampe
Carlos Santiago-Caballero

1
 The research process
Thinking – reading (and datatyping) – calculating – writing
• Suppose that:
– you have an interesting and original topic
– you have read the relevant literature, chosen your
economic model (or even developed one)
– you have centered and operationalized it to being of
theoretically manageable depth, width and length
– you have found a way to make it measurable to
confront it with reality (operationalization)
– you have found the data you need to do that
– you even happen to know how to “test” your model
– and you have done it
• Now, there is just one small step: convert it into a “story”
“If you can’t think of anything to say, you might well
read more, calculate more, and in general research
more. Most research, however, turns out to be
irrelevant to the paper you finally write, which is
another reason to mix writing with researching. (…)
The guiding question in research (…) is So What?
Answer that question in every sentence, and you will
become a great scholar, or a millionaire; answer it
once or twice in a ten-page paper, and you will write a
good one.
If after all this, though, you still have nothing to say,
then perhaps your mind is poorly stocked with ideas in
general. The solution is straightforward. Educate
yourself.” (McCloskey 2000, p. 29)
 Get started (McCloskey 2000, ch. 7)
• Our subconscious is dismayed by the anxiety of filling up blank
pieces of paper
• It would prefer to check your email, do the dishes, track down one
specific point in the library or internet, see a friend (suggesting you
talk about your thesis), etc.
• The only way to get around it is: sit down, get your notes, get your
text processor started, and write
• Your subconscious might say that – in the light of all the great work
you have read in early stages, you are producing rubbish – have
faith in that a) you will have the possibility to revise later, and b)
there is no alternative to start writing if you want to hand in.
• If we wait for the last moment to write things down we might get
incredibly productive, but also teach ourselves that writing only
works with deadlines, which makes our subconscious associate it
with stress, which does not help overcoming its usual anxieties
• Hence, the keywords are: organization, discipline, and regularity
 “I write when I’m inspired, and I see to it that I’m
inspired at nine o’clock every morning.”
• So inspiration is endogenous Peter de Vries
• And creativity is scarce: Don’t expect to write easily all the time, writing
(and thinking in general) “fizzles like a candle”. If you are inspired, be
selfish and finish the thing you are doing (McCloskey 2000, ch. 10)
• There are techniques to start working:
– start each writing session with a warm-up: review your notes and
organizing ideas for the piece of the day from your last writing
session (but do not get distracted by wanting to track down minor
points)
– Typing in ideas from your notes can help you to get into it. You can
then immediately revise them to make a more coherent text.
– At the end you will also need time to leave your text in an
appropriate shape to go on (not perfect, but understandable!)
– Try to get the material and transition to the next piece ready, that
will make starting much easier next time.
– Make a plan, but make it realistic (if not, nothing is more
discouraging than a series of failed self-imposed deadlines while
the “real deadline” is approaching)
 Writing down – stages
• Read and make notes
• Have a plan
• Write the first draft
• Upgrade / Revise
– Include ideas and material that are missing
– Exclude text that is redundant or a barrier to the train of thought
– Rewrite to assure “flow”
– Unify style, tense, etc.
• so, writing is at least a two-stage process, but in reality it
is cyclical with an upward spiral (revision improve your
work), the number of stages depending on the size of
the project
 Reading notes
• While reading, take notes, make summaries, write down
the most important quotes.
• Don’t forget to write down what was actually the source
of this information, and whether your notes are quotes or
summaries in your own words.
• If you do not do this properly, in the end you might get
confused about the origins of the words in your notes,
and you might actually think that something you read
somewhere was your own idea (and this might lead to
“unintended” plagiarism), or confuse someone else’s
words with your own.
Plagiarism can be the child of bad reading notes
“Many of my colleagues [business professors] work like him [KTzG].
First you assemble a rough framework of a manuscript. The
electronic age makes it all too easy to unite all the relevant in the
literature into a document, using copy and paste. In further
integration loops one can decide what will be used literally, what will
be paraphrased (providing the appriate references, of course), and
what discarded to construct one’s own argument.”

Dirk Matten, “Wider die akademische Vetternwirtschaft”, Spiegel

Online, 27 february 2011,
http://www.spiegel.de/unispiegel/studium/0,1518,747408,00.html

NB: While we recommend to start writing from your reading notes,

we are not convinced that Matten’s “short-cut” approach should be
used in the construction of research. / UC3M uses turnitin!
 Have a plan
• Organization is necessary a) to allocate time and efforts
efficiently and b) to assure that your work has a coherent
internal structure
• After having read “everything on the topic”, don’t start right away
– the result is most an unsorted collection of unprocessed ideas
• Sort your stuff, find out the main arguments, construct your “train
of thought”, find out what is are your main points; sketch a
“roadmap”
• But regard the outline as an aid, not a master (be ready to
change it if your advances require you to change).
• At this stage you can discard all material that does not really
help to answer the question (don’t sent it to the wastebin,
however)
• Establish proportions for your sections
 Familiar roadmaps in economics
• Introduction
• Literature review / Theory (especially if you do
not make a theoretical contribution in this work)
• Data / introduction of your case study
• Methods
• Results and what they mean (and
robustness/sensitivity tests)
• Discussion
• Conclusions
 “Don’t get it right, get it written.”
James Thurber
• The purpose of your first draft is to make text where there is none
• You will discover that there is not “one best way” to do things, but
many ways. Maybe your first choice is not the best, but it is better
to revise later than to search for perfection from the start (you will
get better with practice, and your text will be better after
revisions!)
• So, your draft will be preliminary and have things in the wrong
order and with inadequate weights (you might lose sight of your
main train of thought at times)
• You don’t have to write sections in the right order
• You might discover new implications of your work
– You should start to write fragments out of your reading notes
(don’t forget to write them!) as early as possible
– You will need a roadmap in this process, but one you are
drawing as you drive, no definite outline.
 Revision
• After finishing your first draft, read it again
• Share bits of the text with others, their lack of
understanding might help you to clarify (this is normally
NOT the work of your supervisor)
• Check whether your work actually answers your
research question (and delete everything that does not
contribute to this aim)
• Check whether your argument “flows” – moving parts
where they fit better or removing them if they stand in the
way
• You will also discover possible omissions and
inconsistencies
• Contrary to oral presentations, repetitions/restatements
are often boring in a written text, but sometimes you
need them to make sure the reader does not get lost
 Techniques for revision
• Reading out loud is a good idea for revision (especially at the level of
sentences/paragraphs after correction)
• Imagine the same persons you would show your work to (supervisor,
instructors from your old university, Deirdre McCloskey; what would
they say if you read it to them, regarding structure, content/argument
and style? [on style see below])
• A common recommendation is to make a summary of your own work.
– A general rule says that every paragraph should contain one idea,
so sum it up in one sentence/bullet point
– This way you can find the reason for “macro-discontent” (all
rubbish) at the “micro level” (just sometimes poorly crafted and
incoherent)
– Check whether this is possible (if not, you might have to rewrite
the paragraph)
– Check whether the “text flow” works: Do the arguments follow
from/connect with each other? (In case the don’t,
rewrite/rearrange!)
 How to shorten How to augment
• Sentence level: get rid of • Deepen your understanding
unnecessary words, of important aspects,
repetitions, adjectives, etc. include more bibliography
(but with a positive marginal
• Paragraph level: if a contribution)
paragraph can be unified • Deepen an argument by
with another, reduce it to extending phrases into
one sentence paragraphs (separating
• Get rid of unnecessary ideas)
references • Add sections containing
• Replace descriptions with unused (but useful) material
graphs or tables • Include appendices into the
main text
• Get rid of sections that are • Provide a more detailed
not fundamental to the discussion of your method
understanding of your work and its application to your
specific case
Blaxter et. al. (2000), p. 286)
Revision, major issues (Dunlevy 2003, ch. 6)
• This is more intended for PhD theses with several chapters of considerable
length (30-40 pages), but might help you if you believe that your
macrostructure and roadmap make your thesis “unreadable”)
• Ask yourself:
– Is the chapter structure simple (good) or complex (bad)?
– Is the argument pattern clear and logical (good) or unclear (bad)?
– Do the current sections and subsections divide up the chapter text
evenly (good) or unevenly (bad)?
– Does the chapter’s argument have a developmental or cumulative feel
about it (good) or does it by contrast seem recursive and repetitive
(bad)? (the latter means that you have arguments and evidence
scattered all over the text, refocus by concentrating arguments on one
issue in one place)
– Is the underlying mode of exposition argumentative/analytical (easier to
organize and make coherent regarding the author’s POV) or descriptive
(generally worse for organizing and personalizing your argument).
– A mismatch between your initial plan (roadmap) and its implementation
as you read it does not necessarily mean that your text must be
changed to meet the plan; perhaps the plan needs an adjustment, too?
 For whom do I write?
• First of all, for your supervisor, but also for the committee
• All of them might be regarded experts on some field in economics
(development, growth, economic history, econometrics, etc.), but not
necessarily on your topic
• So your “representative agent” should be seen as an informed
economist, but as somebody who knows as much as you about every
aspect of your topic
• This means that you should explain what you are doing and the key
concepts you are using, but if an economics undergraduate should
know them (textbooks) don’t be too exhaustive (don’t repeat an
extensive proof of the Gauss-Markov theorem), but meet the
expectations of your supervisor (justify why you use a specific time-
series model or your economic model)
• In general, economists believe that it is a virtue to be clear and
understandable than to hide behind boilerplate and oversophistication
(see below; but some also hide behind mathematics)
 Style (mostly McCloskey 2000)
• Get a copy of Deirdre N. McCloskey, Economical Writing, Prospect Hights 2000
• Writing is thinking: If you do not make clear what you have to say, readers might
suspect that you do not have it clear (but if you explain things like for idiots,
readers might suspect you take them for idiots)
• It’s a good rule not to write anything you would be embarrassed to say to the
intended audience. So, use your ear (what sounds like flow, flows even when
read in silence)
• Avoid boilerplate, that is, excessive introduction and summaries (like the “index
paragraph” in many articles’ introduction), useless all-purpose fillers (“This
paper…”, instead use a “hook”), and “background” (material you have worked
hard to collect, but that is not really necessary for your story, but documents your
efforts, like derivations of estimators or microeconomic models found in
textbooks – it might also obscure your train of thought), and things like “as I said
earlier”, etc.
• Avoid extensive quotations, since other people’s voice might disturb the flow of
your argument, unless you have a good justification for including it
• Make your writing cohere: write “transitively”, that is, connect the sentences
grammatically (so, the reader always knows who is the subject, etc., and does
not have to “make” sense scanning backwards). Do that effortlessly, i.e., avoid
“boilerplate” connections and excessive linking words, when the connection is
already clear.
 Style (II, still McCloskey 2000)
• Avoid elegant variation, that is, do not use different words that supposedly
mean the same thing, but might confuse the reader (a reason for doing this
might be the fear that repeating the same word/term might be considered
boring/bad prose)
• Use active verbs (not “active verbs should be used”), talk to the reader (“then
divide both sides by X”), avoid “there is” and excessive nominalization.
• If you write alone (as you do in the thesis), use “I” when you document what
you are doing, and “we” if you include the reader (because you document the
progress in understanding that leads to)
• Be concrete: name the things you are talking about as specific as possible, so
people can “imagine” them, use singular, not plural when possible)
• Be plain: write things as understandable as possible, oversophistication
obscures understanding (and hence makes your thinking difficult to follow),
and generally does not make people believe you are cleverer than they are
(see point 1)
• Umberto Eco says: Don’t be Proust (10-page sentences), don’t be e.e.
cummings (experimental language), don’t use metaphors that are not
immediately clear to the implied reader, and therefore don’t explain
metaphors (or jokes, since the actual reader might think you take him as an
idiot)
• Break any of these rules if necessary (but only then)
Other easy mistakes (Eco 1990, ch. V.5/6)
• Don’t be too modest. If you have contributed an original thesis
you should not write that you are “not qualified to provide
conclusions”, etc. Be bold (but not presumptuous)!
• Do not provide references and sources for things everybody
knows
• Do not attribute ideas to an author who is presenting them as
those of someone else
• Do not confuse Anglo-American numbers (“.” for decimals and
“,” for thousands) with other systems
• Do not succumb to “physics envy”, that is fetishism for precise
numbers (1,215,689 unemployed persons), economic reality
is not measured with such precision (Dunleavy 2003).
• Avoid false friends in translating (“ilustración” translates as
“enlightenment”, not illustration)
 Handling graphs (“attention points”)
• Your arguments should be developed in the text, but
especially for economists tables are often important attention
points, and in empirical works readers often first jump from
table to table before reading the text
• An illustration, graph, table that the text does not refer to
should be omitted, the same is true for one that adds very
little to the information provided in the text
• A bad graph or a table that is not understandable can cause a
very negative first impression (Dunleavy: “how you handle
attention points will strongly influence readers’ views of the
professionalism of your approach”)
• There are some easy ways to make graphs and tables
understandable, reducing data and communicating it
effectively, instead of throwing an unprocessed mass of
information at readers (Dunleavy)
 The ‘need to know criterion’
(Dunleavy 2003, ch. 7; Miller 2005, ch. 2, 5, 6)
• Key: What will my readers need to see or need to know in order to
accept the conclusions of my analysis?
• “A table is a tool for presenting numeric evidence, not a database
for storing data or a spreadsheet for doing calculations. (…) Like
a carpenter, do the messy work (data collection and calculations)
correctly, then present a clean, polished final product. ” (Miller
2005, 111)
• Make one set of easily accessible, but summary tables in the text
(for the “general reader”) and provide more detailed information
and the complete set of results for alternative specifications, etc.,
in appendices (for “experts”)
• For details, Dunleavy has good instructions about when to use
which kind of chart; Miller has entire chapters on this (important
advice: 8 lines are the absolute maximum for a chart, more then 4
are already “many” for our brain)
Bad table
Better
table
 Further suggestions
• Miller warns us not to confuse statistical significance with
significant importance (or “substantive significance”): always
provide calculations about what a typical (e.g. 1 standard
deviation) change in an independent variable means for the
dependent variable (“how much does it matter?”)
• Also, do not confound causality and correlation. Since the 1960s,
some criteria for causality have been established (apart from
statistical tests for causality):
– consistency of association (in several studies with different
populations), strength of association, temporal relationship
(cause precedes effect), mechanism (so there is a consistent
theory which predicts that A systematically causes B)
– Hence, before-after observations are better than cross-
sectional yes/no data. And a well-specified model with control
variables is better than partial correlations.
– For writing: “Verbs such as ‘affect’ or ‘cause’ and nouns such
as ‘consequences’ or ‘effects’ all imply causality. ‘Correlated’
or ‘associated’ do not.” (Miller 2005, p. 39)
 Avoid plagiarism, put references
• Make sure you include references to works of others you are referring
to, whether in their own or in your words
• Mark direct use of other’s words with quotation marks (however, you
should avoid abuse of direct quotation and make sure the ideas are
embedded into the flow of your writing).
• Undisputed facts and common knowledge do not need to be referenced
• If you find an idea/argument that you have formed yourself in another
text, and you are proud of having had this idea, you can state that it was
yours, but that someone also also had it (using “cf.” (“compare to”))
• If you are in doubt about whether to put a reference or not, a good way
of “self-education” is to learn from the authors you read (Eco 1990, ch.
V.3 and V.4 has a list of ten rules about references and additional
information about unintended plagiarism and the use of footnotes)
• Necessary, but not sufficient: Include a bibliography of all cited works at
the end of your work
 Sources
• McCloskey, Deirdre N.: Economical Writing (Prospect
Heights, Illinois: Waveland, 2000). Older version here:
http://www.deirdremccloskey.com/docs/pdf/Article_86.pdf
• Blaxter, Lorraine, Hughes, Christina, Tight, Malcolm: Cómo
se hace una investigación (Barcelona: Gedisa, 2000)
• Dunleavy, Patrick: Authoring a PhD. How to Plan, Draft,
Write and Finish a Doctoral Thesis or Dissertation
(Basingstoke, New York 2003)
• Eco, Umberto: Cómo se hace una tesis : técnicas y
procedimientos de estudio, investigación y escritura
(Barcelona: Gedisa, 1990)
• Miller, Jane E.: The Chicago Guide to Writing about
Multivariate Analysis (Chicago: University of Chicago
Press, 2005)