Ingar Haaland
@Ingar30
Followers
3,214
Following
1,181
Media
96
Statuses
1,953
Associate Professor at the Department of Economics @NHHEcon
Bergen, Norway
Joined May 2016
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Telegram
• 1688465 Tweets
WE LOVE YOU YOONGI
• 660099 Tweets
ORM YOUNG OTOP 2024
• 353340 Tweets
WE TRUST YOU YOONGI
• 288895 Tweets
SB19 RED CARPET PREMIERE
• 243498 Tweets
#VALORANTChampions
• 151205 Tweets
1Z BALIBAGAN SA RED CARPET
• 138764 Tweets
ARMYS PRIDE YOONGI
• 95383 Tweets
MILE x McLarenArtura
• 89315 Tweets
#光る君へ
• 87424 Tweets
#DutchGP
• 74246 Tweets
Good Sunday
• 62311 Tweets
ST☆RISH
• 61806 Tweets
Wolves
• 52473 Tweets
Norris
• 45869 Tweets
最高の夏
• 25670 Tweets
#ブラックペアン
• 24312 Tweets
Ruさん
• 18201 Tweets
Lavia
• 12007 Tweets
#マイクラ肝試し2024
• 10898 Tweets
Pinned Tweet
How should one design survey experiments for maximal impact in economics? Here's my slide deck from a recent workshop in Uppsala. Thanks to everyone attending for making it a great experience
@EconomicsUU
3
39
165
Ph.D. students in economics: How to organize your projects and not make the replication files a huge pain - a thread with a minimal working example with some potential extensions.
13
232
1K
1/ People often say that null results are penalized in the publication process. Is that true? In a new paper with
@cp_roth
,
@FelixChopra
, and Andreas Stegmann, we examine whether and why that's the case! Read on to learn more!
19
276
1K
"fre" is my new favorite Stata package: it simultaneously displays the value and the value label. No more "tab x" followed by "tab x, nolab"
35
180
949
GPT-4 is *eating* all other software on the planet.
It can now with ease convert quite complex tables into LaTeX.
12
42
420
Some thoughts on Econ PhD/early career progression
1. The system is rigged towards spending too much time on JMP. You kind of have to embrace it, but only if you have an extremely good paper
1
48
373
1/ New paper alert: People often say they want the most accurate news possible, but in practice, they often consume ideologically aligned news. Why?
8
92
355
1/ More than 70% of the academic researchers who answered my poll "rarely or never ChatGPT". Perhaps some of you are unaware of how good it is. I use it to assist me with all sorts of tasks, including programming assistance and proofreading. See some actual examples below!
3
43
260
I'm introducing Proofreader Pal, a GPT designed to identify grammatical errors and awkward phrasing. It goes straight to the point and formats the output in structured lists. Try it now:
6
31
253
Say goodbye to Beamer slide headaches! With ChatGPT, you can easily create complex bullet points and sub-bullets in your presentations. Plus, it might even catch your spelling mistakes!
#Beamer
#ChatGPT
#PresentationTips
6
29
210
It's one of the best times ever to do a PhD in economics. Learn to work with ChatGPT and its API integration and you can suddenly do amazing stuff that's until recently been totally not feasible. Amazing opportunity to push the frontier
9
20
210
Today's prompt to restore law and order to your LaTeX preamble:
adjust my preamble. remove redundant packages etc and make small comments about what each package does
5
16
209
1/ Why do people invest in actively managed funds when index funds consistently give higher returns after fees? In a new paper,
@OleAndreasNaess
and I examine whether misperceived beliefs can explain this "active investing puzzle".
3
37
196
Indeed, it is important to filter out the students who are only in it for the money. In Scandinavia, where Ph.D. students tend to unionize, earn a living wage, and get an office with windows, morale tends to be low. Case in point: During summer, everyone even goes on vacation.
Graduate students are primarily students and trainees, not ordinary workers. Academia is a calling. I suspect unionization will tip the balance so faculty prefer, at the margin, to hire post-docs to do research in their labs rather than students.
394
69
366
4
7
196
For prospective Ph.D. students: Is it a good or bad experience? It depends, but it will probably be both... You need to write at least three papers. Coming up with novel research is hard. If you do something "new", people might tell you it's too exotic.
2
23
182
ChatGPT can boost research productivity and creativity. To get started, check out the thread below and my new slides deck where I provide guidance on integrating ChatGPT into your academic research.
4
39
174
Proud to publish "Beliefs about Racial Discrimination and Support for Pro-Black Policies" in
@restatjournal
together with
@cp_roth
. We first ask: how much disagreement is there about racial discrimination in the United States? [1/n]
2
40
163
Why do you tend to read ideologically aligned news? Because it feels good to have your beliefs confirmed – or because you tend to perceive aligned news as more accurate? Our recent
@EJ_RES
paper (with
@FelixChopra
and Chris Roth) uses an experiment to tackle this question.🧵
5
39
151
LaTeX packages I like
\usepackage{microtype} % Improves spacing
\usepackage{clipboard} % Copy-paste within and across docs; \Copy{key}{text} & \Paste{key}
\usepackage[plain]{fancyref} % Better cross-referencing; \Fref{tab:regressions} rather than Table \ref{tab:regressions}
4
15
147
A GPT for converting academic references into BibTeX entries
7
22
141
Considering doing a PhD in behavioral economics?
@TheChoiceLab
has three openings. Extremely generous pay, extremely generous research support/funding, and extremely competent research group. Apply now!
4
90
139
We should abolish the "job market paper", encourage PhD students to finish in four years with a dissertation of three papers, and move on to a much better paid job 🥳
What is your academic opinion that would leave people like this?
136
25
274
5
5
141
Are you applying for a PhD in Economics? Apply to NHH - great pay and unbelievable nature. We have an amazing group in behavioral/labor and deep pockets for research funding. If you're interested in media, polarization, and AI, I'm eager to work with you
1
28
131
A nuanced take on the R vs Stata debate: for most people, both work equally well. It's like the Oxford comma debate. Inconsequential stuff that people like to debate.
8
5
127
ChatGPT used to be useless for Stata, but has rapidly improved with GPT-4. It is particularly useful for manipulating string expressions and assisting with complicated loops (graphics remain a challenge). Check out a few new examples in my updated slides.
2
20
125
It's a big win for economics that we can follow these debates, on the researcher frontier, live, on Twitter
Let's talk about Auten and Splinter's work on US inequality 🧵
20
276
914
1
12
121
What makes a good research project in economics?
- You can tell your mother without being embarrassed
- You don't need "perfect results" for it to be a paper
- The setting is not too stylized
- The identification is clean
- It has some policy implications and is newsworthy
6
12
115
Congrats to my incredible friend and co-author
@cp_roth
for becoming co-editor of
@JEEA_News
. JEEA already has an amazing editorial board and Chris is a fantastic choice to strengthen it further, in particular its behavioral and experimental arm. 🥳🥳🥳
📢
@JEEANews
editorial board announcement! From Jan 1
@maiaguell
will be JEEA's 1st Data Editor, while
@mdoepke
Thierry Mayer
@sciencespo
@CEPII_Paris
@cepr_org
Christopher Roth
@UniCologne
become co-editors of JEEA. Full announcement (& more news)
2
15
175
0
4
110
We're hiring 4 predocs at FAIR, a center of excellence at NHH with deep funding and a *fantastic* behavioral/labor group. I'm hiring for my project “Media Bias and Political Polarization” in which we will use AI tools to open the black box of news demand:
2
36
107
Preparing the syllabus for my bachelor elective course at NHH this Spring. We'll do 1 paper per lecture. Students are only supposed to read the intro. The point is to showcase topics not typically explored in undergrad econ courses. Always a very fun course to give!
6
2
101
3/ Studies with null results are perceived to be less publishable, of lower quality, less important, and less precisely estimated than studies with statistically significant results, even when holding constant all other study features, including the precision of estimates (!).
2
15
83
If you're using
@overleaf
, be sure to pay attention to the Keyboard Shortcuts:
You might have guessed that "Ctrl + B" gives you bold text, but, for instance, using "Ctrl + /" to comment out several lines at once is a real killer.
1
10
80
Prospective PhD students should read this thread but also consider options in Europe. There are some world-class PhD programs in Europe that might be a better option than many of the "top" programs in the US.
If you're in a top-10-ish US undergrad, there is a playbook which still gets you a good shot at top-10 econ PhD programs straight out of undergrad
(I think it's extremely unfair that this essentially only works for top US programs, but, info is info)
59
74
626
1
11
79
7/ Our findings highlight the potential value of pre-results review in which the decision on publication is taken before the empirical results are known. The full paper is available on this link:
1
2
76
Awesome econ postdoc opportunity at the University of Bergen! 4 years, low teaching load (but enough to gain valuable teaching experience), and you're 100% free to follow your own research agenda. Apply now, and DM if you have questions.
1
26
70
Very important paper. I think a good start towards a better Ph.D. would be to change focus from writing a top-five worthy JMP (a goal massively out of reach for most PhD students) towards writing several papers and give people a sense of real progress along the way.
Forthcoming in the JEL: "Graduate Student Mental Health: Lessons from American Economics Departments" by Valentin Bolotnyy, Matthew Basilico, and Paul Barreira.
4
154
453
3
8
70
Talking about job market inefficiencies, committees should start to evaluate the candidates based on all their papers—not just the JMP—and adjust their expectations based on years of research time. That's a good incentive structure
3
3
69
Periodic reminder that if you're not using BibTeX Converter for adding BibTeX entries, you're doing it wrong
1
3
66
Tips to avoid common LaTeX mistakes
Dashes: Use --- for em dashes (—) and -- for en dashes (–)
Quotes: Use ` `these' ' for quotation marks, not "these", to get “these” instead of "these"
Citations: Use \citealt{key} over \citep{key} in parentheses to avoid double parentheses
5
8
66
4. Asking good questions is difficult. Good taste trumps everything. Learn how to acquire a good taste. A good question, as Jesse Shapiro puts it, is motivated by econ and not the econ literature. If the FT/WSJ isn't interested, work on something else (theory is an exception).
1
5
65
2. It's important to write many papers to learn how to actually write a paper and organize a research project. Invest in a good workflow early, it will come to hunt you otherwise
1
1
65
Spending the next six weeks in Copenhagen visiting
@CEBI_UCPH
! Spent my first day eating amazing herring smørrebrød and catching up with CEBI friends in the good Copenhagen weather. Drop me a message if you're around and want to go out for excellent coffee.
1
0
57
5/ Do people want to reward surprising results? Actually, the null result penalty is even larger when experts predict a non-null result, inconsistent with people perceiving the publication process to favor surprising results.
3
8
58
Many think of LLMs as search engines (e.g., "How old is the king?"). That's the wrong mental model. They excel as "input-output transformers": transform input A to output B (e.g., "translate this block of text from English to Norwegian" or "make this Stata script run in Python")
5
6
55
6/ Does the null result penalty depend on how statistical uncertainty is communicated? Yes! We find that the null result penalty is larger when the uncertainty is communicated in terms of p-values rather than standard errors.
2
4
53
Are you on the market and interested in media/behavioral econ? Do you plan to start a tenure track position in the Fall? Delay it for one year and come work with me (& Chris Roth+
@FelixChopra
) in Bergen
@TheChoiceLab
@NHHnor
We will study news demand with AI tools. Apply on EJM!
2
19
51
7-year tenure-track associate professor position at the Econ Department at the University of Bergen; DM me if you have questions about Bergen/Norway or the department!
2
20
50
Pro-tip for using GPT-4 when you're stuck with writing:
1) Give it what you have so far (e.g. a half-written sentence, a full sentence with one missing word that you fail to nail)
2) Add an "X" for the missing part/word
3) Ask it to provide 5 suggestions for X
1
4
48
I also want to emphasize: these are probably not all best practices and they can be extended in lots of different ways. But it provides a basic structure that makes submitting replication files easy. A very good best practice guide is found here:
1
7
47
Representative survey across *125* countries finds that 89% demand more action on climate change, yet they severely underestimate global support for action.. Correcting this pluralistic ignorance could be a powerful tool to spark action on climate change!
Hello world! We can share some good news today.
🌍 Support for climate action is widespread across the globe and much larger than it is perceived. 🌍
Our new article is out in Nature Climate Change:
w/
@TeodoraBoneva1
@FelixChopra
@Armin_Falk
2
31
78
2
6
47
When writing the paper in LaTeX, I can then just use these macros instead of actually looking up the numbers manually (an insane process). Want to change a control variable? No problem, all p-values and point estimates are adjusted with the push of one button.
2
2
46
7. Co-authoring is extremely important. Find good good co-authors. Attend good summer schools, visit good conferences, be in a good PhD program, etc. Make yourself useful as a co-author: Invest in quality code and quality writing. You will get far if you can do both.
1
2
46
8. You can ask questions across many different settings, but you should not invest in 5 different methods. Find your hammer. You cannot do theory, experiments, and big data. If you invest in multiple skills, focus on complementarities
1
0
45
After going to work in darkness for several months, the feeling when suddenly realizing that those days are finally over (picture taken 8.25am this morning in Bergen) is just sensational
3
0
45
First, create four folders: one for raw data (csv files), one for code (your do files), one for data (your cleaned files), and one for documentation (e.g. qsv files for Qualtrics)
1
1
44
The best approach is to *diversify* your portfolio: work on several projects simultaneously, and aim to have two or three papers ready after a few years. The JMP model where you're supposed to spend all your energy on a single "perfect" paper is just a terrible model for research
how do people deal with emotional ups + downs of research. like one minute everything is awesome and exciting and you love your job, and the next minute your variation is useless and you're having an existential crisis about your agenda. How do you people live like this!?
82
59
959
1
0
42
To save time and reduce the error rate in your papers, find a system to save all your statistics as macros. In Stata, this is pretty straightforward. Here is an advanced example from one of my studies in which I save point estimates, p-values, confidence intervals, etc.
1
3
41
We had a nice equilibrium where the sophisticates customized the color scheme and the naives went with the blue default. With this legit default scheme in Stata 18, the signalling value of slick figures is gone 😂
The default color palette has been completely revamped to better and more readible colors and this really give a clean sharp look 😍
1
1
22
4
0
41
@vectornomist
Behavioral economics as done by actual economists is always about foundations. The standard model is usually just a special case of a richer model (eg no present bias, no loss aversion, pure self interest). Psychologists who brand themselves as behavioral economists are different
3
0
40
More importantly, find someone you genuinely enjoy working with, have fun, and enjoy the process. Substitutes in production often turn into complements in idea generation
In many parts of economics building a research team is becoming invaluable. Two keys: put people around you who are complements in production not substitutes. Hire high IQ people who think they are middling types.
3
8
65
0
0
38
3. The publishability of a paper is a combination of 1) the question, 2) the method, 3) and the setting.
1
0
39
6. The setting is super important. Econ puts A LOT of premium on great settings. Clean identification is v important, but it's equally important to have a setting that people find exciting. There's a clear "effort heuristic" involved: it should be seen as hard. Invest accordingly
3
0
37
@AlexJParton
Most of these files are available in our replication package for a recent EJ paper
2
3
36
Interestingly, I have the exact same "Haaland" face when my papers are rejected
1
1
35
1/ Is clean identification overrated or underrated in economics?
My sense: both. A short thread 🧵
2
9
35
Bertil Tungodden and Alexander Cappelen for mentoring and building an incredible & extremely including hub for behavioral econ in Bergen
@TheChoiceLab
. Matthew Rabin & David Laibson for being amazingly generous with their time and organizing the incredible RSF Summer Institute
Today, let's also highlight the GOOD actors in economics. To get it started, I am grateful to Guido Imbens, who kept me from dropping out of graduate school,
@lkatz42
who serves as an advisor LONG after graduate school ended, and
@HilaryHoynes
, who I aspire to be.😊
32
40
594
1
2
34
Someone should write a paper on inflation narratives! (oh, wait: )
2
3
34
However, after 2.5 years, I collected data for 3 different projects during a 3-month period, which ended up as three papers in my Ph.D... The nonlinearity of this process is hard to understand before you have experienced it, and induces lots of stress before things "take off"...
1
0
33
4/ Is this effect specific to particular groups of economists? Interestingly, the penalty for null results is of similar magnitude for various subgroups of researchers, from Ph.D. students to journal editors.
1
1
33
2/ We recruit a sample of more than 500 economists and ask them to evaluate different hypothetical research studies. We vary whether a given study had a large and statistically significant main effect or a low and not statistically significant main effect.
1
1
33
It's amazing to have so many Ph.D. students working on research related to beliefs and economic behavior together in Copenhagen. There will be some fun days ahead!
We are incredibly excited to kick off a three-day Ph.D. course on Subjective Beliefs, Attention, and Economic Behavior
@CEBI_UCPH
in Copenhagen. We have an amazing group of Ph.D. students from all over the world attending the course!
@cp_roth
@Ingar30
@SettSonja
@JoWohlfart
2
9
149
0
2
32
Excellent advice
When former President Obama was asked to share his most important career advice for young people, he replied...
"Just learn how to get stuff done."
114
2K
8K
3
3
32
What are the secrets to success in academia, including economics? My guesses:
1) You have good ideas
2) You execute the most promising ideas without too much fuss
3) You can actually write a decent draft
Nothing fancy, but still ...
7
1
31
A point about LLMs: they rarely criticize you very much. If you ask "Is this good writing?" they might say yes and suggest some potential improvement, but it won't tell you " this is shit, you should start over". So: very helpful at the margin, but not good for critical feedback
5
0
31
Econ phd student? Apply for the Spring School in Behavioral Economics in San Diego! Amazing line of speakers and a great way to expand your network and meet new potential co-authors
0
16
30
Thanks for a great day at
@HelsinkiGSE
and an absolutely amazing sauna/Baltic sea experience. Cannot recommend this highly enough!
Fantastic to have seminar guests who insist on getting to sauna and to dip into the Baltic sea (still 10*C) before dinner. Thank you for your visit, great presentation, and for taking us there 👇
0
0
17
2
2
29
We should own the fact (and educate the public) that most experiments have ethical issues, but that very often the learning dimensions very much trump these ethical concerns (it's unethical not to experiment in some sense)
3
3
28
It's really awesome that more good journals are adopting the AERI format (also now possible in ReStat, EJ, and JPubE) – this greatly reduces the risk of writing short papers. More readable papers, faster path to publication, they're a great option for many projects.
1
0
27
In your code folder, include a "setup" file that defines all global commands that you will use between do files, including all paths
2
2
26
Ph.D. students, create a webpage where you include your contact information, a link to your CV, and links to all your papers. Keep everything on one page with no tabs. Super simple, and even if you have no traffic, you always have a convenient place to download your own papers
2
1
25
@AnthonyLeeZhang
The whole "spend 7 years to write an amazing paper that showcases how great you are" is just a terrible model for students. Spend 4 years, write 3 actual research papers, pick the best one as your JMP, get out and get credit for all your work and for finishing on time
0
0
25
At
@TheChoiceLab
the activity level is higher than ever; come visit if you want to meet good people doing behavioral or labor economics!
(Now that faculty aren’t coming into the office anymore) Will universities ever recover?
7
20
97
2
1
24
5. The method is also quite important, but there's lots of freedom here. You can do well with an RCT or you can do an advanced econometric technique. If you have a *new method* to exploit (e.g. new uses of AI tools), great. If not, it's not the end of the world.
1
1
24
@mushfiq_econ
Adding to this, I think people sometimes do not realize or internalize that Twitter is actually a public space where the people you ridicule actually see your messages
1
1
23
I find ChatGPT incredibly useful for all sorts of tasks, especially coding assistance. The one exception is Stata, for which it is mostly useless (makes it tempting to switch to R). If this is an open/closed source software difference, I'm very bearish on closed source software.
2
1
21
A $200 fee for a grad school application makes no sense
Standard advice is to "apply to as many schools as possible." Taking the GRE is $$, sending scores is $$, and each uni's app fee is +$75. Last year, Chicago Harris' fee was $200! The potential implications for inequalities w/i the discipline merit further thought
5
18
148
5
1
21
@kearney_melissa
@bsosumd
For supposedly being experts on social norms, it's funny how many sociologists somehow think norms do not apply to them on the Internet
1
0
20
Empirical data tools/data analytics as well as causal inference/experimentation are safe paths both for research and industry jobs. Economic theory/math is another track that could work, but you need to learn it really well... You pick up skills by doing research, not by courses
1
0
20
Over 70% of the researchers who answered my poll "rarely or never" use ChatGPT for research-related tasks. I use it daily for all sorts of tasks and benefit a lot from it. It will only become better with time (and your prompting skills will improve with practice), so why wait?
As an academic researcher, how often do you use ChatGPT for research-related tasks (e.g. coding assistance, brainstorming for ideas, proofreading your papers)?
3
0
6
1
0
20
PhD student? Come to Bergen in August to attend this amazing course with George Loewenstein (and then you can visit the fjords afterwards!)
FAIR organizes a PhD Course with George Loewenstein. The course will provide a PhD-level introduction to Behavioural Economics – the application of insights and research findings from psychology (and other disciplines, including neuroscience) to economics
0
12
32
1
9
19
I think it's generally good advice to invest hard in learning some skills that are either in demand at private institutions or really take you to the research frontier at some technique/method. You don't realize it during the Ph.D. but you are extremely rich in time; use it well
1
0
19
2/ I'm building a Qualtrics survey and need to validate that a response option is a valid URL. Let's ask ChatGPT for help.
1
0
18
@guygrossman
I like coefficient plots as they clearly communicate the confidence interval while a table seductively makes you focus (too much?) on the point estimate. But my preference is to show a figure without controls & a table with controls etc
0
0
19
6/6 Our structural estimates suggest that the two motives have a similar quantitative importance on people's demand for news. Full paper available on this link:
0
1
18
In your code files, run this setup file before you load and save your data/output
1
1
18