The art of proposal writing
Why am I reading this?
Astronomers use telescopes. But telescope time is a limited resource, that means, access to the telescopes is strictly regulated. The most common way to get access to the big machines is by asking for time through an observing proposal, which is then reviewed by other astronomers. For that reason, proposal writing is part of our job. Over the course of the last two decades I’ve authored dozens and co-authored more than hundred proposals for observations at small and large telescopes. Most of these proposals failed to get time. But it’s not a lottery, at least not always, and good proposals get time more often than bad ones. Therefore it is of obvious practical use to learn how to write a good observing proposal. If you cannot do that, it is less likely that you will be able to do the science you want to do. Coincidentally, proposal writing is also good preparation for writing applications for other useful things, like money, or jobs, or promotions.
In a nutshell
Observing is like asking the Universe a question. Now, everybody can just ask questions, point the telescope somewhere, take an image, done. A good observation is a question for which you can actually expect to get a meaningful answer from the Universe. And writing a proposal is the way to get there.
With that in mind, the three main features of a good observing proposal are: focus, details, and novelty.
Focus means that an observation needs to be able to answer a specific question. It needs to have a clear realistic goal. This is unlikely to be a big question about the Universe (like, where does the world come from), because big questions tend to be vague. Those big questions, however, provide the framework for the proposal. It is important to ask for results that inform a wider context of research, not just a tiny niche.
Details are necessary when setting out the exact plan for the observations. When observing with big telescopes, you cannot just wing it. Every parameter of the telescope and instrument has to be thought of, simulated, justified. Targets have to be vetted carefully. The literature has to be checked for previous results. You have to know more about the observing program than you can fit into the limited space of a proposal.
Novelty means that the observation should reveal something about the Universe that wasn’t known before. You very rarely get time to go and repeat exactly the same observations as before, unless you expect to get a different answer (for example, because your object is variable) or because you want to solidify a previous interesting result with a different method or a different telescope or a different sample. You are not necessarily aiming for a big breakthrough, incremental steps are the norm, but something new has to come out of the proposal.
The review process
Observing proposals are typically reviewed by a panel of astronomers. Some telescopes offer subject-specific panels, but in general you cannot expect the people on your panel to be experts in your sub-discipline. Some panels use external referees to provide a peer-review. Panels are typically three to ten people large. First, individual panel members review all proposals. Then the panel meets, discusses, and decides on a ranking. The best proposals according to the ranking tend to get time. Applicants get sparse feedback from the panel. That feedback may or may not be helpful in the future.
The most important point to realise about the review process is that panel members do not have a lot of time for each proposal. My guess is that the review for a given proposal takes maybe between ten and thirty minutes, plus a short panel discussion. That’s not a lot to digest an entire science case in a field one might not be familiar with. You can expect that people will carefully read the abstract, they will look at figures and check some aspects, but parts of the text might only get a superficial glance. Keep that in mind when writing the proposal.
A path to the first proposal
It is not easy to ask a question, and then come up with an observation to solve it. Many questions are unanswerable with current technology. Others have been answered a long time ago. Instead, it may often be easier to work backwards, to start with a feasible, novel observation and then figure out what can be learned from it. This is a bit like playing Jeopardy, to start with the answer, and work your way to the question — a shortcut to a working observing proposal.
Step 1. Take an object you like. A quasar, a cataclysmic variable, an exoplanet host star. It helps if it is a target you know well. Most PhD students in observational astronomy will have a large pool of interesting objects. Start with some of those. Find out more about this object, using tools like SIMBAD.
Step 2. Find a type of observation that is feasible, but has not been done yet for this object. A high-resolution spectrum. An image with Adaptive Optics. A high-cadence time series. An image, but deeper than ever before. It helps to be familiar with that specific technique, or to team up with someone who is familiar with it. Use observatory websites. Start playing with Exposure Time Calculators. Pick the telescope and instrument that you need for this observation.
Step 3. Figure out what kind of question you can possibly answer with this type of observation. If you got nothing here, change step 1 or step 2 and try again. Re-iterate until it makes sense.
Practical advice on proposal writing
Typical proposal are made of a science case and a technical case, each maybe one page long. That’s the relief — it’s not a lot of writing. Writing a proposal takes a couple of hours, after that it’s polishing. Before that, it’s legwork. Proposals are ideal to write as a team — someone does the science, someone else does the technical parts, for example. Other team members revise and edit.
An ideal science case starts with a big question and then whittles it down, step by step, to one simple question that can be answered by this specific observation. It works like a funnel: At the beginning, many paths are possible, but at the end, you are limiting yourself to one narrow line of attack. The science case needs references, but you don’t get points for citing as many papers as possible. Somewhere, either in the science case or the technical case, you need to explain why you want to observe target X and not something else entirely. What makes your target special and suitable?
The technical case needs to show the way from a specific telescope setup to the results that you promised in the science case. It needs to include and justify all parameters for the observations, exposure time, moon phase, filters, seeing, and so on. That means, you have to prepare just like for an observing run, just without packing suitcases. You are going to spend a lot of time playing with an Exposure Time Calculator, to simulate your observations, to find out what works and what doesn’t. Here you also show that you are prepared to work with the data, perhaps familiar with the pipeline, ready to make sense of the images or spectra. The technical case can be creative, it can be optimistic — pushing the telescope to its limits is good — but it should not be delusional. Reviewers know that you can’t observe a lump of rock in Andromeda with a 1-m telescope.
The proposal is complemented by an abstract and one or two figures. Spend lots of time on these two items. The abstract is simply a short summary of science case and technical case. The most basic feature of an abstract: It makes the reader curious, but it also conveys the basic idea of the course of action. If you are stuck right at the beginning of the abstract, begin with “we propose to do X to achieve Y”, or something of that nature. Figures need to be very very clear and self explanatory. A figure can absolutely make or break a proposal.
Language matters. Good proposals use short and simple sentences. They explain all acronyms. They make assumptions explicit. All that is geared towards avoiding confusion. Confusion is bad. Also to avoid are conditionals: “We will do”, instead of “we might do”. You want to show that something works. If you have doubts, work on them, and then tell the reader why it works. Be specific, firm, and not vague and handwaving.
The pressure factor
The panel ranking is not the only parameter that determines the final allocation of telescope time. Scheduling constraints also matter, a lot — where are the targets on the sky, what observing conditions are required, when is the target visible, and so on. Therefore it makes sense to think hard about the requirements for success. Can the observation be done during full moon, a very unpopular time for many projects? Does it really need excellent seeing? Is the applicant willing to go observing when others prefer to stay at home? As a PhD student and postdoc I had a lot of success with proposals which invariably were scheduled over Christmas. It took me years to figure out why.
The pressure factor is the critical metric to look for on observatory webpages. It tells us how oversubscribed a telescope is. Sometimes it is given for the number of proposals — the ratio between proposals submitted and proposals which got time. Sometimes for the observing time itself — the ratio between hours requested and hours awarded, overall. A pressure factor of 2 is really good. A pressure factor of 10 is almost insane. If your proposal can be done with multiple telescopes, it makes sense to go hunting for the one with the lowest pressure factor. Sometimes that is not an option. Hubble exists only once. Same for ALMA or JWST. In that case you have to go and compete with millions of other astronomers. Another informal rule: Observatories that do not publish this information usually have low pressure factors.
When writing a proposal, failure is normal. Playing the odds is normal as well. Good luck!
