Image: ESO/C. Malin (christophmalin.com)How to read an astronomy paper as a beginner
Almost every astronomy paper written since the early 1990s can be read free of charge on arXiv, usually months before it appears in a journal. Few other sciences are this open. What keeps beginners out is the format, the jargon and the feeling that a paper must be read start to finish like a novel. Working astronomers almost never read papers that way.
Where the papers are
The preprint server arXiv was started by the physicist Paul Ginsparg in 1991. Its astrophysics section, astro-ph, receives dozens of new papers every working day, and the daily list is the morning newspaper of the field. The section is split into six areas: galaxies, cosmology, Earth and planetary astrophysics, high energy phenomena, instrumentation and methods, and solar and stellar astrophysics.
The search engine of astronomy is the Astrophysics Data System, or ADS, which NASA funds and the Smithsonian Astrophysical Observatory runs. Type in an author or an object name and it returns the papers, a link to the free arXiv version, and the list of later papers that cite each one.
A preprint on arXiv has not necessarily passed peer review. Look at the comments line under the abstract. "Accepted for publication in ApJ" means referees have approved it, while "submitted to MNRAS" means they have not yet. The main journals are The Astrophysical Journal (ApJ), The Astronomical Journal (AJ), Monthly Notices of the Royal Astronomical Society (MNRAS) and Astronomy & Astrophysics (A&A), plus Nature and Science for results with wide appeal.
The anatomy of a paper
| Section | What it does |
|---|---|
| Abstract | The whole paper in about 200 words, with the main numbers |
| Introduction | What was known before and which gap this paper fills |
| Observations or data | Which telescope, which instrument, how long, how the raw data were cleaned |
| Analysis or methods | How the data were turned into physical quantities |
| Results | The measurements, mostly as figures and tables |
| Discussion | What the results mean and how they compare with earlier work |
| Conclusions | A numbered summary of the findings |
Theory papers replace the observations section with a description of the model or the simulation. The order is otherwise much the same.
Read in three passes
On the first pass, spend ten minutes. Read the title and the abstract, then go straight to the figures and their captions, then to the conclusions. Astronomers put their result into a plot, and a good caption tells you what the axes are and what to look at. After ten minutes you should be able to say what was measured and what the authors claim.
The second pass takes an hour. Read the introduction properly. It is a short review of the subject written by experts, and its references point you to the two or three earlier papers that matter. Then read the results and the discussion. Leave the methods alone for now. Every unknown term goes on a list that you look up afterwards, so that you do not lose the thread.
The third pass is the methods section, and it is optional. You need it only if you want to judge whether the result is sound or to do something similar yourself.
Finish by writing three sentences in your own words: what was measured, how, and how sure the authors are. If you cannot write them, go back to the abstract.
A worked example
Take the paper that reported the most distant galaxy known, by Rohan Naidu and collaborators. Its title is "A Cosmic Miracle: A Remarkably Luminous Galaxy at \(z_{\rm spec} = 14.44\) Confirmed with JWST", and it is free on arXiv.
The title already contains the result, once you decode it. The symbol \(z\) is the redshift, and the subscript "spec" says it was measured from a spectrum, which is far more reliable than an estimate from colors. The abstract adds that the galaxy is seen "a mere 280 million years after the Big Bang" and that the redshift rests on "a sharp Lyman-\(\alpha\) break". Neutral hydrogen between the galaxies absorbs almost all light with wavelengths shorter than 121.6 nanometers, so the spectrum of a distant galaxy drops to zero on the blue side of that wavelength. At \(z = 14.44\) the drop is stretched to 1.88 micrometers. Finding the step in the spectrum gives the redshift.
The abstract also mentions "\(\sim 3\sigma\) detections of five rest-UV emission lines". That phrase is a statement about confidence. Each line is weak on its own, and the authors say so. The break carries the claim and the lines support it. The skill to practice is reading for the strength of the evidence. The scientific background is in what the James Webb Space Telescope has found so far.
Jargon that trips up everyone
| Term | Meaning |
|---|---|
| \(z\) | Redshift. Wavelengths arrive stretched by a factor \(1 + z\) |
| mag | Magnitude. Smaller numbers are brighter, and 5 magnitudes are a factor of 100 in flux |
| dex | A factor of ten on a logarithmic scale. 0.3 dex is a factor of two |
| \(\sigma\) | Standard deviation. By a common convention 3 sigma counts as evidence and 5 sigma as a detection |
| Metals | All elements heavier than helium, including carbon and oxygen |
| \(M_\odot\), \(L_\odot\) | Mass and luminosity of the Sun, used as units |
| erg | The energy unit of the older cgs system. One erg is \(10^{-7}\) joules |
| pc, kpc, Mpc | Parsec (3.26 light years) and its multiples of a thousand and a million |
| et al. | "And others". Large collaborations can have hundreds of authors |
Several of these units are explained with examples in ten equations every astronomy student should know.
How much to believe
One paper is one group's result. Check the size of the error bars before the size of the claim, and ask what else could produce the same signal. Then open ADS and look at who cited the paper. In 2025 a reported hint of a biological gas on the exoplanet K2-18 b was followed within weeks by independent analyses that found the evidence too weak. Anyone who checks the citations sees that exchange.
Two aids are worth knowing. Astrobites is a site where graduate students summarize one new paper each day for undergraduates. Review articles, such as those in the Annual Review of Astronomy and Astrophysics, give an overview of a whole topic and are the best starting point when you are new to it.
Reading gets easier quickly. The tenth paper on a subject goes much faster than the first, because introductions repeat and the same plots return.
In the A&A Masterclass, each unit contains a research spotlight that presents one published paper by an active researcher, with a link to the original. Unit 16 includes a section on reading research papers and on where a student can go next. The curriculum lists the units.
