What if the paper is behind a paywall and I can only see the abstract?

Search for the authors' names alongside the paper title, since many researchers post free copies on their university page or a preprint server. If you truly cannot get the full text, treat the abstract as a claim to be curious about rather than a fact to repeat, since you cannot check the data or methods behind it.

How do I tell if a journal is any good?

Look at whether the journal is peer reviewed, meaning other scientists checked the work before publication, which is true of most established journals but not of preprint servers. Beyond that, a journal's general reputation in its field matters more than any single number attached to its name, and that reputation is something you build a feel for over time rather than something you can check in one search.

What's the difference between a preprint and a published paper?

A preprint is a draft that has been posted publicly before, or sometimes instead of, formal peer review by other scientists. Preprints let research circulate faster, but they have not been checked by independent experts yet, so treat their claims as more provisional than those in a peer-reviewed journal article.

Should I worry about the number of citations a paper has?

Citation count is a rough measure of how much attention a paper has received from other researchers, not a direct measure of whether it is correct. A highly cited paper has usually been scrutinised more, which is useful, but plenty of solid, useful research sits with modest citation counts simply because the field is small or the paper is recent.

Is it ever fine to just read the abstract and move on?

For most everyday purposes, yes, especially if you are trying to get a general sense of a field rather than make a decision based on one study. The fuller reading order matters most when a single paper is being used to justify a specific claim, a product, or a change in behaviour.

How to Read a Scientific Paper: The Order That Works

21 September 2026

Most people open a scientific paper at the top and give up somewhere in the introduction. Read it out of order instead: abstract, then figures and results, then methods, then discussion, and the introduction last. This gets you to the actual finding fast, then lets you judge whether to trust it.

Why the usual order fails

Papers are written for other specialists, not for you. The order on the page follows the logic of a formal argument, not the logic of a curious reader trying to get an answer quickly.

The introduction is written to justify why the study was worth doing at all. It is full of background, citations to other work, and careful hedging. None of that tells you what the researchers actually found. By the time a non-scientist reaches the results, they have often run out of patience.

Reading in the order the authors intended means you absorb their framing before you have seen a single piece of evidence. That makes it harder to judge the work on its own terms. A different order fixes this: see the finding first, then decide how much to trust it.

Start with the abstract, but treat it as a headline

The abstract is a short summary at the very top, usually one paragraph. Read it first, but read it the way you would read a newspaper headline. It tells you what the paper claims, not what it proves.

Abstracts are written to sound important. Words like significant, novel, and robust get used loosely. Note the claim being made, note the general size of the study if it is mentioned, and move on. Do not decide yet whether to believe it.

One thing worth checking here: does the abstract state a clear result, or does it hedge with phrases like may suggest or could indicate? Heavy hedging in the abstract is often a signal that the finding itself is thin.

Jump straight to figures and results

Skip the introduction entirely for now and go to the results section, or better, straight to the graphs and tables. This is where the actual data lives, before anyone has told you what to think about it.

Look at the scale on the axes of any graph. A dramatic-looking bar chart can be hiding a tiny real difference if the axis has been cropped to start at, say, 90 instead of zero. Look at the size of the groups being compared too. A striking result from twelve people means something very different from the same result across twelve thousand.

If there is a table of numbers, look for a measure of uncertainty next to the main figures, often written as a range in brackets or with a plus-or-minus sign. A result with no sense of uncertainty attached to it is a result you should treat with caution, whatever the summary sentence says.

Check the methods before you trust anything

Once you know what was found, go to the methods section to see how it was found. This is the least readable part of the paper for a general audience, full of technical detail, but you do not need to understand all of it. You need to answer three questions.

First, what was actually measured, and does that match what the headline claim says was measured. A study that measures a change in a blood marker is not the same as a study that measures whether people got better. Second, what was compared against what. A group that took a new treatment needs to be compared against a group that did not, ideally chosen at random, for the result to mean much.

Third, how many people or samples were involved. Small numbers are not automatically wrong, but they make single studies far less reliable, and they make it more likely that an interesting-looking result happened by chance.

Read the discussion for the spin

The discussion section is where authors interpret their own results and argue for why they matter. Read it after you have already looked at the data yourself, not before, so you can spot the difference between what the numbers show and what the authors want them to mean.

Watch for the gap between the data and the claims. A paper that found a small effect in mice is not the same as a paper that found a cure for a human disease, even if the discussion section gestures hopefully in that direction. Good authors are explicit about the limits of their own study, in a section usually called limitations. If that section is short, vague, or missing, treat the paper's conclusions with more caution.

This is also where you will find the authors comparing their result to earlier work. If they say their finding contradicts several previous studies, that is worth noting. One paper against a pile of others pointing the other way is not yet a reason to change your mind about anything.

Leave the introduction for last

By the time you reach the introduction, you already know what was found and roughly how solid it is. Now the background information is genuinely useful, because you have a place to put it.

The introduction explains why researchers thought this question was worth asking, and it usually cites the existing evidence on the topic. Reading it last lets you check whether the framing matches the result. Sometimes a paper is introduced as answering a big open question, and then the actual result is a small, narrow piece of that puzzle. Knowing this order helps you spot the gap.

If you are short on time, the introduction is the section you can skip most safely. You lose context, but you keep the finding and the evidence for it, which is the part that actually matters.

Common questions

What if the paper is behind a paywall and I can only see the abstract?: Search for the authors' names alongside the paper title, since many researchers post free copies on their university page or a preprint server. If you truly cannot get the full text, treat the abstract as a claim to be curious about rather than a fact to repeat, since you cannot check the data or methods behind it.
How do I tell if a journal is any good?: Look at whether the journal is peer reviewed, meaning other scientists checked the work before publication, which is true of most established journals but not of preprint servers. Beyond that, a journal's general reputation in its field matters more than any single number attached to its name, and that reputation is something you build a feel for over time rather than something you can check in one search.
What's the difference between a preprint and a published paper?: A preprint is a draft that has been posted publicly before, or sometimes instead of, formal peer review by other scientists. Preprints let research circulate faster, but they have not been checked by independent experts yet, so treat their claims as more provisional than those in a peer-reviewed journal article.
Should I worry about the number of citations a paper has?: Citation count is a rough measure of how much attention a paper has received from other researchers, not a direct measure of whether it is correct. A highly cited paper has usually been scrutinised more, which is useful, but plenty of solid, useful research sits with modest citation counts simply because the field is small or the paper is recent.
Is it ever fine to just read the abstract and move on?: For most everyday purposes, yes, especially if you are trying to get a general sense of a field rather than make a decision based on one study. The fuller reading order matters most when a single paper is being used to justify a specific claim, a product, or a change in behaviour.

Lucent, once a week

We read the new research and write up what actually holds, in the same plain language as this page. No jargon, no hype, three papers a week.

Read the latest issue