Now that you're on board with the need to read research articles and the benefit of writing summaries of experiments from those articles (and if not, then try following the preceding links), on to how to go about doing it.

Which article? Which experiment?

• You are seeking a research article (not news, not review), one of interest to you, possibly related to your future proposal.
• A research article can be recognized by its presentation of actual experimental results (not merely conclusions) and a description of the methods used to obtain them, sufficient to enable someone else to replicate the experiment. If you can't identify both, you haven't found a research article.
• Find one or more using previously described search strategies.
• Your life will be easier if you choose an experiment that is relatively self-contained, one that can be described without needing to slog through other experiments in the article.
• Most importantly, choose an experiment that suits your purposes, one that is central in some way to your research proposal or can teach you something important about your chosen topic.

What if I don't understand half of the experiment?

Don't panic! Focus on the other half, and ask what, if anything is missing from the story you are trying to tell. If you are stumped by a technique used in the article (and this may be a common experience), first ask yourself if you need to understand this technique in order to tell your story. If not, then forget about it. But if you judge the technique to be essential for your understanding (and your reader's understanding) of the experiment you've chosen, then do whatever is necessary to gain the insight you need. Hit the web. Read another article. Talk with your mentor. Talk with your colleagues. Talk with me or a TA. Whatever it takes.

How do I write a summary of a research article?

Your summary should be aimed at an audience of your peers, students like yourself except without the benefit of having experienced this course. It should consist of a few well organized short paragraphs that address the following:

• Introduction of problem
  What is the big picture surrounding the experiment? How is that area of general interest logically connected to the question addressed by experiment? What hole existed in our knowledge before the experiment was performed?
   
• Description of experiment
  Describe how one experiment was performed in sufficient detail that the reader will thoroughly understand the principle. There is no value, however, in providing detail that is useful only in replicating the experiment (e.g. the composition of a buffer, etc). Avoid terms that will be unfamiliar to the reader, or if they're useful, then define them.
   
• Result of experiment
  Describe what was observed, not the authors' conclusion based on that observation. You will almost always want to present a figure or table from the article, incorporating part of it into your summary. Don't be afraid to relabel or even remake it so that it may be more readily understood by the reader.
   
• Observations and connections
  What insight can be drawn from the experimental result? Does it answer the question? Does it have higher implications? It may or may not be useful to briefly relate other results from the paper, but do this only to the extent they illuminate the experiment you've chosen to present.

What NOT to do?

• Don't paraphrase. You might think that paraphrasing is a reasonable strategy. "After all," you might say, "the authors know what they are talking about and I don't." That may have some truth, but the authors are more ignorant than you in two important respects: They don't know the story you are trying to tell, and they don't know your audience. The authors had a much larger story to a much different audience. It's just as likely that you can piece together a coherent and pertinent story from theirs as you could piece together a hummingbird from a Boeing 787. The authors did not writxee a summary of a single experiment. That's your job.
  
  How to avoid paraphrasing? Easy. Put the paper down and walk away from it. Imagine someone who knows nothing about the article, e.g. yourself a few days earlier, then explain the experiment in simple terms to that person. If you can't, identify why you can't and fix the problem -- read the relevant section of the article more carefully, go on the web to find a good explanation of a technique, find someone who may help you understand what's going on. Then walk away and try agin. Finally, write down what you said. Go back later to fix the parts that sound funny.
  
  More importantly, take responsibility over the task before you. No one in the history of the world has ever been called upon to do what you need to do: to explain this one particular experiment to your peers. That wasn't the authors' job and they didn't do it. Keep your audience in your mind at all times -- what do they need to know right now in order to understand the particular point you need to explain?
   
• Don't quote. This isn't a literary analysis. Quotations are almost never used in scientific writing, because it is generally the content, not the phrasing, that's important. I don't care what you learned in English classes. In scientific writing, authority doesn't matter. It's the content that counts. Just give us what you want to say in your own words. But what if you don't understand the content...?
   
• Don't pass your ignorance on to your reader. If you don't understand what the authors are saying, neither will we. Find a way to avoid a topic you're confused by, or if there's no way around it, do whatever is necessary to gain insight.
   
• Don't pass on useless terms to your reader. Just because the authors used a term, that doesn't mean you should. Their audience is not your audience. Every term puts a burden on your readers. Weigh that burden against the benefit of using the term. If the benefit isn't big enough, don't use the term. Use simple English instead.
   
• Don't merely recite conclusions. Interpretation is fine, but only after you've given us the facts of the matter, the results. What was the actual data? For us to understand, you will need to show a key result, in a way that is comprehensible to your audience.
   
• Don't forget your audience. Your audience is people like yourself. I am not your audience. Always ask yourself whether you would have understood last year what you are writing.

Examples

• Here is an example of a good summary.
• Here is an example of a bad summary.
• And here's one in between (though not just right).