How to read an original scientific research paperWord格式文档下载.docx

1、, i.e. to answer questions which interest you in a broader context （even if its simply the context of getting an essay written）. There is a place for browsing, but blotting-paper minds soon get overwhelmed and confused, and their owners often find there is too little time and energy to complete the

2、main tasks they are set!2. How a paper is written2.1 Authors ought to start by assembling their data concerning the central observation into a Table or Figure, as part eventually of the account of their Results. The other supporting observations （controls, preliminary investigations, extensions of t

3、he main finding to other systems or by other techniques） are also organised into Tables or Figures.2.2 Then they do Materials and Methods （because its easy）, and Introduction （because they should have thought long ago why they are doing it）2.3 Then Results. These work outwards from the key Table/Fig

4、ure, i.e. with pilot/preliminary ground-clearing experiments; other supporting observations; extrapolations to related systems; control experiments.2.4 Then Discussion. This ought to start with a re-hash of the main observation, recapitulating simply what was found. Then the limitations of the obser

5、vation （perhaps technical, perhaps conceptual） should be discussed. This could lead to a piece about controls, and the relative emphasis to be put on the different observations presented in the results. Then there must be a consideration of the findings of other authors which relate to the present f

6、indings （contradictions as well as agreements）; then perhaps a more speculative part about theories/models/future hypotheses worth testing, and how to go about it. A conclusion should summarise the relative weight which can be put on these different aspects （how much is pretty certain; how much is s

7、peculative still）. Good papers usually end on an upbeat note to show where the work is leading to.2.5 References, an abstract and a title follow.3. Deconstruction of the finished product - the readers job3.1 The finished article is a mixture of dispassionate observation and opinionated interpretatio

8、n, which is normally very heavily influenced by the current fashions. It is also an advertisement of the authors opinions （it even sometimes says that: for legal reasons American journals mark articles advertisement when page-charges have been levied for publishing an article）. Be aware that there i

9、s not a little selling going on. Your job is to disentangle what is reliable fact, solidly based in good experiment, and what is inspired guesswork or opinion or prejudice or unfounded faith or vapid verbiage. Off you go: 3.2 Familiarise yourself with what the authors say the paper is about. Read, i

10、n the following order: the abstract the part of the Introduction that is properly background （be careful - many Intros these days become glorified Abstracts, with an anticipation of the results. This is part of the selling process） the first paragraph of the Discussion: this should identify what the

11、 authors think is their main finding.DO NOT READ THE PAPER LINEARLY THROUGH FROM BEGINNING TO END;you are much too likely to fall for the authors views this way.3.3 Identify the Figure or Table which contains the meat of the paper - the key result3.4 Use the Materials and Methods and the text of the

12、 Results to understand sufficiently how the experiment was done to produce the key result, but skim the details3.5 STOP AND THINK4. Your own interpretation: If you had got their result what would you have concluded?4.1 Is it reliable? How reproducible? What is the variance of the measurements? Does

13、it look like a substantial effect? （Log scale?, linear scale? expressed as a percentage of some control?） In vitro artefact? Could a skilled but unpractised observer get the same results? Is it internally self-consistent?4.2 In interpreting it, what assumptions need you make? Is it representative of

14、 the whole system being studied? Is it physiological （concentrations used? cell populations highly selected? cell lines - how near normal?）4.3 What about controls - positive （to show that an effect could have been clearly seen if present）; and negative （where a key reagent/condition was altered to s

15、omething similar which had no effect）. Absence of evidence is not evidence of absence: if something doesnt happen, it doesnt mean its not there.4.4 Continue reading the Discussion, and those other Results that seem material.5. Comparison between your interpretation and theirs5.1 As long as you have

16、not taken textbooks too seriously （avoid fundamentalism!） and are open-minded, you should have the great advantage of zero prejudice: not wanting （needing?） it to turn out one particular way. You are coming to it fresh and baggageless. You get this chance only once! Soon, you too will have prejudice

17、s! Just because the authors are professionals and have wriggled it past a couple of referees, does not mean your opinion is inferior.5.2 Do your and their views agree? If not, is it just on relatively minor points, or is it major? If there is disagreement, are you making different assumptions?5.3 Do

18、 their views and those of other authors agree? If there is a conflict between papers, are they really addressing the same issue? Is one or other set of results too broadly interpreted? Which approach to the question is more direct and more physiological?5.4 What strategy would you adopt to test furt

19、her any discrepancies and distinguish the truth of the alternative hypotheses? How would you falsify the model/hypothesis? Has anyone done such experiments already?6. Salesmanship - points to beware6.1 Assertive Sentence Titles （ASTs in the trade; e.g. Eating spinach improves splenic IL-2 production

20、） - the headline says it all!6.2 Introductions which summarise the conclusions - plugging the message early and repetitiously!6.3 Glib referencing in the Intro to background reviews and not to original papers7. Quality of writing（This section is more relevant to the fine analysis which reviewers sho

21、uld undertake）7.1 Are the references accurate? Check with bibliographic database.7.2 Are techniques described in enough detail （or sufficiently well referenced） that a competent scientist could repeat the method?7.3 Is the style clear? Frequent usage of strings of nouns in apposition is a bad sign a

22、nd often leads to ambiguity （e.g. High Class Family Butchers7.4 Is the Abstract a fair summary of what was done?With luck, and practice now, these skills will become second nature when you read a report of any kind. You will doubtless have plenty to read in your lifetime, and important decisions may hang on them. Dont become a pedant; but do get critical!Feedback on this document will be much appreciated. E-mail me on simon.huntpath.ox.ac.uk.G:SVHUNTTEACHINGFHS_IMMIMMLECSPAPRREAD.DOC 23 May 2000