Reading academic papers

Finding papers to read

There are many ways to find academic papers. As well as using your favourite internet search engine, there are many indexes of medical and biological papers with effective search functions. Pubmed is by far the most popular and very comprehensive, maintained by the NIH, part of the US federal government, thank you. Indexation by Pubmed does lag publication by six or more months, and this will be true of all the academic search engines, like… Jstor, also popular.

So if you simply have to read it as soon as it is published you will have to subscribe to the journals, or having read the abstract in a truncated version of a journal online pay for the whole paper.

When you use the medical research databases remember to use multiple different search terms when looking for papers on a particular topic. You must also use the same terminology as those favoured by researchers and academics. To an extent their use of specialised language makes sense, unfortunately many researchers seem to have never met a complex word they did not like, regardless of the utility derived from using one. So, as well as searching for keywords such as “helminths” or “hookworm” remember to try using scientific names of organisms such as “necator americanus” and “trichuris trichiura”, or using the academic style for contractions as “t. trichiura”. Of course you can simply use “trichiura” instead, but remember with medical research a great deal is conducted in animals. Particularly at the early stages, as we are now with helminthic therapy. So the animal equivalents are often used, as in T. Suis for the whipworm species having pigs as their definitive host.  So search using terms starting at the centre, and move out.

Terms like “definitive host”, and “zoonotic infection” are also terms you will find helpful to master, but most will come from reading on the subject. Or, you can buy a parasitology text if helminthic therapy is your area of interest, and texts in the human immune system, etc.

There are various online medical references, but the best are not online, or not for free. Amazon again provides great choices, but textbook pricing is an outrage, be prepared for texts, new ones and used, to cost over a hundred dollars, although medical dictionaries are usually less than that.

It is also helpful to start by limiting your search to review articles to get a hold on a subject before moving into the primary literature. Review papers are digests of current research, often commissioned by the journals themselves, that are as the name suggests and are a distillation of current research and thinking on some usually fast developing field that is getting a lot of attention. Because of this they are very easy to read, and are a good starting point for the lay reader.

If you like an author’s work, as I do with Correale for MS and helminths, using their name in any search will quickly lead you to other papers from that author. Research in any area that is interesting and complex develops over a series of papers. Often the same cast of characters are involved, and you will also see different groups working with and in competition with one another, but often working in variations of the same group of people.

The order of authors at the top of the paper indicates a lot about their role. The first author of a primary research paper will be the one who did the majority of the experimental work, usually a postdoctoral researcher (postdoc) or PhD student. Second authors will typically have contributed some experiments to the work. Authors in the middle of the paper may have contributed materials, helped with methods or been part of the larger team to contribute intellectually to the paper, or simply ran a sample through PCR analysis – think of Executive Producer credits on movies. The authors at the end of the papers are usually the research group leaders, so they approved the PhD candidate or approved the budget or use of facilities, or actually contributed. The final author being the one who was primarily in charge of the research, it was usually their idea, or they claim it anyway, poor old postdocs.

At the end of each paper is also a bibliography of the research that the authors have read or referred to, or that contribute to the area of study, and are also a great idea to look at for related research if you find a paper interesting. This can lead to some great material in my experience, and papers that are referred to often are clearly thought to be important in that field. The frequency with which a paper is referenced in this way is used to calculate a paper’s impact, see below. But if you see the same paper a few times referenced in papers you read it is a sign that it bears further investigation by you.

When searching for papers you will be able to view the abstract for free but then to get the full text article you often will have to pay for the article if you don’t have access to a subscription from a library. The abstract is an overview of the paper, typically 300-500 words, and will give you a flavour of the content of the paper. Recently there has been a change in the way journals publish papers, as many funding agencies now demand that papers produced from their funding be freely available. You will therefore often see the “free fulltext paper” icon on pubmed searches. In this case the research institution has paid for the article upfront, so you or your library does not have to pay for it.

It also helps to know someone at university, or who works in a library. Students are able to download and print any paper contained in the index, like JStor, that their university subscribes to. Thank you here to all those people who helped me obtain papers, you saved me thousands of dollars.

Peer Review

Most academic journals follow a process of peer review. Once a research group has written up its paper for publication, it will submit its journal of choice. The journal will then select several reviewers from the scientific community. Those selected will have experience of that particular field of study, but will be independent from the researchers. The reviewers identity is kept secret from the research group. The reviewers will then read and pass comment on the paper, giving suggestions and improvements for the paper. These comments are given back to the the researchers and they will then either make adjustments to the paper in line with the reviewers suggestions, or they will perhaps argue there case for how they think the reviewers are incorrect. The editor of the journal will manage this process and when satisfied with the negotiation between reviewer and authors they will accept the paper for publication. This process means that when you read a paper you can have some confidence that the paper has reached an acceptable standard, that the methodology is robust and the conclusions have validity. However, once again, when reading papers you should read with a critical eye and check the results yourself, not simply believing every word of the conclusion.

Impact Factor

Academic journals have a ranking system know as the impact factor. This is measured by looking at the number of times that a journals papers are referenced in other papers. A journal with an high impact factor publishes many papers which have lots of other references. There is the perception that a paper in a high impact journal is of better quality than a paper in a lower quality journal. This is a very general rule with lots of exceptions. It is my view that you should read the paper yourself and make your own mind up on the relative merits of the science.

Types of Studies/Papers

Clinical studies for a new drug or treatment follow the following scheme.

  1. Pre-clinical studies, often testing the treatment in an animal model system.
  2. Phase I clinical trial. This phase includes trials designed to assess the safety, tolerability and the pharmokinetics (how it bounces around in the body, how it interacts with chemicals in the body, half life, LD50, etc.) of a drug.
  3. Phase II trials are performed on larger groups and are assess how well the drug works, in addition to continue Phase I safety assessments on a larger group.
  4. Phase III trails are often large, multi-centre studies aimed at determining the efficacy of the treatment.

Epidemiological studies

Epidemiological studies are studies on human populations, often trying to link cause with effect. eg looking at cancer rates and cigarette smoking or in the case of helminths, a typical epidemiological study would look at the rates of autoimmune disease in populations with and without helminth infection.

Basic Biology/Immunological studies

These studies are often designed to look for underlying biological mechanisms in the normal functioning of the organism and how these mechanisms are perturbed in the disease state. These studies often involve animal models, laboratory cultured cell lines or donated human tissues such as blood and biopsy samples.

Review Papers

These papers are often commissioned by journals. The journal will contact a leading academic in a particular subject and ask them to put together an overview of the latest research in a specialist area. These papers are often a good place to start when approaching a new area, especially for non-scientists as they do not involve descriptions of methodology, statistical arguments or complex mathematics, and are akin to a Reader’s Digest approach to an area of study. They are much more accessible, easy-to-read, and provide a good overview and summary of what can often be a lot of research or study.

Paper Structure

The names I use here are common ones, but others are used, and structure varies, you will also find name and structure varies more between disciplines than within one area of study. Since you are not a peer reviewer you can safely ignore Methods, any statistical discussion, everything really except the Abstract, Introduction and Conclusion or Discussion. That is where you learn why something is being studied, what they hope to find out, and what they did actually find out. You will quickly find that the custom of being conservative in asserting discoveries or observations as actual fact or phenomena to be taken, as is the vocabulary by some, to an absurd extreme. There seems to be some correlation to how interesting a finding is to how conservatively it is expressed. Someone somewhere someday will publish Jasper’s Coefficient of Academic Import. The lower this value is for any paper based on analysis of the author’s certainty of the results, the more important the paper will one day be found to be.

Abstract: This gives a brief overview of the paper as whole in 300-500 words. The idea being that you can read the abstract and decide if the paper is going to give you the information you want.

Introduction: Firstly, the introduction outlines the current state of knowledge in the subject area and then describes the findings that led to the work described in the paper. It lays out the why and the what of the study. In addition, the introduction should also included the background information needed to understand how this new research fits with or contradicts the current paradigm. A good introduction should be well referenced, so that you can then look up previous research.

Methods: In most journals this will follow the introduction, however in some this section is the last one. The idea is to outline the materials used in the experiments and the methodology by which the experiments were carried out. In principle, the methods should be detailed enough to allow other research groups to replicate the work. However, due to space limitations the methods are often compressed, and they often refer back to the authors previous publications. Obviously the methods section is full of detailed technical information, most of which can be bypassed by the general reader. A good idea is to look at the outline of an individual experiment in results section along with the figure legends, and then refer to the methods if you feel you need more information on how the experiment was carried out.

Results: Here the authors will outline the experiments they have conducted and the present the data they have gathered. In good paper the experiments should flow easily from one to another and tell a story as they go. Do look at graphs and tables as there is often more information/conclusions to be draw than the authors will tell you in the text due to space constraints, hence the old adage that a picture tells a thousand words. Don’t forget to check the figure legends for details of the experiment.

Discussion/Conclusion: The authors will here attempt to draw conclusions from the data they have presented and then show how the work advances the current knowledge in the area. They should also acknowledge the limitations of the data and outline what further work needs to be done to solve the problems.

References: As mentioned in searching for papers to read above, the authors will list publications which they have used and referenced in the text. Once you identify a good article, the reference list at the end of the article can be used to find related articles, and then use the reference lists in those articles to find yet even more articles. In this way you can work back in time through the published literature. You can also use pubmed to search forward in time by looking for citation index on your paper of choice. This shows you all of the papers who have referenced your paper of interest in the time since it was published. You can therefore see who has taken this work forward and what the current state of the research area is.

Statistics: When I started I knew nothing about statistics, or how they were calculated or what they meant. I worked to rectify this, spurred on by my inferiority complex, to learn that knowing what they meant added nothing to what I obtained from reading papers. Statistics, like methods, etc., can safely be ignored, unless you intend to reproduce the paper’s results, peer review, etc. Peer review exists so you do not have to read or care about anything but the Introduction and Conclusion, unless you want to.

Bias is something that is remarkably pervasive, and common to see in science, often even in how the question the paper purports to answer is framed. Particularly the kind that we are all prey to, scientists for all they or we might think or wish otherwise, are human. For instance the assumption for hundreds of years that hookworms and whipworms were nasty parasites of no interest or value to humans, and fit only for eradication. Another example is the belief that the appendix was a vestigial organ of no value. In addition, authors are often so intent on testing one idea/hypothesis that they don’t acknowledge that the results are also consistent with an alternative hypothesis, one different from the one they propose.

Bias is very common, sometimes to a comical degree. You may find you want to shout at the authors.

You can often find editorial pieces in journals, or debates raging in the letter pages, on such debates. Letters debating the merits of each others research findings fly back and forth, and get quite snarky at times.

A common bias is for the modern. Many scientists believe that newer data or observations, made or gathered with new equipment and techniques is automatically superior to older examples. Readers often overlook older papers in favour of newer ones.

There are other well documented sources of bias, some inherent to the modern system of science publishing, such as positive results bias.

This is a subject too large to discuss here, but even a casual search using Google will return a great deal of information on research bias.

My Conclusion

I hope you get as much out of reading science as I have. Besides allowing me to restore my health, and that of many hundreds of others, it has become a source of pleasure. I wish I had more time to read it for pleasure, and one day I will. Anthropology, history, biology, physics, metaphysics, there are some extraordinary ideas out there.

While waiting for Horizon (Nova in America) might be perfectly satisfactory, I find I prefer to read the source, and to see the publication bias of television programs as well as the foibles of the program writers.

I also think getting more people to read and understand science is important. A lot of very fundamental questions have been asked and answered, but no one mentioned them to me when it was going on. For instance the choice to go with nuclear reactors capable of producing waste products that could be fashioned into weapons, rather than using alternatives that were built, ran for years, and were decommissioned that no one has heard of, that do not.

What questions are being asked now, and answered while most people remain ignorant that a question ever existed?

Perhaps the most dangerous and damaging bias is the most common. That science is extraordinarily complex and difficult. That it is an area reserved only for the most intelligent or highly educated among us.

I am convinced that bias is actually an enormous factor holding back our ability to enjoy, or even recognise the potential of a great deal of important research.

After all I was not the first to read about the hygiene hypothesis, not the first to realise my allergies and asthma, as well as dozens of other diseases, might benefit from a deliberately obtained helminth infection. Nor the first to think helminths could be easily obtained, worked with, and even sold. I was however the first to act on that knowledge in the way I did.

Heroes and villains are great, for comic books. But not if we choose to abandon whole areas of human endeavour to their realm.

I am certain that there are many, many other examples of powerful research and concepts hidden from most of us, buried in piles of academic research that scientists, because of their inculcation in their profession’s mores, cannot or will not act on or use. I have found a few good commercial prospects for instance, far better in potential and ease of application than helminthic therapy.

I do not have the time or interest right now, but you might.