SD or SE?

18 October 2017

Should I be using the SD or SE? Standard Deviation or Standard Error?

I’m surprised that there are so many poor answers to this question out there, so I’ve decided to write my own, just in case my poor answer helps you any better.

The Standard Deviation

The Standard Deviation is a description of how much spread there is around the mean. You can find the formula all over, so I won’t repeat it here. A small SD indicates that your data points are closely clustered around the mean, and the larger your SD gets the more the data is spread, in a normal distribution. However, there is some intrinsically useful information that the SD carries which may help you decide whether or not you should be using it. If your distribution is normal (and yes, you should have already tested for this), then the SD tells the reader that around two thirds of the data points in your data set fall within one SD of the mean: that is the mean plus 1 SD and the mean minus 1 SD. In addition, the units of SD are practically the same as the units that you have used to collect your data and display your mean. Thus, the SD is providing your reader with a descriptor of the mean that is quantifiable and can be easily interpreted.

You can report the SD in text as follows:

Our data show that males (n = 353) were larger (387.5 ± 49.37 mm) than females (n = 321; 245.4 ± 27.61 mm).

Note that I’ve not used the terms mean or standard deviation in the text above, this is because you should set this up in your material and methods section. If you need to change between the SD and SE in your results, then you will need to indicate which is which when you report it. I have used different levels of accuracy (decimal places) for means and SD. This should be relative to how you actually measured your data. In this case, I measured animals to the nearest mm, so report the mean to one decimal place and the SD to two. Although there is no strict rule, you should not be reporting a mean with far greater accuracy that you actually measured it (e.g. to four decimal places). Note also that the sample size is given for each set. It is very important to provide this information somewhere. There may be instances (such as an experiment) where these numbers are set throughout the document, and so don’t need to be repeated in the text each time you report a result. 

Standard Error

The purpose of the Standard Error is to inform the reader on the likelihood of the mean. In many biological studies, we take the sample from a population that is much larger because we can’t measure all individuals. In our example above, we see that males were larger than females, but how likely is it that the mean we obtained reflects the true mean of the entire population, both sampled and unsampled?

By now, you should have come to the conclusion that the SE is useful when comparing means. In a graph or table where you are interested in demonstrating whether or not the means are different, you can use the SE as error bars around the mean. Note that the convention is that you have plus or minus two SE in error bars (two SE above and two SE below the mean). This is because 2 SE is equivalent to a 95% probability that the mean falls within this range, meaning you’d be very sure. Showing 1 SE either side only shows that you are 68% sure, which doesn’t help much if your test statistic is 0.05. Either way, just make sure that you clearly indicate what you have done in the figure legend. In addition, you will probably want to conduct some statistical test to show that the data are indeed different. Interestingly, statisticians are moving away from many of these tests (or at least the test statistic). If you have used the SE correctly, your graph should speak for itself and there is no real need to carry out a test. In cases where it appears marginal, a test can be useful. Alternatively, you should go back and measure more animals!

In summary

In summary, the SE is telling us about the variability of the population mean (in relation to the one we measured), while the SD is giving information on the variability of the data points that we collected around their mean.

Standard Deviation or Standard Error?

To try to answer the original question posed above. You are most likely to report the SD in your text when describing the data you collected, and the SE on a graph demonstrating how likely it is that the means you obtained represent the entire (only partially sampled) population.

Read on...

If you didn’t understand the above, or want to read more on the (relatively) simple logic of how and why the SD and SE are calculated and derived, I suggest that you look for:

Streiner DL (1996) Maintaining Standards: Differences between the Standard Deviation and Standard Error, and when to use each. Can J Psychiatry 41:498–502.

I am indebted to Don Kramer for pointing me toward the above paper.

  Lab  Writing

A rant for Open Access week

15 October 2017

I put my journal behind a paywall, so why am I talking to you about Open Access?

It is easy to forget who does what in the world of publishing, so I begin by refreshing the minds of older readers, and informing younger ones. I will take the process of submission, peer-review and editing as understood. Although I dismiss it quickly here, this is really the crux of the work involved in science, and what follows is simply the recording of this process.

Academics are not (usually) superstars, nor looking for enormous numbers of readers, but there would be little point to recording our work if we had no readers, or if our work were inaccessible, and so publishing is a necessity. However, we have got into a state in which much of our work is behind a pay wall, and thus inaccessible to most. Whether or not we need our work to look pretty and appealing speaks more to our readers as humans than academics. Perhaps it goes without saying that an audience is likely to be larger the more appealing the presentation, and that’s not just the writing style but the layout and presentation of the text itself. And this is not new.

Right back to the first scientific journal, Philosophical Transactions in 1665, it is clear that the papers were type set and presented in the manor of a book, perhaps analogous to a collection of short stories. At the start, these were reports of studies that were presented at meetings. Producing proceedings of learned societies became the way in which most scientific journals began. Only later did it become possible to submit a manuscript that had not been presented. And later still when publishers began to manufacture their own scholarly journals in the absence of any academic society.

Being the editor of a society journal means being elected by members of that society, and being responsible to an editorial board, normally made up of the society’s members. Until very recently, and I’m thinking back to my first interactions with editors for my first few publications, submitting to the journal meant producing three (or sometimes more) double spaced copies of a manuscript and mailing them to the editor. Editors of bigger journals had secretaries dedicated to handling the administration of the paper. Following a telephonic enquiry, the copies were sent out to referees by post and sent back to the editorial office with a typed report often together with the marked up manuscript. Once the editor had received all reports, they communicated their decision back to the corresponding author (i.e. the one to whom correspondence was addressed) and, once accepted, the article went into production. Prior to personal computers being commonplace (only 25 years ago), each journal would have had to have had a publisher to set the type and print the pages. Clearly, this was beyond the scope of individual societies and the publisher was a necessity. Libraries had to pay for copies of journals, as the cost of publishing had to be offset by the society.

The advent of desktop publishing changed the need for publishers and brought many small society journals onto a larger platform where they could produce attractive content (over the typewritten documents that had been stencil duplicated or similar) and sent around to members. However, for small societies there was no professionalism involved as it devolved to the editor to publish society material. This is where I entered the stage in 2009 when I took over as editor of the African Journal of Herpetology. Thankfully, email had taken the place of the postal service, but once a manuscript was accepted I was the one who needed to type-set the documents (in Quark) and send out proofs to authors. Once proofs were corrected and the issue was ready, I had to find quotes from 3 printers, deliver discs and ultimately collect boxes of printed journals. Back at home, I also packed all of the copies into labelled envelopes (with some help from friends) and carried the boxes once again to the post office.

  

After the first issue, I realised that I could not do all of this work indefinitely. I knew that there were publishers who were interested in acquiring the journal into their stable, and I contacted them and started negotiations. In 2011, the first copy of AJH from a professional publisher emerged, and allowed me to go back to editing the content through peer review.

At the time, I was aware of Open Access and considered this as an option for the journal. Open Access would have required that someone pay for the type setting, and the society would still have to pay for an online dissemination platform, given that they did not have their own platform. This would have meant that authors paid for getting their manuscripts into print. And then, like today, the decision was that our authors would not be willing to pay. Other, richer, societies were able to go Open Access with the costs being covered by their members. For one, Salamandra, this became an incredibly successful model, with submissions increasing as well as their Impact Factor. But it did become too much, and eventually a charge is now levied to the authors (albeit small). However, they (and other societies) demonstrate that Open Access is possible on an independent platform.

Why don’t all society journals do without publishers, and go Open Access independently?

The first problem is that societies generate income from journals. Subscribers to print or virtual copies pay the society, and this can defray a large part of the cost of publishing otherwise carried the members. Going Open Access means losing this revenue, as well as taking on the extra costs.

The second problem for many societies is that their members are paying, and the councils or committee’s elected to represent the members do not feel that it is fair for their members to pay for open access for everyone else. Part of the privilege of being a member of the society is having a free (or more accurately paid through membership) copy of the society’s journals. The costs are not high, and the exclusivity of members having Open Access while it is denied to others is perhaps just a hang over from the days when the only other copies were in the library. Certainly, the costs are nothing like those which authors are now charged by publishers to turn their accepted manuscripts Open Access.

In my opinion, the current situation is truly crazy. Tax-payers (in the main) pay for science to be conducted at universities and other institutes. In places where Open Access is mandatory, the tax-payer pays again to have the research published at a cost that is far inflated from the actual production costs. Publishers are getting fat, and the losers are the scientists (whose funding is reduced to pay publication costs) and the tax-payers who end up feeding the greedy publishers.

What will it take to break the vicious cycle?

We need new models for publishing. Society journals are still kings in this game and ultimately hold the cards for moving away from filling the pockets of publishing companies. What we have seen in recent years is that journals can come from nowhere to become dominant players in the system. Think PLoS ONE, and the even more recent Scientific Reports. These mega Open Access journals didn’t exist 10 years ago. And they don’t need to exist 10 years from now. What is needed is for governments to fund societies for the actual costs of publishing their own content. This could cover type-setting and the additional IT infrastructure (on-line submission system and online dissemination platform). Most (if not all) societies are not-for-profit organisations, and only need to cover the costs of publishing.

Why would a government in one country pay for the costs of researchers from another publishing in their sponsored journal?

Soft power. It is not in the interests of governments to have any scientific content behind a pay wall. The relatively small costs of grants to societies, together with some simple accounting for papers published, would be tiny compared to the current costs that many are handing over to publishers. Have your country’s logo stamped all over the pages if you must, but it really wouldn’t cost what you are currently paying. Think of the soft power involved in diseminating peer reviewed independent science.

Clearly, scientists in some countries might be worried about handing their publications to their government to disseminate. For this reason, perhaps governments might not be the correct long-term solution for many scoiety journals. Instead, there is potential to obtain grants from philanthropic societies: The Open Society Foundations might be one such source.

But wouldn’t that put all editors back in the position that I desperately wanted to get out of in 2009?

Yes, clearly there needs to be a not-for-profit clearing house that societies can use to pass on their manuscripts. Such places exist. Most governments have their own governmental publishers. Although it would be understood that academic freedom might fly in the face of some governments being held responsible for publishing content. Many countries have institutes with publishing wings. We could even use commercial publishing houses, if we could only curb their greediness.

Alternatively, we could get more savvy as authors and start using La-Tex software to compose or at least submit manuscripts. Most of us have become too dependent on word processors, and we could step up to submitting in La-Tex. Submission software could allow us to cut and paste our content into La-Tex editors online. It’s all possible. I think that many academics would be happy to submit via La-Tex if this meant that their content went Open Access.

Can we afford not to change?

If you are from a rich country or institution, then you can probably afford the current system. Those who cannot are researchers in disadvantaged countries. In some cases, the cost of publishing Open Access is greater than the cost of the research. These are insurmountable costs for many researchers. We have a massive hole in scientific contributions from the poorest of nations, and the current Open Access models will see their work being the most hidden from view, while the countries paying for their work do so disproportionately. But even developing countries could be winners in a new Open Access model. By sourcing the relevant IT skills in country, governments of middle-income countries could facilitate the content of their own society’s with relative comfort. In my opinion, everyone should publish work from scientists in the poorest countries as Open Access without any charges.

What is needed for this change?

  1. Societies need money. Editors can’t be publishers.
  2. We need free software. We desperately need good, free editorial management software. There are some free versions out there, but what we need are free versions that are at least as good, if not better, than existing platforms (e.g. ScholarOne; Editorial Manager).
  3. We need a free La-Tex interface with robust templates for all societal journals. Ideally, this would be packaged with the above editorial management software. This must have the ability to cope with figures and equations, and the unusual demands that some society journals have.
  4. We need a solution for hosting and disseminating the Open Access society journals (and their supplementary information if not hosted elsewhere) in perpetuity. This last point is perhaps the most expensive, and almost certainly requires government assistance. Maybe this is an interesting use of the block-chain with libraries keeping the data. It would be an interesting way to build a doi with editor and referee unique IDs, and the document's information hanging off. 
  5. We need to take back our content stuck behind pay walls. Yes, it’s time for you to all to dig up those old submitted manuscripts and submit them to an institutional repository where it can be accessed for free.

  Lab  Writing

No paper should be a puzzle

28 September 2017

Did you ever feel that reading a paper was like trying to solve a puzzle? I did.

During my PhD I read a lot. My study species was the subject of tens of thousands of papers, and I was convinced that I'd find what I'd need somewhere in those musty old reprints and heavy volumes that came from the library shelves. It was easy to believe that the authors of those papers had set great puzzles for me to try to understand their content. The realisation that this is not the best form of communication came as a personal epiphany some years ago.

“The goal of good writing is straightforward: to make your reader's job as easy as possible” Kevin W. Plaxco (2010).

This great quote from a paper by Plaxco (2010) cuts to the heart of why it’s essential to write well. As an editor, I once experienced an author who thought the opposite, and for me this was an epiphany. The submitted manuscript was an impenetrable mess. I could tell that there was good work in there, but as an editor I felt that the information had been made so obscure by the authors that my readers were unlikely to get much out of it. The author was a colleague, and so I decided to phone him and chat about the need for much greater clarity in his manuscript. Experience with email has taught me that they can often be taken in the wrong way; usually in the worst possible way.

The response I got surprised me. The author recognised that his text was dense and was unapologetic. “Let the reader work to make sense of my data,” he said. My epiphany came not because this was a totally alien concept, but because during my PhD days this was exactly how I had thought scientific writing should be done. I had spent so many hours slogging through dense and dreary papers by well-respected figures. The reward when I finally understood what it was that they were trying to communicate felt so great that I believed the puzzle they had set me was what I should set my own readers. Happily, my advisor didn’t feel the same way, but for years I continued to believe that a paper should be a decent puzzle for my readers to crack.

There are still authors out there who attempt to set puzzles for their readers, but they aren’t in the mainstream any more. Instead, biological sciences has some inspiring writers, and many more are taught how to inspire future generations towards communicating great science. That’s not to say that there are no puzzles left. Much of what we do requires great puzzle solving abilities. However, let’s keep the puzzles away from communicating with our audience, make their job as reader easy or even pleasurable, and they’ll keep coming back for more.

Plaxco, K.W., 2010. The art of writing science. Protein Science19, 2261-2266.

  Lab  Writing

What's in a name?

23 September 2017

Species names (Latin or common), taxonomic authorities and acceptable conventions

There are quite a few conventions that it’s worth knowing about when it comes to using names. When talking about common names, it is a bit of a quagmire as there are few standards that are followed by all journals. Scientific names can also be in great flux, but at least there is definitive help. Here I will outline my interpretation of what’s what in using names.

Scientific names

There is only one valid scientific name for a species. This is expressed in a binomial: two names, first the genus (with an initial capital letter) and then the species name (lower case). This name is italicised by convention (or in some cases underlined). The single name is a hard rule as taxonomy can’t survive in any other way. The most recently published version is the one to use, but happily with both reptiles (http://www.reptile-database.org/) and amphibians (http://research.amnh.org/vz/herpetology/amphibia/) there are well curated databases to which you can refer for the latest taxonomic treatments. My suggestion is that you defer to these (with acknowledgement) and you won’t go far wrong. If you do this early on in your thesis, you may need to update it later.

Consider Mac’s example:

When Mac started his PhD in 2013 he worked on three species, and over time these changed so that at various points in his thesis he had the following combinations:

 
 
 
Amietophrynus rangeri
Amietia angolensis
Xenopus laevis
Amietophrynus rangeri
Amietia quecketti
Xenopus laevis
Sclerophrys capensis
Amietia delalandii
Xenopus laevis

The taxonomy of two species changed such that he had to constantly revise the names in his thesis. Luckily, he could simply conduct a “replace all”, but it did become quite confusing. The lesson is that if you want to avoid this, work on Xenopus laevis or conduct your PhD in as short a time as possible!

In taxonomic papers, and certain journals with a taxonomic focus, there is an insistence that you cite the “taxonomic authority” after the species name when used for the first time. This is essentially a citation in which you acknowledge the original description that accompanies the name. Let’s take Mac’s examples again:

Sclerophrys capensis Tschudi, 1838

Amietia delalandii (Duméril and Bibron, 1841)

Xenopus laevis (Daudin, 1802)

It might seem that there’s a mistake above, because Tschudi, 1838 isn’t in brackets. This is actually deliberate. Tschudi, 1838 isn’t in brackets because the genus name hasn’t changed since the description by Tschudi placed S. capensis in Sclerophrys  in 1838. But it is clear in the table above that the genus name did change to Amietophrynus, and it was recognised under a different species name (A. rangeri). Unlikely as it may seem, the original description of the species was by Tschudi in 1838, but this was long forgotten (or ignored because the type specimen was so poorly preserved), and the species was regularly referred to as Bufo rangeri. When the genus Bufo was broken up into lots of newly named genera by Frost et al 2006, Tschudi’s description was once again disregarded. However, in 2016 new work on the type specimen revealed that it was the same as Amietophrynus rangeri, but because Tschudi’s description was older, the taxonomy deferred to the rule of precedence. If you want to read more about it, you can look at all the changes on the AMNH database.

To clarify, the use of brackets around the taxonomic authority only happens when the species is no longer in the genus in which it was originally placed by the person who described the species. Daudin (1802) originally described Bufo laevis, and Duméril and Bibron (1841) originally described Rana delalandii.

The generic name of a Latin binomial is often abbreviated to the first letter (e.g. X. laevis). This should only be done once it is well established in your writing what the full name is. Or, in other words, you should refer to the species in full on the first mention, and then switch to the abbreviation. Because sentences shouldn’t start with an abbreviation, you might need to relax this from time to time. You also need to treat the abstract as a stand-alone document in this respect. You might find that you have two species, which belong to different genera that share the same first letter. In this case, most authors use a different abbreviation: for example Ap. rangeri and At. angolensis. Keen readers would have noticed that Mac doesn’t have this problem any more!

I have seen some comments that suggest that once a generic name is abbreviated in a paper, it shouldn’t be written in full again. In my opinion, this is rather a silly idea. It is up to you (the writer) to avoid ambiguity. If there is no ambiguity, then you may write the species name in full to increase the readability of your piece.

Lastly, there is a convention that when referring to the generic name alone you should always mention that it is a genus. So you shouldn't write about those Bufo which hop around the house, but instead about those members of the genus Bufo that hop around the house.

Common names

Common names have a different set of problems to those that we’ve seen for scientific names. These are largely focussed on: what name to use, and whether or not to capitalise the common name.

  1. What name to use - This is tricky. My suggestion would be to take the most commonly used common name, as that’s likely to be the one that most people will recognise. However, it’s not always that simple and because there’s no “correct” common name, you have some leeway to choose the one you prefer.

Let’s take an example: The African Clawed Frog is a very widely used name for Xenopus laevis which is used all over the world as the model amphibian. However, African Clawed Frogs is also the commonly accepted name for the entire genus, Xenopus. Thus, shouldn’t we avoid using this for the species X. laevis? So, let’s look at the alternatives. Frost (2017) lists 13 common names for this species. Names such as Platanna and Common Platanna are difficult to use in scientific papers as they are only regularly used in South Africa. Clawed Toad is not appropriate, as it is not a toad, a confusion that dates right back to Daudin’s 1802 description. Clawed Frog has the same problem as African Clawed Frog, and other variants such as Upland Clawed Frog and Smooth Clawed Frog also seem inappropriate as this species is neither confined to uplands or the only member of the genus which is smooth (although ‘laevis’, the species name, does mean ‘smooth’).

There has been a recent movement to formulate common names into genera (or even families) such that the specific common name takes on a specific epiphet of the generic or familial common name. In a world of stable taxonomy (such as that enjoyed by ornithologists), I can understand that this is possible. But imagine the requirements in common name changes for Mac in his thesis over only a few years. Surely the useful thing about common names is that they are used commonly by normal people who don't worry about taxonomy? If we insist on common names constantly changing, don't we risk alienating the public?

I eventually decided to use ‘African Clawed Frog’ as this is the most widely recognised common name for this species. It does become an issue when writing about X. laevis and other Xenopus species (as we have), but we’ve managed to resolve this without too much difficulty.

  1. Should common names have capitals?

View 1. Common names are proper nouns and proper nouns start with capitals.  The Western Leopard Toad is not any toad, and the African Clawed Frog isn’t any frog.  The Cape platanna is any platanna living in the Cape, but the Cape Platanna is Xenopus gilli! In this last example, you can see how capitalisation reduces ambiguity and differentiates between adjectives and nouns.

View 2. Common names are not formal names (like scientific names) so why does it matter? Many journals do not capitalise common names and insist on lower case unless they are proper nouns, like African in African clawed frogs. This seems to have arisen from editors who dislike superfluous capitalisation.

On balance, I’d say that it’s likely that your usage will be dictated by the policies of the journal.

  Lab  Writing

Why do I need to cite?

20 September 2017

Why do I need to cite? Standing on the Shoulders of Giants...

Have you ever wondered what the blurb on the front of Google Scholar means? Who is standing on whose shoulders?

Essentially, it's a recognition that all research is built upon research that has gone before it, and this is the basis for citations in the text of scientific papers. Patrick Dunleavy  argues that citations are required to meet 7 criteria with respect to academic writing.

Interestingly, although the phrase is often attributed to Isaac Newton, it turns out that Newton got it elsewhere:

If we dwell a moment longer, we can look back at the World's first scientific journal: Philosophical Transactions 

You can download the first of those papers of 1665 and note that there are no citations (other than to books or letters) because there were no previously published articles from which to draw. However, even then authors noted that ideas came from previous authors and we can recognise that acknowledgement back to Aristotle.

Research is built on existing work and ideas

It would therefore be very unlikely that your idea/ideas has no basis in existing literature. If you can’t find it, the chances are that you haven’t looked in the right way. Try Google Scholar, Scopus, Web of Science and vary the search terms or try searching for articles citing something similar. There are lots of ways to search, but maybe that's another blog post for another day.

Citations demonstrate to readers where your ideas have come from. Citations can also be used to reduce what you need to write – especially with respect to methodology. If you (or others) have already provided the methodology in full then you can give a much simpler description and the citation.

Where to cite?

Automatically, I'd suggest that you cite whenever you write. Using citations becomes habit forming, and you’ll end up wanting to use citations everywhere. Popular scientific writings tend not to have citations, but they can still be there but subtly different. Have a look at this example from The Conversation.

Within a manuscript for a scientific journal or in your thesis, you can expect that your introduction and discussion sections are going to be full of citations. The methodology may also have citations, and these can often save you needing to write a lot of sections which would otherwise be very detailed (see above).

Where within a sentence should the citation come?

There are a number of different styles, and this is likely to depend on the journal that you are writing for. The standard way is to use the name and year in parentheses at the end of the statement to which the citation is relevant.

The impact of all invasive amphibians is similar to that of invasive birds and mammals (Measey et al. 2016).

You’ll see that sometimes the names are brought into the sentence and become the central agents of the text. The phrase et alii (often abbreviated to et al.), literally means "and others" referring to the other authors (typically used when there are more than two - but check the journal format!). 

Sometimes, instead of "et al." you can write "and colleagues" or "and others". This is something to do occasionally when you are looking to diversify some text. Don't over do it though.

Measey et al. (2016) found that the impact of invasive amphibians is comparable to that of birds and mammals.

This technique is very useful when you then want to add another sentence or two about this same study.

Measey et al. (2016) found that the impact of invasive amphibians is comparable to that of birds and mammals. They did this by constructing GISS scores for all individuals in all groups. 

Because the authors are the subjects of the first sentence, the citation becomes implicit in the second sentence.  Then you don’t need to use the same citation again within the paragraph.

What about page numbers?

Sometimes you'll see a citation with a colon and page number after. This really only needs to be used if you are quoting specific text on a particular page:

Measey et al. (2016: 976) proposed that using GISS scores could show that "some amphibians can have devastating impacts to the environment". 

What about citations as taxonomic authorities?

Taxonomists have special rules for this, and I'll do it in another blog post. These are not the same as regular citations (because they don't appear in the literature cited and you don't have to have read the descriptions), and only some journals ask for them. 

Is it possible to mis-cite?

Yes. One of the most common ways in which students mis-cite a paper is to use statements made in the introduction (or discussion) which were not the subject of the study. For example, in the introduction of their paper, Measey et al. (2016) make comments on amphibian decline.

Amphibian populations are currently declining across the globe (Wake & Vredenburg, 2008; Collins et al., 2009; Pimm et al., 2014) and alien amphibians are at least partially driving these declines through competition (Kupferberg, 1997), hybridization (Dufresnes et al., 2015) and introduction of novel pathogens (Berger et al., 1999; Daszak et al., 2003; La Marca et al., 2005; Martel et al., 2013).

However, it would be wrong to give a statement on amphibian decline and cite Measey et al. (2016). They did not study amphibian decline. Instead, you should read the papers that they cite (Wake & Vredenburg, 2008; Collins et al., 2009; Pimm et al., 2014), and read around those to find studies on amphibian decline that are appropriate for your context. This underlines one important aspect of choosing citations where the statement that you make relates directly to the study carried out in the citation.

Another common mistake is to forget which paper has which information. You can try to make sure that you don’t do this by taking better notes or a more accurate plan.

Should I cite without reading the paper?

No. When you are citing a study, you should be sufficiently familiar with the publication that you are endorsing the study in relation with the statement that you make (but see below). If you are not convinced by the nature of the study that you are tempted to cite, then rather don’t cite it and use another one. If you can't get hold of the paper, this is another reason why you might not cite it. This is a regular reason given for why Open Access journals attract more citations. 

What should I not cite?

This does depend on the journal you are writing for. Some journals don’t permit citations to unpublished data or web sites. My suggestion would be to avoid such sources anyway (with particular exceptions – see below), unless it is really important that you include it. Other examples of texts to avoid are: text books (use the original paper instead), newpapers or magazines, predatory journals (or any non-peer reviewed article).

One exception to the not citing websites relevant to our work is the IUCN Red List. Note that all entries on this site now have DOIs, and this might be a good guide for what is available to cite. The DOI is very useful as it means that there is a consistent record of that version. Otherwise, you could cite any website (even this one) and then the owner can go and change the site and it no longer says what you thought it did. The DOI removes this problem there will always be an archived version with that particular DOI.

Do I cite the review or the primary literature?

This depends on the space you have and whether the review contains all the information you need to cite. It’s preferable to use primary literature, but sometimes reviews (or meta-analyses) are actually more expedient to use, especially if they are not the focus of your study. You can even cite both when relevant.

How many citations is enough?

Some journals have a word limit, or even a limit to the number of references that they allow. Others do not, and you should probably use what is recently published as a guide to what is acceptable. For the chapter of a thesis, you should err on the upper end; from 50 to 100 references. Note that citations may well be more as you may cite a paper more than once.

Obviously, everything you cite needs to be in the Reference (or Literature Cited) section, and you may well need to spend time deleting extra stuff. You can get around this common issue by using a citation manager like Mendeley or EndNote. I’d always suggest that you use one of these tools as they can really help with your reading too. These days they are busy turning into a kind of scientific social network. They regularly make suggestions of what you could be reading based on what you read. This can be useful.

If you are looking for multiple examples of a statement that you’ve made and there are many possible, I’d suggest that you aim to produce three. Make sure that you use a suffix (like “e.g.”) to show that you are aware that these are examples of a widely reported phenomenon. You can choose these as you like, but may want to consider using what you consider to be ‘the best’ examples, and/or references that you are planning to use elsewhere in your paper/chapter. This can drive the total number down considerably, and helps to keep the citations more relevant to your work.

Should I cite myself?

If you are publishing relevant and appropriate papers, then there is no reason not to cite yourself. In many cases (such as with your thesis work), your own publications are likely to be more relevant to methodology and subject matter than much of the other work that is out there. However, if you’ve previously published on termite fungus and now you’re publishing on frog toes, it’s unlikely to be relevant.

I would avoid citing your thesis if possible, rather put it all into papers. There are times though when citing your thesis is unavoidable. Within your thesis, I would suggest that you do provide citations to different chapters as this will help the examiners see how the chapters relate to each other.

Take a look at this article on self-citation. They claim that self-citations in the Natural Sciences run at 33% (Centre for Science and Technology Studies, 2007). 

Should I cite my friends?

It may be easier to cite your friends if you already know their work well. You may have heard them talk and know that the subject is relevant. They may be encouraging you to cite them, but should you?

In these days of scientometrics, we do need to acknowledge that citations act as a kind of currency. They count towards your H-index and this can reflect on your prospects as a post-doc or employee. What’s also clear is that cited papers get cited more, so it could really help your friends if you cite them. Obviously, the inverse is also true, so beware of the politics of citing. However, the most important points have already been raised (above). The study must be relevant and appropriate before it gets included as a citation in your work.

Should I cite my advisor?

Yes!

Does the impact factor of the cited article matter?

Papers in journals with high impact factors are more likely to be cited because their contents are already thought to be of interest to a wide range of people. That is something that the editors of that journal will have considered even before sending the manuscript to review. Sometimes (but not always) the impact factor of the journal can be an indication of the quality of the study. But you should judge this for yourself when you critically read the paper.

I find that the first paragraph of a paper is more likely to contain citations of higher impact journals. This is in part as these are likely to be more cross cutting (as is often the case for the first paragraph). In the end, if you need to make a choice, choose the paper that is most relevant, irrespective of impact factor.

  Lab  Writing