As scientists, our main objective is not just to do well motivated and novel research, but also to publish that research in a timely fashion, and in a way that clearly communicates to an “educated layperson“ what is new and interesting about your research, while also providing enough detail in the paper such that an expert in your field can assess the soundness of your work. Remember that, first and foremost, your work needs to be novel with a well posed research question, and must be an interesting and well motivated addition to the body of published literature on the topic. If you don’t have the skills to produce a well written paper, even with a well posed research question you will have a difficult time getting your work published, and/or you will needlessly get bogged down in review.
- Learning to write (and read) a scientific paper
- Tips for making your paper easy for a reviewer to review
- Sections required in a scientific paper, and what they should contain.
- Developing skills to efficiently and quickly write up your work
- Who to include as co-authors
- Your responsibilities as a author/co-author
- Submitting your paper to a journal
- Responding to reviewer questions
- Rejection and Acceptance
Learning to write (and read) scientific papers
Learning to write (and read) scientific papers is a process that definitely requires skills, but unfortunately those skills are rarely taught, even though for the most part they are actually pretty well defined. Here is a tongue in cheek paper on how to write scientific papers (that unfortunately contains a lot of truth). On a more rigorous note, in 2014, Lacum et al published a study where they trained students to look for seven key elements when reading or writing papers. As we will see below when I describe what sections need to be in a scientific paper, these elements are integral in the sections of a paper:
- Motive: Statement indicating why the research was done (e.g., a gap in knowledge, contradictory results). The motive leads to the objective. The motive should appear in the Abstract and Introduction.
- Objective: Statement about what the authors want to know. The objective may be formulated as a research question, a research aim, or a hypothesis that needs to be tested. The objective should appear in the Abstract and Introduction.
- Main conclusion: Statement about the main outcome of the research. The main conclusion is closely connected to the objective. It answers the research question, it says whether the research aim was achieved, or it states whether the hypothesis was supported by evidence. The main conclusion will lead to an implication. The main conclusion is often the last sentence in the Abstract, and is of course also described in the Discussion and Summary.
- Implication: Statements indicating the consequences of the research. This can be a recommendation, a statement about the applicability of the results (in the scientific community or society), or a suggestion for future research. This may appear in the Abstract, and certainly appears in the Discussion and Summary.
- Support: The statements the authors use to justify their main conclusion. These statements can be based on their own data (or their interpretation) or can be statements from the literature (references).
- Counterargument: Statements that weaken or discredit the main conclusion. For example, possible methodological flaws, anomalous data, results that contradict previous studies, or alternative explanations. Counterarguments are sometimes presented as limitations. They are placed in the Discussion and Summary.
- Refutation: Statements that weaken or refute a counter-argument. Refutation appears in the Discussion and Summary
Making your paper easy for a reviewer to review
It is very important to always try to be aware of how easy your paper is for a reviewer to read… do you define all your variables, model parameters, etc early on in the paper in a nice tidy table or two that the reviewer can flip back to should s/he forget the parameter definition while reading the paper? Remember, you’ve probably been up to your armpits in that analysis for months, and memorized long ago what your parameter definitions are… but the reviewer is seeing them for the first time. Be kind to the reviewer, and give them reminders here and there throughout the paper as to what the pertinent variables mean.
Also, help the reviewer by adding a sentence at the end or beginning of most sections of the paper, to let them know what you’ll be describing next. I review a lot of papers in many different subject areas, and all too frequently I find it annoying when I’m wading through 40+ pages of text, and I’m in the middle of some lllooonnnnggg section where the authors appear to be rambling on, and they haven’t given me any indication either at the beginning of the section (or just before it) what the point of the section is supposed to be. That kind of thing makes reviewers cranky, and much less likely to appreciate what might be new and novel about your work, and much more likely to recommend rejection of the paper because they find it unreadable.
In a well-written paper, the reviewer should be able to read your Abstract and Introduction, and flip to the tables and figures and be able to understand them. This means putting descriptive stand-alone captions on tables and figures, and properly labelling the figures.
Also, help the reviewer out by always stating in plain words what can be stated in plain words. Again, because I do a lot of cross-disciplinary research as a statistician, I read papers from a lot of different fields. Each field has its own jargon, and if you are “brought up” in that field as a young graduate student, you can easily be trained into forgetting that there are often simple plain-English ways to say what your field might use a lot of jargon or inscrutable acronyms to say… jargon that is meaningless and/or tedious to understand to a well educated scientific layperson outside of your field. And you may very well get a person outside your field as a reviewer if you are, for instance, fitting the parameters of a mathematical model to data.
Basic elements of scientific papers
The basic elements of virtually any scientific paper are as follows (Introduction, Methods and Materials, Results, and Discussion and Summary):
- Introduction: This section always appears in a paper. At the beginning of the Introduction is where you motivate your work (ie; why should anyone care?). Start from a broad motivation, and move to focus in on the particular motivation of your work. For instance, let’s assume I was writing an Ebola paper that describes a compartmental modelling analysis I did to assess the effects of isolation and/or quarantine on the spread of the disease:
- At the very beginning of the introduction I’d start off talking about the number of people killed in past outbreaks, and the wide geographic spread of outbreak locations, and the ever present risk that Ebola cases can be imported to other areas of the world due to modern air travel (the point being that no matter where you live, you should care about Ebola). Then I’d talk about the high mortality of the disease. These would be among the first few sentences in any paper written about Ebola.
- For my particular paper, I’d then mention that the lack of current treatment options (like vaccines or medications) leave better hygiene, quarantine, and isolation as the only options available to slow the spread of disease.
After motivating the project, you then move on to describing the objective of the paper. This is where you present your research question. And where you give a very short overview of what you did in your analysis, and how it advances the body of work in the published literature on the subject. In the Ebola paper case, I would add some sentences saying that mathematical models are being increasingly used to assess the efficacy of disease intervention strategies (and I would cite a few well known seminal publications on that topic). Then I would state that in this work we use a mathematical model to assess the efficacy of isolation and quarantine, and I would state that no one has ever done that before for Ebola (as of August 2014, this was true). In this part of the paper it is very important to state what is new and novel about your work.
Once you have described your motivation and objective, it is a good idea to end the Introduction with a sentence or two that gives a road map for what the reader should expect in the following sections. Something like “In the following section, we will describe the data sources and mathematical and statistical modelling methodologies used in these studies, followed by a presentation of results and discussion” (this is assuming your analysis uses data, a mathematical model, and statistical methods).
- Methods and Materials: This section always appears in a paper. if you are using data in your analysis, the first subsection in this section should be Data. The Data section should thoroughly describe your sources of data. If you collected it, what were your laboratory or field protocols? If it is time series data, what time steps are used? What, precisely, is the data measuring? If you got the data online, give a reference to the source. Even if you didn’t collect the data, you need to describe the collection procedures of the person or group who did collect the data.
If you are using a mathematical or computational model, the next subsection should be Model. In this subsection, you will describe what kind of model you are using, and give citations to relevant related publications in the field. You will describe what is new and interesting about your model (if relevant… sometimes it is the data that are new and interesting, and what is novel is applying an old model to new data). Here you will give the model equations and compartmental flow diagram (if using a compartmental model), or other details about your mathematical or computational model. You need to give enough details that anyone could reproduce your work based on this information.If you are using statistical methods that are fancier than your usual statistical tests based on Student T, Z scores, Spearman rho, etc etc, you need to have a subsection under Methods and Materials called Statistical Methods. This subsection would be appropriate, for instance, if the statistical methods you use are so esoteric that they are either new, or very rarely used in your field.
- Results: This section always appears in a paper. Here is where, without discussion, you give the results of your paper, often in tables and figures, and accompanied text. Do not discuss the results here!
- Discussion: This section always appears in a paper. Never put results that you haven’t discussed in the Results section here…. they should be in the Results section! In the Discussion section, you talk about notable things revealed by your results and how this fits in with (or contradicts) the published literature.
- Summary: This section is sometimes called Conclusions, and sometimes is lumped in with Discussion (and called Discussion and Summary). It depends on the journal. If there is a separate Summary section, you start off with a little paragraph describing what you presented in the paper, and why it is new and novel. In the summary you detail limitations of your study, possible future work, etc, and usually end with a “feel good” sentence about the utility of studies like yours.
In papers I’ve reviewed for which I’ve recommended rejection, the authors often fail to motivate their work, present a proper research question, have many mis-spellings or grammar mistakes, don’t describe their sources of data properly, and/or present a lot of irrelevant chaff of low-information plots or text. The first two items are pretty much guaranteed to cause your work to be rejected, whereas the last three really don’t help your case. It is a simple matter to, at the very least, run your work through a spell checker, so never submit a paper without spell checking. If English is your second language, have an anglophone proofread your paper before you submit it.
If you are fairly new to publishing have an educated friend of yours, outside your immediate field, read the paper and comment on it. Pay attention to what they have to say about your paper, and make improvements accordingly.
Developing skills to quickly and efficiently write up your work
When I embark on a new research topic, I begin background reading to determine what the state of the literature is on the topic. This is usually an ongoing process that can take days/weeks/months. At the very beginning I start a bibtex .bib file where I not only have the bibtex entry for all relevant papers (even if I think I might end up not citing them), but also comments with the abstract of the papers, and my notes about the salient points made in the paper that I think are relevant to my research topic.
I also create a special directory for each and every analysis I work on, that contains all files related to the analysis. In that directory I have a file called README where I put the URL’s of web pages I’ve found that are relevant to my topic, and notes on what the R/Matlab/C++/etc programs I’ve written in the directory do to perform the analysis.
I create a subdirectory under this directory called tex, and in this directory, very early on in the analysis, I start the skeleton of a paper. In the latex file for the paper, I put the section headings Abstract, Introduction, Methods and Materials (with appropriate sub sections like Data, Mathematical Model, Statistical Methods, etc), Results, Discussion and Summary.
I then write the Introduction for the paper. I am able to do this because I never embark on an analysis without a clear idea of the motivation, and a clear idea of what my research question is. With these two things in hand, I can write the Introduction. If you cannot write most of the Introduction of your paper very early on, you seriously need to rethink what you are doing, and why you are doing it!
Next, early in the analysis, I fill in the Methods and Materials section. I know what the sources of data are (if I’m writing a paper that involves data), and I know what my modelling methods are. This section should be easy to write.
In the Results section, I put a few sentences describing what tables and figures I think would best summarize the results I expect to get. That way, when I get into work every morning, I can look at that bulleted list, and work towards getting one or more of the figures and/or tables prepared. It makes my days much more efficient.
In the Discussion section, I will fill in some sentences describing the previous literature on the topic. Once I’ve finished the analysis, and do the final writeup of the paper, having those sentences there makes things go faster.
Early on in the analysis, you should have a clear idea of what journals you might want to submit to. If you are junior, and unsure about what journals might be appropriate, ask senior people for advice. Go the journal websites, and look at the “Instructions for Authors”. There you will find information about acceptable formats (often journals will have latex style files and templates you can use), article word counts, etc. Keep this in mind as you write, because it will save a lot of time later. Also be aware of what the tone of the journal is; some are more mathematical than others, in which case a thorough description of the model in the main paper is fine. For other, more data oriented, journals, you may have to consider putting the detailed mathematics in a supplement or appendix.
I cannot stress enough that you need to thoroughly document your analysis as you go along. Once you submit it to a journal, it may take months to hear from the reviewers. In the meantime, you’ve been working on something different, and likely have forgotten what all your Matlab/R/C++/etc code did. And if the reviewers ask for cross-checks or extensions to your analysis, you will be in real trouble if you can’t immediately replicate what you already did. So, keep the README file up to date, and extensively comment your code, and use good programming practices. Also, back up your work regularly, to ensure that a disk crash doesn’t doom your analysis to a virtual garbage heap, never to be published. By “regularly”, I mean every day. You can buy external hard disks for your laptop, usually for less than $100. If you don’t regularly back up your work, the day will come when a disk crash will cause you to lose several months of work. I’ve seen that happen several times now.
Who to include on your author list
Obviously, anyone who has contributed significantly to the analysis or manuscript should be included in the author list. But how much does someone have to contribute to be on the author list? Well, it turns out different people have very different standards for that. My recommendation is to always err on the side of being generous in offers of co-authorship. I’ve unfortunately had the experience of being excluded from papers where I helped people by writing code to do their modelling and/or code to do their statistical data analysis. Even if it took me just a few days, it was significant outlay of time on my part (I am always very busy on many different fronts), and having that contribution go unrecognized doesn’t make me feel kindly towards helping the people again, especially when my time saved the researchers weeks or months of work because they didn’t have the requisite skills to do the task themselves. I’ve had several colleagues complain about exactly this problem; some people get very bitter about such things, and the last thing you need as a young researcher is to make enemies. So… err on the side of generosity (and I’m not talking about when you work with me ). It costs you nothing. Plus, it has been shown that your ability to network, as evidenced by the number and variety of people with which you co-author, is a strong indicator of your future publishing success.
In my experience, junior researchers seem to be the most likely to not acknowledge help with their research, even though they have the most to lose by undervaluing the people who help them. If you ask someone to do something for your analysis that you don’t know how to do yourself, you need to either pay the person as a consultant, or offer to include them as as a co-author.
Duties as an author/co-author
As a co-author, it is of course your duty to contribute to the analysis in some way (but keep in mind that few co-authors contribute to all aspects of a particular analysis), and to thoroughly read the paper before submission. Note that if you are working on a similar analysis with another group of people, you need to ensure that all your co-authors on both papers are aware of your potential conflict of interest, right from the very beginning. If the analyses are very similar and/or using the same data to estimate the same quantities (even with different methodology), it is unethical to work on both, and it is up to your co-authors to ultimately decide whether or not there is a serious conflict of interest, not you. If your co-authors believe there is a conflict of interest, you ethically must recuse yourself as co-author on one of the papers.
If you are a junior co-author on a paper that has some analysis methodology outside your current abilities to replicate some of the results (this happens, particularly early on in your career), at the very least thoroughly proofread the paper, and do it within a day or two of the primary author sending it to you. The more senior co-authors will appreciate the thorough proofread.
If you are the primary (corresponding) author on the paper, you will probably be doing or coordinating most of the analysis in collaboration with the other authors, be doing most of the write-up (with help here and there from co-authors, if necessary), and will also be the person responding to most (if not all) of the reviewer comments. Usually you are the person who conceived of the analysis in the first place, and generally, you and/or the senior author have say over who gets co-authorship. It is rude to be a co-author (not the primary author), and invite people to be part of the author list without first consulting the primary author. If you are not the primary author, and are about to ask someone for significant help with some aspect of the analysis, first contact the primary author to discuss the potential of offering co-authorship to the person you are about to ask for help. Do this before you ask someone for help.
Make sure that as you are writing, you include an Acknowledgements section that acknowledges anyone who has given you advice on the analysis, and/or acknowledges grant support.
Ordering of authors in this field usually, but not always, follows the pattern where the corresponding author who primarily wrote the manuscript and conceived of the analysis is listed first. The last author is usually the most senior person on the collaboration (it is usually obvious who this person is). In between the first and last author, the situation gets a bit fuzzy, and here is where sometimes unhappiness can arise. Often the second author is the person who had the most to do with the analysis next to the primary author. Then just list the remaining authors alphabetically. Or list all authors between the first and last alphabetically. Whatever you do, do not arbitrarily order the authors anything but alphabetically, or alphabetically within institute. Some people are sensitive to author ordering, and diplomacy may be needed.
I’ve worked on analyses that are so intensely collaborative that no one author could be identified as the driver of the analysis, so we voted to just order everyone alphabetically, and got one person to volunteer to be corresponding author and do most of the writing up.
For your part as a co-author, try to be somewhat easy going about author ordering. I can assure you that very few people other than you will care whether or not your name was second, third, or fifth… what a future employer might care about is whether or not you can succinctly describe the analysis and what your role in it was. And whether or not people speak of you as someone who is easy going, or hard to get along with.
I need to point out that the primary author (and/or the senior author) is also usually responsible for ensuring that there is no scientific misconduct in the analysis. Even if it was not you, but one of your co-authors who, for instance, manipulated the data, it is up to you as primary or senior author to ensure that they have not in fact done that. It is always a good idea, when possible, to get at least two authors on a paper to independently verify results.
Sometimes conflicts arise between co-authors on some point of the analysis, where one or more co-authors think things should be done one way, and one or more other co-authors vigorously disagree. On rare occasions, this can lead to an impasse where neither side is willing to give, and feel that what the other side wants is not worthy of being published. If you really find that you do not want to put your name on a paper that employs a certain methodology that you find deeply flawed, you have the option to recuse yourself as a co-author. However, this is almost never done… if you do that, you run a strong risk of losing a good working relationship with colleagues/friends, and also for developing a reputation in the field for being hard to get along with. I’ve only been in this situation once. I chose to remain as co-author. But I did politely, but pointedly, ask that a couple of sentences be added to the paper to clarify caveats about the analysis.
For the situation where you are a primary author, and one of your co-authors strongly disagrees with the analysis methods you have employed, the situation is even more complicated; you cannot arbitrarily choose to later exclude a collaborator as co-author once you have invited them to collaborate. It just isn’t done (not that there are co-author police who monitor and enforce this… it is just in appallingly bad taste). It’s also in bad taste to ask them to recuse themselves. For situations like this, perhaps ask a well-respected impartial senior colleague to help mediate the situation, and find at least some common ground that can allow the publication to move forward with perhaps neither party fully liking it, but at least getting enough of their own way to mollify them. A relatively impartial senior colleague may also gently suggest to the parties that perhaps two separate publications are in order, if the situation cannot be resolved. Keep it polite.
If you find that a co-author has an un-hitherto disclosed conflict of interest, the situation becomes extremely delicate, because (as explained above) this now becomes a question of ethics, and universities have specific rules about such things. Conflicts of interest might include an undisclosed source of funding that might encourage bias in research (for instance, getting funding from a tobacco company to study the rates of lung cancer among smokers). Another conflict of interest is if the co-author is participating in a very similar analysis; the problem is that the co-author can either potentially communicate the details of one analysis to the other analysis group in a way that undermines one of the analyses, or even (consciously or unconsciously) try to hold up one of the analyses so that the other can get published first. In these situations, you need to explain to that author how the conflict of interest interferes with the integrity of your paper, and gently suggest the co-author recuse themselves, but offer to acknowledge their contributions in the Acknowledgements of the paper. I stress again, keep it polite. If the co-author doesn’t recuse themselves at that point, mediation by an impartial well-respected senior colleague may be required.
Submitting your paper
As mentioned above, long before you finished the paper, you should have known what journal(s) you were aiming for. Before submitting, double check that the article format, word count, etc conforms to the journal guidelines. You’ll find the journal submission guidelines if you do a Google search on the journal name and “instructions for authors”.
You need to write a cover letter to the editors that gives a brief description of your analysis, what is new and novel about it, and why your paper is appropriate for publication in that particular journal. Here is a hypothetical example of what such a letter might look like (the topic of which may or may not motivated by a problem my next door neighbors are currently having with their kids…). Feel free to copy and paste and modify it for your own uses:
Aug 30, 2014
We are pleased to submit for your consideration, our manuscript entitled “A Mathematical Modelling Analysis of the Spread of Human Head Lice in a Public School Environment”. In our analysis, we employ a Susceptible, Infected, Susceptible (SIS) compartmental mathematical contagion model to simulate the spread of head lice in a small Indiana public school. Using a weekly time series of head lice prevalence data recorded by the school during the 2012/13 school year, we fit the parameters of this model, and determined that the basic reproduction number for the spread of head lice in that environment was R0=2.5. Using our model we also assessed the potential efficacy of intervention measures, such as crew cuts among the males in the student population.
Head lice are increasingly becoming a problem in the US among school children, due to rising resistance of the parasite due to insecticidal resistance. Several million children are affected each year, and the problem is estimated to cost the US economy approximately $1 billion annually in treatment products and parental lost work time. Ours is the first analysis to apply a mathematical model towards assessment of the efficacy of intervention strategies in the spread of the parasite in school populations, and we feel the topic of paper is appropriate for your journal, Quantitative Parasitic Epidemiology.
All authors have read and approved the final manuscript, and declare no competing interests.
If you need further information, please do not hesitate to contact me at LucyLouse@asu.edu or 408-555-5423. We look forward to hearing from your office soon.
Sincerely, on behalf of all authors,
Dr. Lucy B. Louse
I usually list suggested reviewers as a postscript in the cover letter, as well as in the online forms you need to fill out as part of the submission process. You can come up with suggested reviewers by looking at the author list of seminal publications you have cited in your paper and/or by asking the advice of more senior people in your field.
The editors will examine your manuscript, and give it a sniff test to ensure that a) it does indeed look to be appropriate for the journal, b) is not quackery, and c) is a more or less well written paper and apparently sound analysis. If it passes these checks, the editors will assign it a manuscript number (you’ll be notified by email) and they will put it out for review.
If you are submitting to an open source journal like PLoS, which has significant publication charges, and you do not have grant funding available to pay for the publication charges, you can write a letter to the journal to plead poor. State that you are junior, have no grant funding yet, and would be grateful if they could waive the publication charges. I’ve yet to hear of anyone being turned down when pleading poor.
Responding to reviewer comments
After a period of weeks to months (sometimes even longer!) the editor will forward on to you the reviewer comments (usually from two or three reviewers). If you don’t hear back from the editor after a few months, it is perfectly acceptable to send them a polite email inquiring about the status of your manuscript. This usually prompts the editor to send a snarky letter to the reviewer who is dragging his or her heels, and often results in getting the reviewer comments not long afterwards.
Responding to reviewer comments can seem daunting the first few times you do it. It is important to never get defensive! The reviewers’ job is to make your paper better, and their comments are almost always towards that end. Some reviewers are politer than others. I appreciate journals where either the review is double blind (where you don’t know who the reviewer is, and they don’t know the identity of the authors) or double non-blind (everyone knows who everyone else is). This encourages politeness both ways. Unfortunately, only a few journals do it this way.
Even if one or more of your reviewers is snarky and rude, never be rude back. Always be excruciatingly polite in your replies, no matter what.
I usually format my replies in a Word file, where I first copy and paste all the comments made by each reviewer. I preface the file with a short statement thanking the reviewers for the careful reading of the paper, and helpful comments. In the file I highlight the comments of each reviewer in bold black text. Point-by-point, underneath each comment, in red non-bold text, I provide replies.
If the reviewer points out any mistakes in the paper, thank them each and every time for pointing it out. Sometimes reviewers will point out things that they didn’t understand, even though you feel you covered it extensively in the paper. I always thank them for pointing out that the text needed clarification, and state that I’ve added sentences to clarify the text (sometimes the sentences just consist of a little preamble that gives an overview of what the paper is about to describe in detail).
If the reviewer recommends major changes or cross-checks to your analysis, do not submit a new version for review until you address each suggested change or cross-check, even if it means it will take you a few months of extra work. Unless of course you have a very good reason for not doing the cross-checks, in which case you have to carefully, and extremely politely, convince the reviewers and editors why you are right.
Sometimes a reviewer will make a strange comment that will lead you to suspect that he or she may have in fact been drunk when reading your paper. If the comment is really inscrutable, ask your co-authors if they can figure out what is being asked. If you still can’t figure out, politely state that you aren’t entirely sure what the reviewer meant by that comment, then give your best guess at what was being asked, and answer that. The reviewer will get back to you with further clarification if it really was an important point.
Some papers go through several rounds of review. If you are careful in thoroughly replying point-by-point to reviewer comments in the first round, you decrease the probability of second or third rounds of comments, increase the chances of acceptance, and your publication rate will go up.
Above all, pay attention to the sociology and psychology of this process, and get along with people. If you are impolite to reviewers, they will remember your name, and in all likelihood this will cause you problems later in your career if you have a reputation for being anything other than easy to get along with.
Rejection and Acceptance
Sadly we will all, as scientists, go through scientific rejection in our lives. It disturbs me how often I need to recommend rejection of articles I review (usually about 50% of the time… I’ve asked my colleagues what their rejection rate when reviewing is, and they too bat about 500, which tells me that my standards are about right). There are a lot of very poorly written manuscripts (and/or poorly conceived or poorly motivated analyses) that go out for review, and the ones that the reviewers see are the ones that already passed the editorial sniff test to get put out for review in the first place… I sometimes wonder what the papers look like that get rejected outright. If you follow all of the tips I give above, you will significantly increase your chances of successfully getting your paper published.
Also important to successful publication is aiming for the right journal; when I first started in the field of mathematical epidemiology in 2010, I was running about a 30% rejection rate for articles I wrote, which was frustrating. I eventually learned more about the available journals, and knowing which journals were most appropriate to my research topics helped a lot to increase my acceptance rate. Ask senior people in your field for advice about appropriate journals.
Another thing that can help lower your rejection rate is not to aim too high in impact factor; if all you ever do is submit your articles to Science, Nature, or PNAS, in hopes that someday your ship will come in and they’ll publish you and you’ll be famous, you will deal with a lot of rejection (they tend to only publish about 10% or less of submitted articles). Plus you will risk annoying the editorial staff if you are continually submitting manuscripts to their journal that are not high impact material. However, you don’t want to aim too low either, or the exposure of your work will suffer. Like I said, get advice from your more senior colleagues on what journals might be appropriate for your analysis. Then go to those journals, and read several articles to ensure that the content and scope matches the kind of paper you’re trying to write.
If your work gets rejected by a journal, either outright by the editors, or upon review, pay careful attention to the comments that were made, and improve your manuscript and/or analysis accordingly. Then find another appropriate journal to which you can submit your work. Sometimes even well written papers on a well motivated topic will get rejected just because the topic wasn’t considered interesting enough at the moment. For instance, trying to get an Ebola modelling paper published in 2013 would be a lot harder than it would be in 2014. In 2013, editors would likely see the motivation for the paper being somewhat weak, because no one besides specialists at that point cared that much about Ebola; the editors would consider the audience of the paper too narrow to warrant publication.
Some journals are broad in the subject matter they cover, and some journals are very narrow and specific. You need to think about what kind of audience you are aiming for. It’s sometimes nice to aim for a mix on your CV, depending on your field; the specific journals are where you can really get into the nitty gritty of the details of your methodology, and it is publications like these that will impress future potential employers with your technical expertise. The more broad journals often publish things that are of general public interest, and it is through these journals that your work can get noticed by a wider audience.
Speaking of which, if you publish in a broader interest journal, consider having your university write a press release about it if the work is something that might be of public interest (ie; is the topic of your paper something you could discuss with your elderly auntie over tea, and she would be interested in it, and easily understand its significance? If so, the work is probably of public interest). If your work gets press, it can help you in your future job searches and/or future grant requests that are related to that work. All research universities have press/media offices, and they employ someone whose job it is to help researchers write press releases. All you need to do is email them and ask.