5.5- Introduction

In this next module we’re going to talk about writing the introduction section and the good news is that the introduction section is actually easier to write than you might think. The introduction section follows a fairly standard format. Because it follows a definable format it’s easier to write than you may think. It’s also fairly short. The typical introduction section is only three paragraphs long. You may not have realized that before. You may have thought it was much longer but an introduction section should be no more than five paragraphs. It typically ranges between two and five, and typically it’s about three paragraphs long. The biggest mistake that people make when they’re writing the introduction section is that they think it’s supposed to be some kind of long, exhaustive review of the general topic. You’ve spent all this time collecting information and reading papers and you feel like you need to stuff all of that into your introduction but that is not the purpose of the introduction. The introduction section is focused around the specific question or hypothesis or aim of your study. You shape the whole introduction around that specific question or hypothesis. For example, if you’re writing about a study that tested the association between breast cancer and smoking, you are not going to write a whole bunch of background information about breast cancer in general or about smoking in general. You’re not going to talk about papers that are just about smoking and how bad it is or just about breast cancer and how bad it is. Rather, you’re going to keep your introduction section focused on the potential relationship between specifically smoking and breast cancer. You’re only going to talk about previous studies that addressed this link, not studies that looked at breast cancer in general, not studies that looked at smoking in general. So it’s a very narrow section which actually means that it’s pretty easy to write. I’m going to take you now through the standard format. A good way to think of the introduction is as a cone, as it’s pictured here. This is a figure I’m using from one of Thomas Annesley’s papers in the journal Clinical Chemistry. He represents the introduction as a cone. The idea is that you’re going to start with something general and you’re going to narrow down very quickly to your specific study. So you start your introduction by briefly giving some context, some background information, about what is known on your topic, what’s known about breast cancer and smoking about that link. Then you quickly move from the what’s known to the what’s unknown. You pinpoint for your reader the gaps and limitations of all those previous studies on smoking and breast cancer. What were the flaws in those studies that make us still unsure of whether breast cancer causes smoking? Then you narrow down even further to your specific hypothesis or question or aim of your statement. And you’re going to make a statement of purpose where you’re going to say, quite clearly, we hypothesize that or we wanted to answer the question of or we aimed to. It’s good to use that explicit language with those key words in there, we hypothesize, we aimed, because it helps the reader to find the statement of purpose of your study. So you quickly tell the reader what’s the main hypothesis or aim of your study, then you tell them a little bit about your experimental approach. And you say how your experiment on breast cancer and smoking is different and new and better than previous studies. Tell your reader how your study is going to fill in all the gaps and limitations of the previous studies. Similarly, Mimi Zeiger, in her book that I’ve referenced before, gives a similar structure for the introduction. So, she says you start with what’s known– those are the previous studies. Then you give the what’s unknown– those are the gaps and limitations in those previous studies. And then you move into the specific question or aim or hypothesis of your study, so, we aim to, we hypothesize to. Then you give away just enough about your experimental approach to show how your study is going to fill in the gaps and limitations of previous studies. You can think of this as roughly three paragraphs. You don’t always have to have it structured this way but you can see how it neatly fits into three paragraphs. Paragraph one might be the what’s known, paragraph two can be the what’s unknown, and paragraph three can discuss your specific study. Now authors may also go back and forth between the known and unknown in those early paragraphs, you see that as well, but this gives you a rough feel of the breakdown of the introduction section. Now a couple of tips for writing an introduction section. First, I’ve told you it’s only about three paragraphs, so that might make you think that you need to be, to write really long paragraphs. But I actually mean three short paragraphs. They should be crisp and short. Also, as much as you can, try to write for a general audience. Your manuscript, of course, is going to have technical details but save those technical details for the materials and methods, give a friendly introduction to what you did in your study. As I’ve said in the previous slides you want to go from broad to narrow, from the known to the unknown to the question and you want to have that specific purpose statement where you say, we asked whether or we hypothesize that, our aims were, something like that. When I’m reviewing papers I often scan the introduction for that statement so that I can figure out very quickly what was the point of the paper. So make sure it’s there and it’s clear and easy to find. Don’t forget to emphasize how your study fills in the gaps of previous studies and, again, explicitly state your research question, we asked whether, our hypothesis was, our aims were. The introduction section should not contain results or implications. You don’t want to answer the research question, just set up the research question. Finally, in the introduction section you’re going to summarize the previous literature at a high level. That means we’re tempted when we read all these studies to go into details about each individual study. For example, you might say, Jones et al did this case control study but that study had problems with x,y, and z and then Smith et al did this other cross-sectional study and the problem with that study were a,b, and c. You’re going to be tempted to go into this level of individual detail but you’re supposed to actually summarize the study at a high level. So what you want to do is say, you know, if there were nine studies that have looked at breast cancer and smoking, summarize for the reader, well, two found an association and seven didn’t– give a high level summary. And when you’re talking about the problems in the previous studies, again, you don’t want to go into the nitty gritty of each individual study especially if there’s a lot of them, just tell the reader broadly what were the problems with the previous studies. I love the BMJ Christmas issue so I’m going to start with an article from that. This was the study I mentioned before that looked at spam emails to academics and I think this is a great example of an introduction section. It’s nice and short and I can fit it all in one powerpoint slide. So notice it’s only two paragraphs long. This is partly because it’s in the BMJ Christmas issue which is a humorous issue but really introduction sections aren’t going to be much longer than this. Notice that the first paragraph is all the what’s known, the background information on unsolicited academic spam. Unsolicited and unwanted (spam) electronic invitations to speak at or attend conferences or to write for or edit journals are a burgeoning aspect of academic life. Colleagues regard such invitations with wry amusement, intense frustration or resignation. Two of us have reviewed travel grant applications from colleagues who received spam invitations to give conference presentations. Then we get to the unknown. There’s just a one sentence statement of the unknown here. Few studies have focused on academic spam. Here they are identifying the gap in the literature. It’s an easy gap to figure out. The gap is that no one has studied this formally before. Then we get to the specific study in the academic spam study, that’s their study. We investigated the amount, relevance, content, and suppressibility of academic emails. They’re definitely filling in a gap here because they may be the first study to ever look at this and they’ve explicitly stated what they want to look at– amount, relevance, content, and suppressibility. So that’s a very simple example of an introduction section to start with. Now we’ll move on to something slightly more complex. Here’s another example, notice again that this one is only two paragraphs long. It’s slightly longer than the last example but still only two paragraphs. This one was looking at whether or not obesity and overweight was related to mortality from cancer. So they start with the what’s known. We know that if you’re overweight or obese you’re more likely to die from all causes as well as from heart disease. Those relationships are well-established. We also know that excess weight is important in cancer, in general, but there’s a whole bunch of gaps in that. So here’s the unknown. We don’t know the magnitude of that relationship. How much does being overweight increase your risk of cancer? And we don’t have a good idea of exactly which cancers, which individual cancers, are related to excess weight. All right. So now they’re jumping from known to unknown now they’re gonna go back to the known. So previous studies, what do we know for sure? Previous studies have consistently shown associations between adiposity, excess weight, and increased risk of certain cancers, so endometrium, kidney, and gall bladder for women, breast in post-menopausal women, and colon in men. Those have all been well established. Then we jump back to the what’s unknown, so data on cancers of the pancreas, prostate, liver, cervix, and ovary, and on blood cancers are scarce or inconsistent. So they summarized a pile of studies in just this one sentence. We have a bunch of studies on cancers of the pancreas, prostate, liver, cervix, ovary, and blood cancers. They list them all here but they’re inconsistent. You notice how they don’t give any details of the individual studies listed here. They’re just telling you really they’re inconsistent. We haven’t answered the question. And then they tell you specifically what’s been wrong with previous studies. What are the gaps and limitations of previous studies? Why aren’t these questions answered? And so the lack of consistency may be due to the fact that just there aren’t a lot of studies, especially there aren’t prospective studies which are the superior study design. Are people categorized overweight and obesity differently in different studies? There can be bias introduced with respect to smoking. And so there’s a whole bunch of gaps and limitations that they identify and then they jump into how this study will answer the question with better methods. So they say, we conducted a prospective in this investigation, in a large cohort of U.S. men and women to determine their relations between body mass index and the risk of death from cancer at specific sites. So the aim of this study is to look at death from cancer and from specific cancers. And they’re going to use body mass index rather than categorizing overweight or obesity which involve some arbitrary categorization. So they’ve already told you how they’re filling in some of the gaps– it’s prospective, it’s large, it’s a trusted data set they tell us in the next sentence and they’re looking at BMI rather than an arbitrary cut off for obesity or overweight. Okay. One more example. This one is actually three paragraphs long. So this is a study looking at exogenous estrogens in young women to see whether it affects their bone density. So this one starts with the what’s known. So we know that exogenous estrogens, that is giving women estrogen in post-menopause, that increases bone mineral density. That’s well-established. Okay. Now we jump to the what’s unknown. It’s unclear whether when you give exogenous estrogens in the form of the oral contraceptive pill to young women, it’s unclear whether that affects their bone density. And here we get a review of a bunch of studies and you can see in just one sentence we are summarizing all of these studies. Several studies suggest that exposure to oral contraceptives during the premenopausal years has a favorable effect, increases bone mineral density, and other studies show no effect. So we get just a quick summary of what the literature shows us, it’s, it’s unclear, it’s inconsistent. Some studies do show an effect, increase in bone density, but a lot of the studies also show no effect. Then we get the gaps in the previous research. Why isn’t this question answered? Well, past studies of the relationship of OC use and bone mineral density have had several limitations so authors used measures that were kind of crude such as current, past, and never use of oral contraceptives which isn’t going to tell you about the dose of estrogen. And many studies also didn’t account for things like lifestyle characteristics and then finally there aren’t many studies that look in races, you know, women of races other than white. So those are the gaps and then we get to this study. The aim of this study was to evaluate the associations of world contraceptives with spine, hip, and whole body BMD. Notice that it says the aim of this study was in black and white pre-menopausal women. So we get a statement of the aim of the study. Notice that we’re including both black and white pre-menopausal women, that’s filling in a gap in the literature. Our primary hypothesis, the, we get a hypothesis statement here, was that there would be an association between cumulative exposure to estrogen from oral contraceptives and bone mineral density. So they tell us that they’re going to be using a measure of cumulative exposure rather than just a simple ever/never use of oral contraceptives. So that’s the basic structure of an introduction section.