Introduction

Research methods is  a general term that includes several related activities: (1) observing the world around you, (2) critical reading of literature or critical appraisal of the video or lectures you come across, (3) evaluation of the information, (3) writing a research proposal, (4) drafting a research protocol, (5) mode of collection of information relevant for your research, (6) analysis of data, and finally (7) writing the actual research piece that will evoke further studies.  In this tutorial, we are going to discuss how you can critically read and write research. This is a fundamental skill for every researcher in any discipline. We will first lay out the general principles and then we will discuss the more specific issues.

How to read an article or a report

Six types of statements

As you read an article, pay attention to each sentence and how sentences are linked; also pay attention to how paragraphs together hold the entire paper and flow from a discussion of the background of the research, the methods the authors adopted to test or find the key findings, the description of the findings themselves, and finally, the discussion and conclusion. This format, termed IMRAD (introduction, methods, results, and discussion) has standardised the practice of writing research papers and we are not going to elaborate a discussion of this format here. For more information on how to use this format, consult a relevant journal where you want to send your paper. Todorovic (2003) has provided a tutorial on how to use IMRAD format \cite{todorovic2003original}. Taking a modular approach to reading a paper,  we recommend that you note each sentence carefully as you read a paper. Tom Chatfield (2017) describes in his work, "Critical Thinking" that you can classify each sentence/paragraph that does not include any reasoning embedded into one of the following types \cite{chatfield2017critical}:
So, description, summary, opinions, beliefs, clarification, and illustration provide a class of expressions and statements that you should note with the caution that they are not associated with any reasoning. There are places and situations to use them, as descriptions and summaries contain data that are essential for developing further arguments.
You should be cautious about opinions, beliefs and rhetorical manipulations. In general, treat opinions and beliefs with circumspect if not associated with reasoning; they may not be useful for developing arguments or explanations which you will seeking as you read or be ready to learn from academic writing. Besides, you should be cautious about rhetorical manipulations and rhetorical elements. Rhetorical elements are statements associated with "emotional languages". For example, "It is high time we took action against the fracking industry as they are playing with the lives of millions and the life of the planet". Note, that in statements such as this, we do not see an associated rationale or reasoning that we can argue with or use as explanation. Irrespective of their intentions, you should treat emotional laden language that are not associated with any reasoning with skepticism as well (note that this statement itself is an opinion). 
This leaves us with two other classes of statements/passages/paragraphs that you will come across in any academic or scholarly writing or "paper": explanations and arguments. Explanations are statements that rely upon preceding statements that are either descriptions or facts, and explanations build upon them by adding reasoning statements. However, explanations only offer reasoning behind facts or state if a pattern is being observed or some form of generalisation. For example, "Compared with those who did not have breast cancer, those who had diagnoses of breast cancer were 1.56 times likely to consume high cholesterol diet. This suggests that high cholesterol diet on a regular basis is a risk factor for breast cancer." This statement has two sentences, and the second sentence is a "conclusion" that high cholesterol diet consumption on a regular basis is a risk factor for breast cancer; the first statement is a rationale for arriving at such a  conclusion that those who did not have breast cancer were found to consume less cholesterol in their diet. When we have statements such as this, we label these as explanations. Explanations are some of the most common statement sets in scholarly literature and we should pay attention to the various facts and statements that make up explanations.
Now, consider another class of statements: "Compared with those who did not have breast cancer, those who had diagnoses of breast cancer were 1.56 times likely to consume high cholesterol diet. This suggests that high cholesterol diet on a regular basis is a risk factor for breast cancer. Hence we recommend that cholesterol contents of diet should be restricted". This statement or set of statements too, have pointed to facts or descriptions, they have provided rationale or added reasonings, and over and above, they have provided a "call to action" or some form of persuasion, as to why is this explanation contextual. When you have a series of statements that are not only associated with reasons, but also associated with "a call to action" or some form of persuasion, then you deal with "arguments". Arguments, as you may see, are the ones with which you can engage actively as you read literature and you will provide counter arguments or provide alternative explanations to counter their claims. We will deal with arguments and explanations in the rest of this paper. 

How to construct a standard form of argument and why will you do so?

The first thing you do when you read a paper is to scan for these various forms of sentences or statements and look for or scan for arguments that the authors have provided. If you look for arguments and explanations, then the authors would have provided you with a final conclusion or a final concluding statement, and several supporting statements to support the conclusion as reasoning. The statements that the authors have provided as reasoning to support their claim for hte conclusion they have made in the paper are referred to as "premises". We will differentiate between two different classes of premises: explicit premises, and implicit premises. Explicit premises are those statements that act as premises and those that hte autors have articulated in the paper. You can easily identify them as they are associated with the conclusions. You spot conclusions in various sections of the paper by noting identifier words or phrases. Look for phrases such as "hence", "given", "given that", "therefore", and so on; these words and phrases are those that signal that a qualifying statement is going to follow, and a set of statements that have supported the statement will be found either next to the concluding statement or preceded the conclusion. Look for these statements. A paper is based on more than one argument and explanation. Each argument will have exactly one final conclusion, and several premises. 
Compared with an explicit premise, which you will find in the article, you will need to supply the implicit premises by closely reading the explicit premises and your own common sense. For example, consider the following statement:
"Since the beginning of the agricultural and industrial era, humans have generated greenhouse gases and emitted the gases in the atmosphere. As a result, the global temperature has increased resulting in global warming". 
If you read closely, you sense that this statement is actually made of three separate sentences:
Now this example shows you how a standard format of argument is constructed. You start with a conclusion, and there is only conclusion per argument, and the conclusion of an argument can well become a premise in a following argument that is connected to this argument and so the paper flows as a series of connected arguments or independent arguments that act as premises of each other resulting in a final conclusion of the paper. But each argument will have only one conclusion. You always start with a conclusion and you always first identify the more obvious, explicit premises. Then, after you have identified all explicit premises that lead to an argument, you fill in the implicit premises that lie in between the explicit premises and the conclusion. In the process, you may see that you have supplied some intermediate conclusions as well and that is OK. For example, in the statement above, where we supply the implicit premise "Greenshouse gases cause rise in temperature", could well be an intermediate conclusion. The implicit premises are also referred to as "implicit assumptions" or you can call them "assumptions".
Why are we breaking our heads over these things? Are these not common and can we not always assume them? The answer is "No we cannot". We are not allowed to go beyond what the authors have stated or supplied as facts, BUT, we can always argue by linking these explicit assumptions with other reasonings that either follow from them or are implied by them, which we term as "implicit assumptions". The implicit assumptions are the ones that we will need to examine. For instance, here, we would need to verfiy or justify that the claim, "greenhouse gases, when they accumulate in the atmosphere, will actually lead to increased temperature". It turns out that this fact is "absolutely correct" or "certain" as this "proof" is obtained from physical experiments where Svante Arrhenius in the 1700s showed that gases such as carbon dioxide when they accumulate in closed systems increase the temperature of these systems. So there is a rational basis of this argument and we can accept it. But it will not always be easy or simple as this. This is also an opportunity where you can argue the "missing" pieces in an argument that then forms the basis of your own research.
So, here are the rules of developing standard format arguments (follow these steps, we will see in an actual reading piece):
  1. First, identify the strongest argument you can find in a paper or a piece of writing. Also, assume that the author is authentic and is knowledgeable to write what he or she is writing about. This is referred to as the principle of charity. While we label this as principle of charity, it has nothing to do with philosophically being kind to the authors or adhering to anyone's authority. It is done for a practical purpose. If you do not adhere to the principles of charity, and if you do not adhere to the strongest argument that you can examine, then you cannot claim that a statement or thesis is to be refuted, as referring to a weak argument makes no sense. Hence, you should always identify the strongest argument and provide the author the benefit of doubt.
  2. Second, you should always identify the conclusion of an argument. Start with this first. Set up a spreadsheet and enter this as the first element in a cell. You will need to add cells to the spreadsheet on the TOP of this cell to enter explicit and implicit arguments, but do this first.
  3. Third, write the explicit arguments clearly. Identify ALL explicit statements or arguments that the author or authors have stated and write them clearly on the top of the "final conclusion". So, the final conclusion will lie at the bottom and on the top of the final conclusion, you will list all the premises that are explicit. At this stage, do not worry about the order in which you identify them or write them. It is important that you should not miss any of the explicit claims. If you are in doubt, include the explicit premise. You can take it out later.
  4. Fourth, mindfully read the explicit premises and try to identify what implicit assumptions or implicit premises they are based on. Some of these assumptions will be intermediate conclusions, and that is OK. At this stage, you are only going to identify the implicit assumptions or implicit premises.
  5. Now that you have identified ALL the explicit and the implicit premises of the argument as well as the conclusion (final conclusion) of the argument that you are about to examine, read the premises mindfully. Do you see that these premises are linked or unlinked? Linked premises are those premises that should depend on one another to make sense or lead to the final conclusion. For example, consider the following statement: "We found increased concentration of inorganic arsenic in the groundwater of the Region X where we sampled the water. Experimental studies suggest that exposure to high concentration of inorganic arsenic in the groundwater leads to bladder cancer among mammals that were studied. We suggest that we should screen for evidence of bladder cancer among residents in Region X". Let's construct a standard format argument from this (EPx == Explicit premise x where x = 1,2, 3, etc; IPx == Implicit premise or assumptions x where x = 1,2,3, ...; ICx = intermediate conclusion x where x = 1,2,3, ...; FC = conclusion):
Can you see the complexity of information hidden in a simple four sentence statement? In the statement, we have stated that we have found high concentration of inorganic arsenic in the groundwater of region X and that we know that if certain mammals are exposed to high concentration of inorganic arsenic, then they are likely to develop bladder cancer; therefore, we recommend that we should search for people who have not yet developed signs of bladder cancer but can develop and start a programme. Each of these implicit assumptions and intermediate conclusions are open to further investigations and investigations as to whether there are knowledge gaps. For example, where is the evidence that inorganic arsenic acts similarly in the body of all mammals and all mammalian systems are sufficiently identical so that we can extrapolate the results from one species to another. Are there inter-species variation in which inorganic arsenic is metabolised? Is it only inorganic arsenic or are there other compounds that are either associated with inorganic arsenic or their byproducts that can give rise to cancer that we should be mindful of? Besides, are these premises linked to one another? For example IP3 and IP4 (human beings are mammals and inorganic arsenic acts similarly in the body of all mammals similarly) are linked to each other and are crucial for reaching the final conclusion. If one of them fails, the final conclusion will fail. Similarly, IP5 and IP6 are linked to each other. Although other premises may not be as linked and can be treated as unlinked premises and they stand on their own.
So, we suggest that when you read a piece of literature critically, this is the first thing you should do. Start noting the types of expressions, and note the sentences or statements that are associated with some form of reasoning with them. Note which of them are explanations so that they do not call for any form of "call to action" or "persuasion", and note which ones call for action and are aimed to persuade you to take certain action. The ones that aim to persuade you in some ways are 'arguments'. Then, once you have noticed arguments & explanations, you start constructing standard formats to develop explicit and implicit premises, and lay out your argument maps. After this, we will examine the nature of such reasonings. Are these robust or sound arguments? Are these cogent arguments? Are these arguments forceful enough? Can we explain these relationships and if we can explain the relationships and would like to answer the question, "why we see what we see, what can we do about them?" This is where classifying and analysing the nature of the reasoning makes sense.

Logic, probability, and explanation: the three classes of reasonings

After you have constructed the standard form of argument, you identify the type of reasoning that the authors have used to arrive at their conclusions. Do these conclusions make sense to you? We state that there are three classes of reasoning that you should be aware of: reasoning that are pure logic and express relationship between implicit and explicit premises and the conclusions they reach purely by the force of logical thinking. In doing so, you rely on your common sense and your close reading of the statements. This class of reasoning is referred to as "reasoning by deductive logic". Reasoning by deductive logic may sound like detective novels or detective work and indeed they are like that. All we say here are that the statements should make good logical sense, or good common sense to us.
Reasoning by deductive logic: the play of logic
Consider the statement, "Exposure to high concentration of inorganic arsenic through water in all mammals result in bladder cancer. As humans are mammals, when human beings are exposed to high concentrations of inorganic arsenic through drinking water, they will develop bladder cancer". If we break up this statement into a standard form, we get the following:
Now each of these sentences are true. This argument is a deductive argument and note the format:
All humans are mortal.
I am a human.
Therefore I am mortal.
Note also that when we discuss something like this, or a logical set of statements, we also state that there is a certainty that will happen. For example, there is certainty that if mammals are exposed to inorganic arsenic through drinking water or water they drink, then they will develop bladder cancer, and as humans are mammals, this holds true for them as well. This element of certainty differentiates reasoning by deductive logic from the reasoning by inductive logic, we will discuss next. 
These things may appear intuitive, but we need to keep in mind a few other rules as well that will become handy as we examine text where these rules are often violated as people write and as people think. Note that everything we write here not only relates to what we read, but also what we see in the form of video or graphics, what we hear when people speak as in lectures or in everyday language, or what we experience through our thinking. Here then are a few elements of this logic:
If X then Y ... (1); here, in this structure we term X as antecedent, and Y as consequent
X is true: therefore Y is true (this is referred to as affirmation of the antecedent)
Example: If X is a mammal, then exposure to inorganic arsenic through drinking water will cause bladder cancer; X is a mammal, therefore it is certain that X will develop bladder cancer exposed to arsenic through drinking water. 
Here, this argument is a "valid" argument.
Note another one:
If X then Y ... (1)
Y is true: therefore X must be true .... (2) (this is an invalid argument and is referred to as affirmation of the consequent)
Example: "if the surface temperature of the seas increase, then there will be increased intensity of cyclones when they have landfall; we have experienced increased intensity of cyclones during landfall; it follows that the surface temperature of the sea have increased"
This is an invalid argument, even if both the statements are true (that is, there have been increase in the cyclone intensity over the past years, and that the sea temperature has remained an all time high). But just because the cyclone intensity has increased does not imply that the sea surface tempeature is high. The increase in the intensity of cyclones can be due to many other factors. You will frequently encounter these sort of errors in emotion laden languages and in non-fiction written to influence or present specific points of view. For example, in a recent book on Climate change, an author has claimed that he has experienced severe storms in Mumbai, India when he was young; remembering the storm surge, he now states, that the storm surge indicates the play of climate change. This is an example of "affirmation of the consequent" and is an invalid argument. What he argues is something like this:
If the global temperature increases, there will be severe cyclones and their intensity will increase ... (1)
I have experienced heavy cyclones or high intensity cyclones
This implies that global temperature have increased
It might as well be that the global temperatures have risen and also that storm intensity or cyclone intensity have been high, but the way he presented it do not make logical sense. Such logical leaps are invalid logical leaps and you must keep your eyes and ears open to catch this type of logical inconsistency.
Consider a third case:
If X then Y .... (1) 
Y is not true; therefore X is not true as well .... (2) (This is valid argument and is referred to as negation of the consequent)
Example: "If there are sudden increase in the concentration of PM2.5 in the air, there will be abnormally heavy cases of deaths due to heart attacks in the city;  deaths due to heart attacks have been steady or at least not in record high levels: hence, the sudden unexpected rise in PM2.5 has not happened". This statement may be convoluted and we will see in the next section that these things fit well with reasoning by induction, but this illustrates the situation with negation of the consequent that this is a valid argument. The fact that Y is not true signals that X did not happen if X was the only reason Y should have happened.
But, on the other hand,
If X then Y ... (1)
X is not true: therefore Y is not true as well ... (2) is invalid and is referred to as negation of the antecedent. 
If we were to continue with the previous example, it would be stated something like as follows:
"Sudden increases in PM2.5 levels will cause high rates of admission due to heart attacks in emergency rooms; PM2.5 levels have not risen high today, therefore we will not expect high rates of admission due to heart attacks in emergency rooms". This assertion is not true as there may be other reasons why there would be high case loads of heart attack related hospitalisations.
We have here scraped the surface of what may happen with logic and how you should keep your eyes and ears open to the possibility of how people play with logical conclusions. However, in health sciences, these errors are relatively rare because the scope of deductive reasoning is not very popular. Instead, in health sciences and health care we are interested in another form of reasoning; this is reasoning by induction.

Sound and unsound arguments

In the above sections, we dealt at length about validity of the logic and conclusions we have drawn from the premises. However, along with the fact that the logical structure of the arguments be valid, the statements themselves should also be true. If the arguments are not valid in themselves, then the question of evaluating whether the statements themselves are true or not does not arise. Consider the following argument:
If an alien spacecraft with yellow colour hovers over the North Island of New Zealand, one person will prematurely die
We have not heard about premature deaths so far; therefore an alien spacecraft with yellow colour has not hovered over New Zealand. (This is taken as a spoof on a theme from "The Hitchhikers' guide to the galaxy")
Well, we cannot find fault with the logical structure of the statements, but the statements do not make sense. This sort of a situation, where either the logical structure is false, or the sentences do not make sense, are referred to as unsound arguments. Sound arguments are those arguments where you get both logical sense  & that the facts are true as stated.

Reasoning by induction: the case of probability & sampling

In reasoning by induction, we also set up premises and conclusions but the premises and conclusions take the shape where premises are facts or observed phenomena. The observed phenomena from specific examples or specific instances then lead to generalised statements. The generalised statements are not absolute statements, we express the generalised statements using probabilistic language. Consider the following example:
P1: In our study, we found that those who had lung cancers were twice as likely to be exposed to environmental tobacco smoke than those who did not have lung cancer.
P2: Other investigators have noticed similar associations with other respiratory diseases as well.
C: We conclude that exposure to environmental tobacco smoke is a likely risk factor for lung diseases
Such statements are commonly expressed in the discussion or introduction section of studies, or theses. What's going on here? In the first two sentences (P1 and P2), the authors describe two specific observations. Then in the third sentence, "C", they make a generalisable statement, that based on P1 and P2, they can conclude that what they observed can be generalised to a wider truth (that is exposure to environmental tobacco smoke is a risk factor for lung disease). But in stating that, they do not write in absolute terms; instead they mention certain levels of uncertainty in the concluding sentence. This is the essence of inferential logic and this is the essence of inductive logic: we start with observations of the particular truth and from there, we extend to the larger or more generalisable truth. But because we are limited to specific observations, therefore, we cannot make statements in absolute terms. Consider the following statements:
P1: For each of the last ten years, we have experienced, using the averaged annual temperature trends, each subsequent year has been warmer than the previous one;
C: Hence, it is likely that the next year will be hotter than this year.
Here, we see that there has been an observation and then based on the trends observed over the past, there is a statement that it is likely that the next year will be hotter than the previous year. But how likely? This is a good question and we will reserve the meaning of this term till the next example. Consider the same or similar statement as above but stated in a different manner:
P1: Last year was hotter than the year before.