Thesis Definition Tutorial 2019
Simply defined, a thesis is an extended argument. To pass, a thesis must demonstrate logical, structured, and defensible reasoning based on the credible and verifiable evidence presented in such a way that it makes an original contribution to knowledge, as judged by experts in the field.
This Tutorial explains the Thesis Definition with best examples. And also explains types of the thesis with sample templates.
Among the many types of scholarly productions, the thesis is an oddity: each one is different, and there are no standard or generic constructions. From the start, it is good to have a solid idea of what a thesis is, and perhaps the best place to start a discussion of the thesis is with their purpose.
Types of Thesis
At Honors level, a thesis—strictly, a ‘minor thesis’ is a work of original research of approximately 10,000 words in length. For many students undertaking a minor thesis, it is the first time that they have conducted original research.
From my experience, one of the main struggles occurs in making the transition from ‘research consumption’ to ‘research production’. Minor thesis is closely supervised and, very often, stem from research that is of direct interest to the supervisor.
An Honors thesis is typically produced within a year alongside the demands of coursework. For the most part, they are assessed within the students’ department; note, therefore, that the readership is well-known and thus the writing can be tailored to fit the audience.
At the Masters Degree level, there are two types of thesis. One is a minor thesis, with length limits ranging from 10,000 to 25,000 words.
It is completed within one or two years alongside coursework, and usually requires one or two semesters of full-time effort. Much like those at the Honors level, the minor thesis is assessed within the department by a set of internal criteria.
The second type is a ‘Masters by research’ thesis of 30,000 to 40,000 words. It is much more substantial than those that are written by coursework students as it is the result of full-time research over one to two years.
This thesis is examined by experts in the field outside the department. In some departments, students first join the field by writing a Masters thesis; if it is considered to be of high quality and can be extended, it can be converted into a doctoral thesis.
A ‘Doctor of Philosophy’ is earned by the successful completion of a Ph.D. thesis. For Ph.D. students, the word limit of a thesis is 1, 00,000 words; most students write approximately 80,000 words.
The standard thesis structure has four parts: an introduction, the background, the core (for want of a better word), and a synthesis. Note how, as illustrated in the following figure, the sections are connected to each other. A conclusion responds directly to an aim, for example, and the background must directly foreshadow the core.
Some of these parts might contain more than one chapter, and the core might be more than half the thesis. Each of these parts has a distinct role.
Your thesis should be organized as follows.
1.An Introductory Chapter
Tell the reader the problem you are tackling in this project.
State clearly how you aim to deal with this problem.
Limit the scope of your study.
Sketch out how the thesis is structured to achieve your aim.
2. Background Chapters
Include in these chapters all the material required to lead up to your own work.
Ensure that there is a flow of narrative that explains why each topic is being discussed.
3. A ‘Core’ Account of Your Own Work
Begin with a formal statement of your hypothesis or research questions.
Follow this with an account of the methods you chose to test your hypothesis or answer your questions, and why you chose them.
Report the results of applying these methods.
You are now ready to pull the whole thesis together.
Discuss the implications of your results.
Draw strong conclusions backed up by your discussion.
Check that they respond to the aim stated in your introduction.
Things to consider:
Are you are blocked in your writing, or procrastinating? Do you understand why? If not, discuss it with someone.
Think about how your thesis will work as a narrative.
Decisions about the organization should have a rational foundation. Satisfy yourself that you have good reasons for your chosen thesis structure.
The thesis demonstrates authority in the candidate’s field and shows evidence of command of knowledge in relevant fields.
It shows that the candidate has a thorough grasp of the appropriate methodological techniques and an awareness of their limitations.
It makes a distinct contribution to knowledge.
Its contribution to knowledge rests on the originality of approach and/or interpretation of the findings and, in some cases, the discovery of new facts.
It demonstrates an ability to communicate research findings effectively in the professional arena and in an international context.
It is a careful, rigorous and sustained piece of work demonstrating that a research ‘apprenticeship’ is complete and the holder is admitted to the community of scholars in the discipline.
Aim and Scope of the Study
Your stated aim should have three characteristics:
It should follow as a logical consequence of the problem statement. You identify a problem, and your aim is to address it; as just noted, you have to be clear about what the problem is.
It should be singular. You must identify only one aim. This is not easy to do. Students often show magnificent ingenuity in stringing all the aims they want to include in the same sentence, as
The introduction explains what the thesis is about: the problem that the thesis is concerned with, the aims and scope, and the thesis structure. In some disciplines, it includes an overview of the findings.
An introduction is typically written for a wider readership than the bulk of the thesis and may use illustrative examples to help underpin the reader’s understanding of what you are trying to achieve. Such examples help to create a narrative that a reader can use as context for your work.
However, an introduction isn’t an essay—the only purpose it has is to introduce the research. You should outline the problem you have investigated, explain the aim of the research and any limits on the scope of the work, and then provide an overview of what lies ahead. Five to ten pages are ample.
Your introductory chapter should consist of five brief elements:
1.The context of the Study
Provide a brief history of the issues to date.
Situate your particular topic within the broad area of research.
Note that the field is changing, and more research is required on your topic.
2.Problem Statement (or Motivation for the Study)
Identify a key point of concern (for example, increasing use or prominence, lack of research to date, response to an agenda, a new discovery, or perhaps one not yet applied to this context).
Refer to the literature only to the extent needed to demonstrate why your project is worth doing. Reserve your full review of existing theory or practice for later chapters.
Be sure that the motivation, or problem, suggests a need for further investigation.
3.Aim and Scope
Be sure that your aim responds logically to the problem statement.
Stick rigorously to a single aim. Do not include elements in it that describe how you intend to achieve this aim; reserve these for a later chapter.
When you have written the conclusions to your whole study, check that they respond to this aim. If they don’t, change the aim or rethink your conclusions.
If you change the aim, revise the motivation for studying it.
Be sure to establish the scope of your study by identifying limitations of factors such as time, location, resources, or the established boundaries of particular fields or theories.
4. The significance of the Study
Explain how your thesis contributes to the field.
There are four main areas of contribution: theory development, tangible solution, innovative methods, and policy extension. One of these contributions must be identified as the basis of your primary contribution to the field.
In contrast to reports for the industry, theory development is an expected and required contribution; for PhDs in particular, it must be ‘original’.
5. Overview of the Study (or Structure of the Thesis)
Sketch out how the thesis is structured. Don’t confine yourself to a list of the chapters, but show how they are linked and that one section logically leads to another.
The background is the knowledge required before a reader can understand your research: relevant history, context, current knowledge, theory and practice, and other researchers’ views.
In the background, your purpose is to position your study in the context of what has gone before, what is currently taking place, and how research in the area is conducted.
It might contain a historical review. If the research is location-specific (an investigation of diet in low-income suburbs, for example, or an examination of how a dialect is changing) you will need to describe the study area and its characteristics.
If the research is technology- specific (such as a study of food packaging or the yield of a harvesting machine) you will need to describe the specifics of this technology and how it affects the questions you can ask.
The background usually contains a chapter reviewing current theory or practice and may include the results of preliminary experiments or surveys carried out to help you feel your way into the problem.
Experiments may also be used to establish benchmarks based on other work against which your work is to be measured, and these too form part of the background.
The core concerns your own work: your propositions or hypothesis, innovations, experimental designs, surveys and reviews, results, analysis, and so on.
The synthesis draws together your contribution to the topic. It will usually contain a discussion in which you critically examine your own results in the light of the previous state of the subject as outlined in the background, and make judgments as to what has been learned in your work; the discussion may be a separate chapter, or may be integrated with the detailed work in the core.
Finally, it is where you summarise the discussion and evaluation to produce conclusions. These should respond directly to the aim of the work as stated in the introduction.
Overview of the Study
The overview of the study should follow on logically from your statement of the aim. In other words, it is an annotated version of the table of contents. However, rather than writing it as a list of chapter and section headings.
You should write it in the form of interconnected sentences and paragraphs to ensure that the logic flow is clear to the reader. That is, it should be a synopsis of the storyline that the thesis follows. Here is an example of an overview of a study.
The significance of the Study
The scope is linked to the significance. A way to think of the significance of your thesis is to equate it with potential impact: Where do you think your study will make the most difference to current thinking? There are four primary lines of argument that may be used to establish the significance of a study.
First, it may advance knowledge in the applicable field; that is, it revises or creates new knowledge (for example, the results will extend what is known about the applicability of a theory, the results are widely generalizable, or, for qualitative studies, transferable to other contexts).
Second, a study may contribute to the solution of a practical problem faced by many others in the field (for example, the control of bacteria in food preparation, or the development of sustainable policies of food consumption).
Third, it may demonstrate a novel use of a procedure or technique (for example, a statistical test, a projective technique, or an instructional procedure).
Fourth, a study may contribute to part of a programmatic research effort; that is, when the results of the study are considered in relation to other studies, there may be theoretical or practical applications of major proportions. Each line of argument that is applicable should be pursued.
Data comes from sources and experiments described in earlier chapters. Only include data that is derived from a process that you have described, that is, the reader must understand where your data comes from. Describe the data fully.
Have clear criteria for inclusion and exclusion of data and results. These should be independent of what the data shows, that is, it is not acceptable to only include data that confirms your hypothesis!
Make sure you have used the right kind of analysis mechanism for your data. For example, tools or approaches for large data sets may be unsuitable for sparse or irregular data.
The conclusions in your last chapter must respond to this aim. Obvious? Each of the examiners of my own thesis pointed out that I had promised one thing, and delivered another. Over the months you work on your thesis, it is easy to forget the relationship of the introduction to the final conclusion.
When you have written the last sentence of your conclusions, go back and re-read your aim. If the conclusions don’t respond to the aim, you had better rewrite it; and don’t forget, you will also need to rewrite the problem statement that leads up to it.
Mechanics of Writing
Most people do all their own typing and word processing—a dramatic change from just a decade or two ago, and nowadays the old approaches to thesis writing (which included the use of professional typists and a great deal of writing by hand) are almost forgotten. All students are familiar with the challenges of using a word processor to produce a well-written short work such as an essay.
They are also familiar with the modes of writing that computers encourage: frequent revision, writing of sections in any order (or of several sections at the same time), ease of change of style and layout, and so on. The use of word processing has become universal, profoundly affecting the way research can be carried out and reported.
However, the task of writing an extended document such as a thesis is a very different process to that of writing a shorter work. Many students know the elementary features of word processors that are sufficient for a 3,000-word essay.
But not the more advanced features that help authors to maintain consistency of style and presentation over 50,000 words or more, a scale on which manual checking can become painfully laborious and where it is essential to have automatic maintenance of elements such as section and figure numbers.
Likewise, good presentation requires software that automatically maintains bibliographies; mechanisms that create indexes and tables of contents; tools for professional-looking illustrations; and strategies for keeping versions and back-ups.
Today, the most widely used general-purpose word processor is Microsoft Word or the OpenOffice equivalent; in the mathematical and physical sciences, many researchers use the more technically oriented markup-based LaTeX.
I do not explore the specifics of these word processors, but encourage you to use resources such as advanced guides and manuals to ensure that you are using them well—even an occasional revisit to an online tutorial can be surprisingly rewarding.
The fine details of the style of your thesis may not become settled until it is almost finished, and I strongly encourage that, in the early stages at least, your focus is on the creation of content rather than on how it appears on the page.
Nonetheless, right from the start, you need to write within the constraints of a style—for example, so that all headings of the same kind, such as chapter titles, are displayed in the same way.
Which format do you adopt? In principle, you have great latitude in your choices, so long as you are consistent in what you do. In practice, however, you should adopt a thesis style that is already in use in your department, and it may well be that your supervisor has specific requirements.
You should also be aware of field-specific guidelines, such as those published by the American Psychological Association. That said, some standard styles are less than eye-pleasing, and you may want to refine the appearance.
The key thing is that you use a style from the start so that you can easily change the appearance of the whole thesis if you want to do so.
Using a style means that, once you have established a pattern, you can easily stick to it and the reader will get the same message every time. For example, main section headings, wherever they appear in your document, will always be in the same typeface (font) and of the same size.
They should always be preceded by the same space separation from preceding text, and always be followed by the same space separation.
If the style you choose is clearly different from that of other headings, the reader will quickly understand ‘We are starting a new main section’ or ‘This is a sub-section within the section’. This is especially true if you use a style that is familiar to most readers in the field.
A thesis consists of several different parts that need to be tied together by a set of conventions. Without a standard format across the entire document, the work will appear random and unprofessional. For example, you should put all chapter headings on a new page, using the same style; that is, the same font and paragraphing.
You should give all major section headings a style that is different from that of the chapter headings. Captions to figures should all have the same style, but be different again from section headings and different from the main text.
All new paragraphs should begin with the same indentation (except for the first paragraph after a heading, which may have no intent at all), and so on.
All this will help your readers to navigate their way through your thesis. This styling is provided with templates, which govern the appearance and numbering of every element of a document.
After creating a style, you can generate a thesis structure, with a few empty chapters and perhaps some subheadings and so on. You can then use the style to generate a table of contents, and begin to get a sense of how the final thesis will appear.
As you proceed, you will use the table of contents, or other outline tools, to get a sense of the current structure of the thesis and where it may need revision—extra chapters, moving of material from one section to the next, changing how headings relate to each other in the hierarchy, and so on.
[Note: You can free download the complete Office 365 and Office 2019 com setup Guide for here]
Most word-processing programs include a facility for checking to spell. It checks every word you have typed against a dictionary built into the program. Do not ignore it!
Few people are infallible spellers or proofreaders and it is only rational to have oddities questioned. However, although the spell-check is good at picking up typographical errors, it can’t make decisions for you.
Typical problems are proper names (people’s names or place names), and words for which there are alternative spellings. In the case of proper names, the temptation is to tell the program to ignore its questionings and go instead to the next area of doubt.
This is a mistake: you should check any proper name the first time the spell-check comes to it and when you are satisfied that you have got it right, add it to the dictionary.
The second problem is words for which alternative spellings are permissible. The most important constraint here is that you be consistent. Before you start, determine your preferred spellings for these words, and keep to them.
On a related note, don’t rely on the spell-check to proofread for you; although it will pick up misspell words, it won’t distinguish between, for example, there and there, or affect and effect. Similarly, it won’t tell you if you have left a word out.
The name and year (or Harvard) system is the most popular reference style for the thesis. It works well for readers, because it names references in an understandable way, and also works grammatically. Another form of referencing is the numbered note, or footnote, system, which is used in many books.
Such notes are usually collected at the end of each page. When the page is printed out you will see the superscript number in the main text and the footnote text at the bottom of the page, separated from the main text by a dividing line.
If you have more than one reference number on the page, the footnotes are all collected automatically on that page, or instead, the notes can be collected at the end of a chapter or in a consolidated listing at the end of your thesis.
Note that tools for maintenance of references provide mechanisms for changing from one style to another, but, if you do make such changes, make sure the sentences with citations still parse correctly.
Wherever you put them, the notes have to be backed up by a consolidated alphabetical listing of all the references. A typical Ph.D. thesis will end up with two hundred or more references (yes, you will read that many, and understand them). Even a minor thesis may have thirty to sixty references.
Keeping track of these is a daunting task. For this reason alone it is worth learning how to use an effective bibliography tool. Word-processing programs can collate and maintain references, using bibliography software that builds up a catalog of references.
Each entry consists of elements such as, in the case of a journal article, the author (or authors), the title of the article, the journal name, year of publication, and publisher and place, together with an optional abstract and keywords.
This reference database can be used independently of your thesis as a way of recording the papers you have read, where they can be accessed, and your views or comments on the content.
Whichever system of referencing you use, the word processor offers the advantage that it helps you to maintain the match between the references cited in the text and the references appearing in your consolidated bibliography.
It helps to prevent you from inadvertently omitting references from your list that have been referred to in the text, and also helps to prevent you from retaining references in the list that are no longer referred to.
Tables and Figures
Word processing software includes rich mechanisms for assembling tables, whether of numbers, images, survey responses, or some other data. It is up to you to ensure that your tables have an obvious logical structure that readers can easily understand, and it also up to you to ensure that you make good use of the software, not simply use the defaults.
Consider the breadth of uses of Microsoft Word: school children drawing pictures; managers dashing off memos; journalists typing up articles; clubs producing membership lists; and on it goes.
It is hardly surprising that the default settings aren’t particularly well suited to the specific, niche task of thesis writing, and yet in many theses, the author has made no effort to improve the look of the work.
Tables in particular often seem to be poor, with uppercase headings as if the author is SHOUTING, the bad vertical alignment of values in the columns, illogical and inconsistent organization, and heavy lines everywhere.
Such tables are a sad contrast with the presentation in a typical professionally typeset journal, and yet the effort to turn one into the other may be only a few minutes.
I recommend that you find good models and imitate them. Remember too that tables are sometimes copied and used out of context—in slides, lecture notes, and so on—so they should be reasonably independent of the text. That is, take the effort to create captions and headings that make them at least somewhat comprehensible on their own.
Similar comments apply to figures, that is, graphs and artwork. If your word-processing package has a reasonably sophisticated graphics or line art system, you might consider using it to draw your figures.
This method has the advantage that you can edit the figures at a later date in the light of rewriting or alteration of the text. However, you will have to accept the limitations of the inbuilt graphics system, which may be significant.
Alternatively, you can draw all your figures using a separate specialized graphics package, and import them electronically into your text. You can’t edit them while they are in your word processor, but you can delete them, go back to the original version in the graphics package, edit that, then re-import.
If you are using charts generated by a spreadsheet program or a statistical package, these too can be imported into your text.
If you want to plot your data, enter it into such a package rather than attempting to create a graph with a line art tool— such plots are amateurish at best—and take advantage of the facilities that these packages provide for displaying data in a variety of useful formats and perspectives.
Don’t assume that the standard layout provided by the software is going to be best for your data; you may want to use colors, different kinds of labeling, different kinds of ways of representing quantities, and so on.
Tips for Writing Style
Most books on writing thesis deal with the art of writing and presentation. They usually deal also with the conventions that support good expression, namely grammar and punctuation.
This is not a blog about writing style, but rather about structure and coherence, and for advice on good grammar and so on you should look elsewhere. Likewise, I cannot deal here with all the errors that I have come across in my reading of draft thesis.
I recommend that you buy at least two writing books: a style manual and a guide to good writing. Read them thoroughly, and keep them on your desktop. Such books will tell you all you need to know; more than you can take in at first. We develop a writing style long before we start to write a thesis.
Some people can effortlessly write beautiful, clear, direct English that aids communication. Others have writing styles that hinder the reader: verbose, ungrammatical, turgid, labored.
Use of the Passive Voice
You will write more clearly if you use the active voice for verbs rather than the passive voice. Although it is not always appropriate, active voice should usually be your first choice.
Use of the First Person
As the structures of science—papers, reviews, disciplines, and so on—became elaborated during the nineteenth century, the idea developed that science was impersonal, that a scientist was a disinterested observer of the unfolding of new knowledge.
It followed that scientific researchers could not claim any personal credit (or even display excitement) over their discoveries when they came to report them. thesis, reports and scientific papers had to be written in the third person as if someone else had made the discovery. To remove the presence of the writer from the text, scientists resorted to using of the passive voice.
Some non-English languages don’t use tenses but rely on the context to indicate whether something happened in the past, is happening now, or may happen in the future.
In English, we have such a rich collection of tenses that we often get them wrong. This creates problems not only for students whose native language is not English but also for English-speaking students.
Long quotes from the work of others, say longer than thirty words, should not be designated by quotation marks and contained within the normal text, but instead should be presented as a separate block.
The whole block should be in slightly smaller type, indented, with space above and below. Quotation marks are not needed, and should not be used. And the quote should not be in italics.
Quotation marks are used to indicate that the enclosed words are the title of a chapter in a book, a paper in a journal, a poem, and so on.
Quotation marks were used to indicate colloquial words in formal writing or technical words in non-technical writing. However, it is now common to use italics for this purpose. (On this point, note that use of underlining is obsolete.)
After the first use of the word, the quotation marks may be omitted. Many writers extend this use by putting pet words or humorous expressions in quotes. It is best to avoid this as much as possible: it can become a bad habit.
We use link words to indicate the logic flow in a passage of text. They are of two kinds: conjunctions, which are used inside sentences to link clauses, and transitional words, which are used to link a sentence to the one that preceded it.
Many writers seem to use them interchangeably. This is a great source of confusion. Commonly used conjunctions are but, although, unless, if, as, since, while, when, before, after, where, because, for, whereas, and, or, and nor.
These words are used to link one sentence to the next. Commonly used transitional words are, however, thus, therefore, instead, also, so, moreover, indeed, furthermore, now, nevertheless, likewise, similarly, accordingly, consequently, and finally. We also make use of transitional phrases: in fact, in spite of, as a result of, for example, and for instance.
The confusion arises because some of the transitional words are commonly misused as conjunctions, as for example ‘in such reports the underlying theory used as a framework for the investigation might be reviewed however it is unlikely that new or improved theory would be developed’. The opposite fault is also common—conjunctions used as transitional words.
Appendices or annexes, as we can tell from the derivation of the two words, are things appended or tacked on to the main text of a report or thesis.
They do not participate in the main thread of the argument but have been included to support it in some way. They might establish the context of an item in the main text, or give the derivation of an equation.
They are often used as a repository for raw data. They might give a sample of a completed questionnaire (in this case the main text would describe how the researcher constructed and administered the questionnaire, and would summarize the results obtained).
How do you decide what you should include in the main text, and what you should relegate to appendices?
In most universities PhD candidates are given a word limit of 100,000 words, exclusive of appendices. Students often find that they have exceeded this limit, and a typical reaction is, ‘Well, I’ll have to put something in an appendix then’. Although this sounds a bit arbitrary, it does make sense.
Most of these issues concern the conduct of research, but one, plagiarism, concerns how it is presented. Plagiarism is a fundamental issue of academic honesty and instances of it can provoke strong responses.
An underlying cause is that academics’ reputations are based on what they have written so that when one person reuses another’s text it is perceived as a particularly threatening form of theft.
Plagiarism may, of course, consist of much less than the reuse of an entire chapter; one or two hundred words of direct copying would usually be regarded as a breach of ethical standards sufficient to trigger some form of disciplinary proceedings.
Nor does it matter whether the original material was an academic paper or, say, a web page; it would still be regarded as copied.
Moreover, considering that academics do a lot of reading and that your thesis will be examined by one of only a small number of true experts in your field, it is surprisingly easy to get caught.
Developments in search technology have made the detection of copied material very easy and accessible. Unless clearly attributed and put into quote marks, all of the text in your thesis should be your own.
Guidelines for Examiners
Does the candidate show sufficient familiarity with, and understanding and critical appraisal of, the relevant literature?
Does the thesis provide a sufficiently comprehensive investigation of the topic?
Are the methods and techniques adopted appropriate to the subject matter and are they properly justified and applied?
Are the results suitably set out and accompanied by adequate exposition and interpretation?
Are conclusions and implications appropriately developed and clearly linked to the nature and content of the research framework and findings?
Have the research questions in fact been tested?
Are the literary quality and general presentation of the thesis of a suitably high standard?
Does the thesis as a whole constitute a substantive original contribution to knowledge in the subject area with which it deals?
Research methods are the building blocks of the scientific enterprise. They are the “how” for building systematic knowledge. Let’s take a moment to think about knowledge. How do you “know” things?
One way you know things is through your own personal experiences. Even as personal experiences are rich in depth and detail, and create a lot of meaning in life, they are also quite limited in scope. If you try to generalize what is true for you, it is easy to overgeneralize and arrive at misleading conclusions for everyone.
Another fundamental way to gain knowledge is through the authority of others—your parents, teachers, books you have read, shows you have watched, news and articles from social media.
This “second-hand” knowledge includes many diverse sources, and often this knowledge is more than one step removed from where it originated. Life is made simpler by inheriting knowledge from humanity’s vast collection, instead of relying only on what you can discover for yourself.
In fact, most people spend years attending school to acquire a basic set of knowledge that seems relevant for a living and working in today’s world. Even though it can still take a long time to learn even a small proportion of the knowledge that is available, the efficiency of being able to gain a lot of knowledge in this way benefits us and allows us to continue to build and further what is collectively known.
However, not all information that is passed along is of equal value. While some of the things that we learn on the authority of others are based on scientific research, certainly there is much more information that is based simply on opinion, common sense, misinterpretation, or skewed information. It takes critical thinking skills to sort this out.
By learning about research, reading samples of research, and practicing research it is possible to expand your ability to think through knowledge and its acquisition in new ways.
When you learn the rules on which research is based, you are learning to generate knowledge in the tradition and practice of science. Regardless of the method selected, social science research methods are designed to be systematic and to minimize biases.
The goal is to produce findings that represent reality as closely as possible, overcoming some of the hidden biases that influence our conclusions when we are not systematic. As you will soon learn, research involves making many careful decisions and documenting both the decisions and their results.
Decisions are important throughout the practice of research and are designed to help researchers collect evidence that includes the full spectrum of the phenomenon under study, to maintain logical rules, and to mitigate or account for possible sources of bias. In many ways, learning research methods is learning how to see and make these decisions.
These days, research is everywhere. Whether you pursue an academic career or enter an applied field, research skills are likely to have a valuable application. In academic research, the application is obvious.
Academic writing nearly always describes research methods because academic work is judged first on the merits of its methods. Findings must be supported by how the information was collected, and whether it was thorough and unbiased, and addressed the research question appropriately.
Outside of academia, more and more careers call on people to understand data, to design ways to solicit feedback or information, to actually collect the information, and to figure out through analysis what the responses mean.
For instance, people in many fields and sectors of the job market want to understand who is using their products or services, how well they are carrying out internal or market objectives, how well their employees are performing, and who is interacting with their website or following them on social media.
It is possible to specialize in research and become an expert in answering questions of this type, but even knowing some basic principles of research can help you to make intelligent and meaningful contributions.
Knowing about research methods can also empower you in your personal life because it can make you a wiser, more critical consumer of all information. It can help you ask better questions about the information you encounter and ultimately act as a better-informed citizen.
The accumulation of knowledge through research is by its nature a collective endeavor. Each well-designed study provides evidence that may support, amend, refute, or deepen the understanding of existing knowledge. However, individual studies, no matter how compelling, are rarely enough evidence to establish findings as “fact.”
It is through the ability to find similar findings across studies, and the variability that studies may find when they ask questions in different ways and of different groups, that theories grow to be established as our working knowledge.
Much like language, scientific knowledge is a living conversation in which new studies and new inquiries allow what we know to grow and change over time.
The empirical approach to knowledge simply means that it is based on observation, direct or indirect, or in other words, on experience.1 In a casual sense, everyone uses the empirical approach in his or her daily life.
For instance, a person may notice that the mail comes at the same time every day, and begin to expect the mail to arrive at that time, based on experience. Or a commute may typically take 30 minutes, so a worker decides to leave for work based on previous experiences of the commute length.
As useful as observation and experience are, they can also be misleading, especially when they are not collected or reviewed systematically. In everyday thinking, we often make mental short-cuts that are helpful but not always accurate. We are all susceptible to biases in thinking that can cause us to overestimate the value of some information and underestimate the value of other evidence.
For instance, confirmation bias is a tendency to recall or favor the information or interpretation that fits with one’s existing beliefs. We give less consideration to information or interpretations that do not fit with what we already believe to be true.
Not all biases are related to one’s personal opinions. Take, for instance, availability bias, in which we tend to rely on the most recent information we have or only consider the immediate examples that come to mind on a topic, which may not represent all the information very accurately.
It is also possible to misinterpret observations. A teacher might observe that students become restless during a particular lesson and interpret their response as boredom.
The teacher may have misinterpreted the reason for the students’ restlessness. The time of day may instead be the reason that students are restless, not the dullness of the lesson.
Even if the lesson in question is boring to these particular students, the teacher might conclude that the lesson is boring for students in general. In fact, students who differ from the current students in ability level, background, or subject matter interests may find the lesson very engaging.
Researchers use the empirical approach as a way to avoid misleading results and poor interpretations. The key is carefully planning why they want to make observations, how to observe when to observe, and whom they want to observe.
Researchers create a plan or design, collect data in a systematic way, document their data collection, analyze data, and report the results.
Empirical approaches to research encompass all research design approaches, including experimental designs and nonexperimental design. They include qualitative and quantitative approaches to research design and analysis. Regardless of which method is used, researchers still need to answer these basic questions about their observations.
Let’s consider a case in which researchers wish to determine which teaching approach will best help students acquire math skills. After considering their own personal experiences with learning math and reviewing the literature on the topic, researchers learn that one effective approach uses “hands-on manipulatives.”
Manipulatives are concrete objects that can be viewed and physically handled by students to demonstrate or model abstract concepts. The researchers prepare a formal statement of research purpose, proposing to test “whether the use of hands-on manipulatives to teach Math Topic X will result in greater student achievement than teaching Math Topic X using a workbook-only approach.”
Now that the researchers have defined their research question more concretely, they must decide how to carry out the research. Deciding why to make particular observations is connected to sufficiently narrowing one’s research interest into a manageable project that has a clear research question.
Approaches will vary depending on the question that is posed, the opportunities the researchers have to conduct research, and what information already exists to address the question.
Planning how to observe is also connected to a matching approach and research question. As a part of the research design, researchers have to answer many “how” questions.
This is because research involves translation between ideas and measures. For instance, in the above example, the researchers have to decide how to measure “greater student achievement.” Figuring out how to measure something may be more or less challenging based on how abstract the concept that must be measured.
Consider the first example of bored students in the classroom. Boredom sounds easy to identify but it can prove hard to measure. A person can “look bored,” but how does the researcher know they are bored, and not simply sleepy? Perhaps the best method to measure boredom is to ask people to rate their boredom.
This self-assessment approach might not work as well to measure differences in student achievement. Measures vary from established, standardized instruments such as psychological inventories, to interview questions that the researcher writes and adjusts to fit the goals of the specific study. Other examples of measures include surveys, scales, direct observation of behavior, and objective tests.
Researchers must also decide when they will use the measures to obtain the most relevant results. If you want to study student achievement, successful weight loss, criminal re-offending, or smoking cessation, your results may have a lot to do with when you ask.
When researchers plan whom to observe, they first decide whether to observe an entire population (such as all fifth-grade students in a school district) or just a sample of the population.
If a sample is chosen, which is often the case, researchers decide how to select a sample that is not biased against any types of individuals or subgroups.
For instance, asking students to volunteer to take a mathematics lesson might result in a sample of students who are more interested in, or better at, mathematics than students in the entire population.
Such a sample might bias the results to look like better performance when compared to the population, which includes those who are less interested or more challenged by math. If a sample is biased and does not include all types of students who might be affected, the findings are less likely to align with reality.
Once observations are made, the researcher has data. Data may be in the form of numbers, which are analyzed statistically. This is called quantitative research.
Some data are not initially collected in numerical form but are translated into numbers. For instance, rating one’s health may be in terms like “excellent, good, fair, or poor,” which is then coded numerically and analyzed using statistics.
Other scientific observations are not reduced to numbers but are expressed in words. For instance, interview data may be described in a narrative that points out themes and trends. Such research is referred to as qualitative research.
THE ROLE OF THEORY IN RESEARCH
Research and theory are interrelated. They form a cycle that is part of the collective dialogue of the scientific enterprise. A theory is a unified explanation of observations, some of which may otherwise appear contradictory.
Typically, a theory will try to explain a relationship between two or more actions or things using rigorous criteria so that it aligns with logic and empirical evidence.
Good theories are also designed to be generalizable to groups or situations beyond those immediately studied and to be testable, so that when they are applied to other situations there are clear ways to determine if they hold true, revealing their applications and limits.
While theories often begin as ideas, they come to make up our working scientific knowledge when they are supported through the findings of repeated experiments or nonexperimental research.
Consider one of the most widely studied theories about learning: reinforcement theory. In this theory, positive reinforcement is defined as anything that increases the frequency of a response from an animal or individual. A common example of this is praising a dog as a reward for sitting on command.
In reinforcement theory, the praise is the positive reinforcement, and it has increased the frequency of the sitting behavior in the dog, a part of learning. At first, the reinforcement theory sounds obvious.
If you reward a behavior, you increase the behavior that results in the reward. In a way, it is self-defining. Why would something that seems so obvious be so carefully studied? Because it explains many apparently contradictory observations.
For instance, suppose an individual praises a dog regularly for sitting at first, but after a while, the person becomes lax and only offers praise for sitting every once in a while. What would you expect the result to be?
Common sense might suggest that the dog’s sitting behavior will decrease with the decrease in praise. However, we might actually observe an increase in the dog’s sitting behavior because the reinforcement theory indicates that intermittent reinforcement is, under any circumstances, more effective than consistent reinforcement.
Sometimes, answers that seem like common sense turn out to be right; sometimes the findings contradict the guess that common sense produces. Theories about behavior often start with common sense guesses, but research helps sort accurate guesses from inaccurate ones using empirical evidence and appropriate analysis of the data.
Reinforcement theory offers an explanation for why certain behaviors increase in their frequency. Without this theory, it might be more confusing to understand behaviors that would appear inconsistent.
Sometimes, research is used to test a hypothesis derived from existing theories. This is a deductive approach. It can also be considered a “top-down” approach because the theory precedes the data collection. Another way to think about the deductive approach is moving from a general principle to examining if it holds in a specific instance.
A hypothesis is formulated to be consistent with the existing theory. For instance, self-regulated learning theory proposes that a student’s level of cognitive engagement is determined by the goals they set.
From this, a researcher might deduce that when students know they will be tested again on the same material, those who have lower goals (e.g., a goal of getting 70% right) should ask for less feedback about wrongly answered test items than those who have higher goals.
If this hypothesis is confirmed by the study, it lends support to the underlying theory. Assuming the study is methodologically strong, failure to confirm a hypothesis calls the theory (or parts of it) into question, causing theorists to consider reformulating it to account for the discrepancy. This is an approach that aligns with quantitative research approaches.
In an inductive approach, research provides a “ground-up” approach, using observations and conclusions to formulate a theory.
It can also be thought of as moving from the specific to the general by adding up observations of specific events or people to notice patterns and using those patterns to develop theories that explain the events or behaviors observed. Qualitative researchers often take an inductive approach to theory construction.
Consider the above example of the dog receiving intermittent positive reinforcement. It is the observations that reveal a pattern in which intermittent reinforcement produces more effective results.
Even though qualitative and quantitative research tend to start from different points, the process of science means that research often employs both approaches. Inductive observations of patterns may result in a theory that is then tested using hypothesis testing in a deductive manner.
Deductive research may result in unexpected findings that the researcher then explores using inductive approaches. In truth, most types of research utilize both types of logic and benefit from going back and forth between data and theory.
Grounded theory is a methodological approach that closely links theory development and research through an iterative approach3 in which researchers alternate between theory development and research collection in an iterative fashion so that each step influences the overall process. Both theory and data collection are regularly revised as new observations warrant.
Students who are looking for a research topic for a thesis or term project would be well advised to consider a theory of interest. Testing some aspect of the theory can potentially make a contribution to the understanding of all aspects of behavior related to the theory.
In addition, students will find it easier to defend their selection of a research topic and write the introduction to the research report if it can be shown that the study has implications for validating and refining an important theory.
When thinking about theory as a basis for research, keep in mind that no theory of human behavior is universal. There are almost always exceptions to every rule.
This is why researchers usually examine trends across groups in order to test or develop theories. However, do not overlook the possibility of designing a study specifically to examine those individuals who do not perform as predicted by theory.
Understanding how the dynamics of their behavior differ from those of individuals who act in the way predicted by theory may help in refining a theory to take account of exceptions.
EXPERIMENTAL AND NONEXPERIMENTAL STUDIES
Studies may have designs that are experimental, quasi-experimental, or nonexperimental. The fundamental difference between experimental and nonexperimental study designs rests on the use of manipulation or treatment. Experimental designs introduce a specific treatment and then measure whether this treatment has an effect on some outcome.
When used in medical studies, the “treatment” in the experiment is often a medicine, but in most social science disciplines, experimental treatments tend to be “manipulations.”
For instance, an educational researcher may wish to find out if students remember basic math skills better if they receive online math tutoring. Like the medicine in the pharmaceutical study, online tutoring is a “treatment” that may or may not produce the desired effect.
To determine if a design is experimental or not, it helps to ask, “Will treatment be introduced in the study?” However, introducing a treatment is not sufficient to make a study an experiment. Experiments must meet a few other criteria.
Because experiments are trying to isolate the effect of the treatment, it is important to be able to rule out other factors. First, when possible, the researcher ideally measures how things were before the treatment was introduced.
Consider the reasons for doing so. It would not mean much to give a group of volunteers a medicine that treats seasonal allergies if it was not known whether they were experiencing seasonal allergies before they took the medication.
If the volunteers all felt perfectly fine after the medication, does it indicate that the medication is effective?
Not if they all felt fine before the medication as well. For this reason, a pretest measure is a highly desirable component of experimental designs.
If we continue with the allergy medicine example, we could find that a group had more allergies before taking the medication and fewer allergies after taking the medication.
Is this sufficient to say that the medicine caused the relief from allergies? What if the pretest was taken at a time with a high pollen count, and the medicine was administered to the group on a day with a low pollen count?
Everyone still reports fewer allergies, but there is another plausible explanation for the effect. Perhaps the difference resulted from the change in weather rather than the medicine.
One way to help reduce the plausibility of other explanations is to have a control group that did not receive the treatment, in addition to the experimental group that received the treatment.
A true experiment will divide participants into the control and experimental groups randomly. Without random assignment to groups, a study does not meet the criteria of an experiment.
In many medical studies, it is possible to make the study blind, which means that either the participant or the researcher does not know who received treatment and who did not.
In a study that is double-blind, neither the participant nor the researcher knows who is receiving the treatment. This element is not required to make a design an experiment, but it can reduce bias based on the expectations of an effect by either the participant or the researcher.
Blinds are easier to introduce when the treatment is a pill instead of a manipulation, but it is still sometimes possible to use in the social sciences. One instance might be an educational intervention in which those grading the assignments do not know which students received the experimental treatment and which were in the control group.
Experiments are generally considered a more definitive and rigorous form of evidence than nonexperimental studies because they answer more directly whether an intervention can help to explain an outcome while controlling for other explanations.
Some studies may be quasi-experimental by meeting some but not all of the criteria of an experiment. For instance, a study may include treatment, before and after measures, and a control and experimental group, but no random assignment.
This may still provide compelling evidence, especially if it can be argued that the control and experimental groups do not differ in substantial ways. At a minimum, a treatment, random assignment, and post-test measuring effects of the treatment can together comprise an acceptable experimental study.
When few of the basic experimental conditions can be met, it is time to consider other possible quasi-experimental or nonexperimental designs because, as the allergy example above showed, each element that is missing or inadequate within an experimental design weakens confidence in the reported results.
Sometimes, experimental conditions are not possible, practical, or ethical, and nonexperimental study is a better design.
In nonexperimental studies, researchers do not give treatments. Rather, they observe participants in order to describe them as they naturally exist, without introducing manipulations or treatments. Some common types of nonexperimental studies are surveys, polls, interviews, and observation.
These methods may ask participants about their attitudes, beliefs, and behaviors instead of observing them. Nonexperimental studies come in many forms. Generally, nonexperimental studies require a greater accumulation of evidence to support claims.
Let’s consider some additional examples of experiments.
Fifty students are divided into two groups at random. One group receives math instruction via a correspondence course on the Internet. The other group is given instruction on the same math skills using a traditional textbook. The purpose is to see if instruction via the Internet is more effective than traditional textbook instruction.
A psychiatrist identified 100 clinically depressed clients who volunteered to take a new drug under her direction. She also identified 100 individuals with the same diagnosis and similar demographics (i.e., background characteristics such as age and gender) to serve as controls.
The study was conducted to investigate the effectiveness of the new drug in treating depression.
The students in one classroom were observed for an hour each day for a week, and their inappropriate out-of-seat behaviors were counted. During the next week, the teacher provided extra verbal praise when students were in their seats at appropriate times.
During the third week, the teacher stopped providing the extra verbal praise. The results showed less inappropriate out-of-seat behavior during the second week of the experiment than in the other two weeks.
In Example 1, the group receiving the new type of instruction via the Internet is the experimental group, while the group receiving the instruction with a textbook is the control group.
This experiment is a true experiment because all four conditions for an experiment are met: there is a treatment, participants are measured before and after instructions, there is a control and experimental group, and participants are randomly assigned using appropriate random assignment strategies.
Not all experiments are true experiments,1 as illustrated by Examples 2 and 3. In Example 2, the experiment compared volunteers who were given the new drug with a group of individuals who were not given the new drug.
In Example 3, the treatment consisted of “extra verbal praise.” The classroom was observed before, during, and after the treatment, but there was no separate control group.
These leaves open the possibility that another factor affecting the entire class could also help to explain the shifts in behavior over the weeks of the study.
Even though experiments 2 and 3 were quasi-experiments by design, they are still referred to as experiments because treatment was administered.
Consumers of research cannot distinguish between nonexperimental and experimental studies based on the type of measure used. Measures such as paper-and-pencil tests, interview schedules, and personality scales are used in both types of studies.
The act of measurement itself is usually not considered to be a treatment. In fact, researchers try to measure in such a way that the act of measuring does not affect or change the participants. This is true in both experimental and nonexperimental studies.
By now, it should be clear that the purpose of an experiment is to explore cause-and-effect relationships (i.e., treatments are given to see how they affect the participants). The next topic describes how nonexperimental studies are also sometimes used for this purpose.
An experiment is a study in which treatments are given in order to observe their effects. When researchers conduct experiments, they ask, “Does the treatment given by the researcher cause changes in participants’ behavior?”
When researchers want to investigate cause-and-effect relationships, they usually prefer experimental over nonexperimental studies. However, for physical, ethical, legal, or financial reasons, it may not be possible to conduct an experiment.
For example, it would be unethical to learn about the effects of smoking by treating some participants with cigarette smoke by requiring them to smoke a pack of cigarettes a day for 15 years and comparing this group to a nonsmoking control group, whose members are forbidden to smoke for 15 years.
Clearly, the harm done to those forced to smoke would create ethical problems with such a research design. For this research problem, researchers cannot conduct an experiment.
Even if it were ethical to conduct such an experiment, it might not be practical because researchers probably would not want to wait 15 years to determine the answer to such an important question.
When it is impossible or impractical to conduct an experiment to answer a casual question, a researcher must settle for information derived from nonexperimental studies. To continue with the above example, one way to learn about the relationship between smoking and cancer might be to identify a group with lung cancer.
The researcher may then recruit a control group of individuals who do not have lung cancer but who are as similar to the experimental group as possible in demographics (basic statistical characteristics of humans that are used as identity markers, such as socioeconomic status, gender, or age).
And compare them on prior cigarette smoking as well as other key characteristics, such as diet, exercise, alcohol use, prescription drug use, and other measures that may help determine exposure to environmental toxins.
By choosing groups that are as similar to one another as possible except for having lung cancer, researchers can feel more confident in suggesting that the differences they do find between the groups may help to explain the difference in lung cancer status.
For instance, if the researchers find that smoking differentiates the two groups who are otherwise demographically similar, it suggests that smoking is a possible cause of lung cancer.
However, there are several dangers in making a causal interpretation from a non-experimental study. First, smoking and cancer might have a common cause. For example, perhaps stress causes cancer and causes individuals to smoke excessively. If this is the case, banning smoking will not prevent cancer, only reducing stress will.
Another danger is that the researcher may have failed to match the correct demographics or other characteristics of the two groups. For instance, perhaps the researcher compared the group with lung cancer to a group without lung cancer based on age, race, gender, alcohol use, and activity level, but did not take into account whether the people in the study lived in urban or rural areas.
If the group with lung cancer mainly resided in urban areas with heavy pollution, while those without lung cancer were all residents of rural areas, this weakens the argument that smoking was the key factor.
These types of problems would not arise in an experiment in which participants are divided by random assignment to form two groups. They would not exist because the random assignment would produce two groups that are equally likely to experience stress and equally likely to live in either rural or urban areas and, in fact, be roughly equal1 in terms of all other potential causes of cancer.
The above example of smoking and lung cancer illustrates a specific type of non-experimental study known as a causal-comparative study (sometimes called an ex-post facto study).
The essential characteristics of this type of nonexperimental study are (1) researchers observe and describe some current condition (such as lung cancer) and (2) researchers look to the past to try to identify the possible cause(s) of the condition.
Notice that researchers do not give treatments in causal-comparative studies. Instead, they only describe observations. Hence, they are conducting nonexperimental studies.
Although the causal-comparative method has more potential pitfalls than the experimental method, it is often the best researchers can do when attempting to explore causality among humans.
Note that when it is used properly, and the comparison groups are selected carefully, the causal-comparative method is a powerful scientific tool that provides data on many important issues in all the sciences.
TYPES OF NONEXPERIMENTAL RESEARCH
In nonexperimental studies, researchers observe or collect information from participants without trying to change them. Nonexperimental studies take many forms because they serve many purposes. The most common types of nonexperimental studies are briefly described here.
Causal-comparative research, described in the previous topic, is research in which researchers look to the past for the cause(s) of a current condition. It is used primarily when researchers are interested in causality but cannot conduct an experiment for ethical or other limiting reasons.
Surveys or polls describe the attitudes, beliefs, and behaviors of a population. To conduct a survey or poll, researchers draw a sample of a population, collect data from the sample, and then make inferences about the population.
For instance, a researcher could survey a sample of individuals receiving SNAP benefits (“food stamps”) to determine what types of food they purchase with this benefit. The results obtained from studying the sample could be generalized to the population (assuming that a good sample has been drawn).
If a researcher is able to interview everyone in a population (i.e., all individuals receiving SNAP benefits) instead of drawing a sample, the study is called a census. A census is a count (or study) of all members of a population.
This is easy to remember if you consider that the United States Census, completed every 10 years, strives to include every single person in the United States.
Unless a population size is small, completing a census study can be quite expensive. The anticipated cost of completing the 2020 census of all 300+ million United States residents is $12.5 billion!
While surveys usually include hundreds or even thousands of participants, a case study usually involves only one. For instance, some important theories in clinical psychology were developed from intensive one-on-one case studies of individuals.
In a case study, the emphasis is on obtaining a thorough knowledge of an individual, sometimes over a long period of time. Researchers do not confine themselves to asking a limited number of questions on a one-shot basis, as they would do in a survey.
In correlational research, researchers are interested in the degree of relationship among two or more quantitative variables. For instance, scores on a college admissions test and GPAs are quantitative (numerical), and because individuals vary or differ on both of them, they are variables.
If a researcher conducts a study in which he or she is asking, “Did those with high admissions scores tend to earn high GPAs?” the researcher is asking a correlational question.
To the extent that the relationship between the two variables is positive—that is, the higher admission scores correspond to higher GPAs—the researcher can assert that the test successfully predicts GPAs.
So far, studies are cross-sectional, meaning they are a snapshot of one moment in time. When researchers repeatedly measure traits of the same participants to capture similarity or change over a period of time, they are conducting longitudinal research.
For instance, a researcher conducting longitudinal research could measure the visual acuity of a sample of infants each week for a year to trace visual development.
Other examples include educational data, such as the Minnesota P-20, a statewide education data system that collects student data from pre-kindergarten to completion of postsecondary education to gauge the effectiveness of various educational programs and initiatives.
Experiments and all the types of research mentioned so far in this topic belong to the class of research called quantitative research. A distinctive feature of quantitative research is that researchers gather data in such a way that the data are easy to quantify, allowing for statistical analysis.
For instance, to measure attitudes toward Asian American immigrants, a quantitative researcher might use a questionnaire and count how many times respondents answer “yes” to statements about Asian Americans and then calculate the percentage who answered “yes” to each statement.
By contrast, in qualitative research, researchers gather data (such as responses to open-ended interview questions on attitudes toward Asian Americans) that must be analyzed through the use of informed judgment to identify major and minor themes expressed by participants.
Most published qualitative research is collected through semi-structured interviews in which there is a core list of questions from which the interviewers may deviate as needed to obtain in-depth information.
In historical research, information is examined in order to understand the past. Note that good historical research is not just a matter of developing a chronological list of so-called facts and dates. Rather, it is an attempt to understand the dynamics of human history.
As such, it is driven by theories and hypothesis. In other words, by reviewing historical evidence, such as newspapers or other archival documents of the past, researchers are able to develop theories that may explain historical events and patterns.
These theories lead to a hypothesis, which is evaluated in terms of additional historical data that are collected. Historical researchers may use qualitative methods (e.g., examining historical documents, using insight and judgment to identify themes) or quantitative methods (e.g., counting certain types of statements made in historical documents).
QUANTITATIVE AND QUALITATIVE RESEARCH: KEY DIFFERENCES
Quantitative and qualitative research differ in many ways. The names derive from the key difference in how research results are presented. Quantitative research results are presented as “quantities” or numbers, which are usually but not always presented through statistical analysis.
Qualitative research results are presented primarily through words, most commonly by interviewing people or observing settings and analyzing the data by reviewing interview transcripts and/or field notes.
In qualitative studies, researchers often identify themes in the data. Researchers must identify concepts that are consistently raised, and the range of responses in relation to those themes, which are written about in the analysis of the phenomena under study.
To arrive at different results, quantitative and qualitative studies begin with different plans and have different challenges in designing a study that will produce credible results.
Most research topics can be formulated into quantitative or qualitative research questions, but each approach has its strengths and weaknesses, and ultimately each type of research yields answers to different types of questions.
Quantitative researchers emphasize studies that seek to generalize and approach methods with goals of objectivity and standardization. Qualitative researchers emphasize and study questions in which observation or interview responses are from participants who are involved in an interpretive process.
Qualitative researchers often approach planning inductively and take an exploratory approach to questions that have not been adequately identified. They start by observing or formulating some well-designed questions to ask those involved in the area under study.
From this initial data collection, the researcher may develop additional questions as themes emerge, allowing them to ask more refined questions about specific dimensions that turn out to be important.
The strengths of qualitative research are in its ability to provide insights on interpretation, context, and meaning of events, phenomena or identities for those who experience them.
Results from qualitative work are often expressed in a narrative format because respondents provide answers in their own words, or can be observed in real settings over a period of time, instead of being limited to specific choices in a survey, poll, or experiment.
Qualitative research is good for research on topics or in settings where little is known, few theories exist, or the population is hard to reach.
Quantitative researchers often plan their research deductively. The most common approach is to evaluate existing theories on a topic and then try to apply those theories to a new or different scenario in order to see if the theories apply or require some adjustments when different conditions are considered.
Quantitative research can help to extend the generalizability of information that was discovered through exploratory research.
Most approaches to quantitative research aim for generalizability to a larger group but generally cannot reach all members of a group, and so the findings are based on a sample. The sample and measures have to follow procedures that improve the confidence with which results can be generalized to a larger group.
Qualitative and quantitative researchers examine previously published literature and include reviews of it in their research reports. However, quantitative researchers use literature as the basis for planning research, while qualitative researchers tend to deemphasize the literature at the outset of research, placing more emphasis on preliminary data.
When deciding on measures (sometimes called instruments) to use, quantitative researchers prefer those that produce data that can be easily reduced to numbers.
This includes structured questionnaires or interviews with closed-answer or quantifiable questions, such as questions in a multiple-choice format. In contrast, qualitative researchers prefer measures that yield words or capture complex interactions and behaviors using rich description.
This type of data collection is typically obtained through the use of measures such as unstructured interviews or direct, unstructured observations. It is always possible to reduce qualitative data to numerical data, but quantitative data cannot usually be expanded to provide qualitative data.
Even though qualitative data can be reduced to numbers, this data may not be particularly useful because it may not meet the criteria used in the statistical analysis of quantitative data.
Quantitative researchers often strive to select large samples to use as participants. Quantitative researchers are able to work with large samples because objective measures such as anonymous, objective questionnaires usually are easy to administer to large numbers of participants in a short amount of time.
Quantitative researchers tend to ask more targeted questions, and they hope to be able to generalize the results from the sample to some larger population, which may not be all people but maybe a defined subgroup of people—for instance, all public school teachers in a specific district.
Statistics can analyze a sample and estimate how well it represents a population. Because most quantitative research uses samples and statistical analysis, the sampling process is key to the credibility of the study.
In contrast, qualitative researchers tend to use smaller samples but spend more time with each participant in their study through extended, in-depth, unstructured one-on-one interviews and extensive observations over time.
Qualitative researchers are more likely to select a purposive sample of individuals. The purpose of the study and the approach to analysis are different, and this means that the criteria for sample selection are also different.
Researchers may select key informants and seek participants who represent the full range of those under study. If studying public school teachers, they might select several dimensions that matter for teachers’ experiences.
Such as the district in which the person is teaching, the grade being taught, the subject taught, or how long the person has been teaching in the school system.
And then ensure that each of these categories is adequately represented in the interviews or observations collected. Results often include verbatim responses from participants as evidence to back up analysis.
Because of the more intensive, sometimes participatory relationship to their research, qualitative researchers may choose to include relevant details about their personal backgrounds (such as having a mother who was a police officer).
In order to explain their position in relation to their research and account for possible sources of bias in their point of view or how research participants responded to them.
While working with the participants, qualitative research designs may allow for the possibility of making adjustments in the measures, such as rewording questions or adding questions based on earlier responses by participants. Quantitative researchers seldom make such adjustments during the course of a research project.
Instead, quantitative researchers plan their research in detail and follow the plan closely throughout the study because midstream deviations might be viewed as introducing subjectivity to the study.
For data analysis, quantitative researchers tend to summarize all responses with statistics and seldom report on the responses of individual participants.
Qualitative researchers, on the other hand, tend to cite individuals’ responses (e.g., by quoting individual participants) in the Results section of a research report.
Finally, quantitative researchers tend to generalize the results to one or more population(s), while qualitative researchers tend to limit their conclusions to only the individuals who were directly studied.
QUANTITATIVE AND QUALITATIVE RESEARCH DECISIONS
To understand some of the major differences in qualitative and quantitative research, consider this research problem: A metropolitan police force is demoralized, as indicated by high rates of absenteeism, failure to follow procedures, and so on.
Furthermore, the press has raised questions about the effectiveness of the force and its leadership. In response, the police commission is planning to employ a researcher to identify possible causes and solutions.
Take a moment to think: If you were the researcher, what approach do you think would best suit the problem?
A qualitative researcher would likely be interested in questions of “why,” and might formulate some preliminary ideas or questions that can help to uncover how members of the police department are making meaning of these events or their reasons for participating in some of the reported behaviors. They might investigate by collecting preliminary observations and informal interviews.
Questions would be exploratory and open-ended. The qualitative researcher would pay attention to topics that seem to come up with regularity—that is, the common themes that arise.
Questions or observations might also attempt to learn more about specific behaviors such as absenteeism and the failure to follow procedures.
Are these behaviors common or just a few people? Do they seem to cause or result in the police force being demoralized? The answers could help the researcher formulate additional questions and pursue those answers through more formal interviews or continued observations.
By contrast, a quantitative researcher would likely begin by reviewing the literature on topics related to organizational effectiveness, leadership, and demoralization in police forces or in other organizations. From this, the researcher might discover theories or combine ideas to hypothesize how leadership, police effectiveness, and morale are related.
To see if the theories were relevant to the current case, the researcher would then formulate some hypothesis that could be tested by conducting research.
For instance, maybe the researcher would read a theory about incentives that have been found to result in improved effectiveness in other studies of police behavior. The researcher could then carefully formulate a hypothesis and create a plan to select a random sample that allows the results to be generalized to the department.
The measure used might be a survey or an experiment that offers incentives to a randomly selected experimental group but not to a control group. In a quantitative study, the data will be used to either support or reject the hypothesis as potential explanations for the possible causes and solutions.
Each of these two research methods will need to carefully consider who is included in the study. Both qualitative and quantitative studies will have to decide how many people to include and how to account for meaningful diversity within the department.
Diversity may be not only differences in race or gender, but also age, length of time on the force, people with or without a record of absenteeism, or people in different places in the department’s hierarchy.
Even though both studies must account for these factors, they will have very different criteria. As a method of analysis, statistics are evaluated on the correct selection and size of a sample, without which the results cannot be trusted to represent the larger population.
The criteria in qualitative studies are not as reliant on the analytical requirements of the method, but the results will also be scrutinized for biases, and if it seems that groups were not included, it might limit the usefulness of the results.
Should the police commission select a researcher with a “quantitative” or a “qualitative” orientation? Clearly, the police commission faces complex decisions about how to improve the department. How would you answer the question in the first paragraph of this topic? What is the basis for your answer?
Some research questions inherently lend themselves more to a quantitative or qualitative approach. For instance, “What is the impact of terrorism on the U.S. economy?” lends itself to quantitative research because economic variables are generally numerical, and the question generalizes to a large group.
“What is the emotional impact of terrorism on at-risk healthcare workers?” lends itself to a qualitative approach because it focuses on subjective effects, a more interpretive and harder to quantify the issue. Note, however, that the second question could be examined with either qualitative or quantitative research.
When little is known about a topic, qualitative research should usually be favored. New topics are constantly emerging in all fields, such as new diseases (such as SARS), new criminal concerns (such as domestic terrorism), and new educational techniques (such as online adaptive quizzing).
On new topics, there often is little, if any, previously published research. In its absence, quantitative researchers may find it difficult to employ the deductive approach or to formulate structured questions.
How can a researcher decide exactly what to ask when little is known about a topic? In contrast, qualitative researchers can start with broad questions and refine them during the course of the interviews as themes emerge.
Theories might be developed from qualitative results and lead to a hypothesis that can be deduced and subsequently tested with quantitative research.
When the participants belong to a culture that is closed or secretive, qualitative research should usually be favored.
A skilled qualitative researcher who is willing to spend considerable time breaking through the barriers that keep researchers out is more likely to be successful than a quantitative researcher who spends much less time interacting with participants and relies on honest answers to impersonal questions.
Consider a quantitative approach when potential participants are not available for extensive interactions or observation. For instance, it might be difficult to schedule extensive interviews with chief executives of major corporations.
However, they might be willing to respond to a brief questionnaire, which would provide data that can be analyzed with statistics.
When time and funds are very limited, quantitative research might be favored. Although such limitations are arguable criteria for choosing between the two types of research, it is suggested because quantitative research can be used to provide quick, inexpensive snapshots of narrow aspects of research problems.
Qualitative methods do not lend themselves to the more economical, snapshot approach. They can be, and often are, carried out by one person in an economical way when that person does the observation, interviews, and transcription of notes, but the work is intensive and likely to take more time than quantitative research.
When audiences require hard numbers (which legislators or funding agencies, for example, sometimes do), quantitative research should be favored or, at least, incorporated into qualitative research, possibly as a mixed methods project. When someone says, “Just the numbers, please,” themes and trends illustrated with quotations are unlikely to impress.
For such an audience, one should start by presenting statistics, when possible. This might open the door to consideration of more qualitative aspects of the findings.
Notice that implicit in this criterion is the notion that both qualitative and quantitative approaches might be used in a given research project, with each approach contributing to a different type of information.
It is wise to remember that personal stories and narratives have a different power than statistics, but both have important contributions to make.