Sociology Research paper
RESEARCH QUESTION: How is race portrayed in news coverage regarding violence in the National Post?
KEY REMINDERS: You will choose articles that discuss violence from Aug.2014- Jan.2015 and please read through ALL instructions thoroughly, it tells you exactly what is expected.
SOC101Y – Quantitative Newspaper Content Analysis
Each student will be designing, conducting and writing up the results of a research project. The objectives are both to apply some of the sociological knowledge learned during the course, and to gain an appreciation for what is involved in the practice of sociology through the pursuit of an actual research project. The particular type of research project students will pursue is a quantitative newspaper content analysis. This document lays out the parameters for this year long project due at the end of March. It goes over the what, how and why of content analysis as a methodology. The content of this document offers a framework and instructions for the work students need to produce. While reading this whole document once at the beginning of the year is a very good idea, it is not expected you will understand all of its content right away. The skills and knowledge necessary will also be developed through weekly tutorials taking place between September and March, and digitally on Portal through your own TA’s Corner. Five small assignments will also be completed that represent various key aspects of the research design process. This document should be used as a reference to be consulted often, as well as the additional insights and advice from your TA and from Jenna.
This quantitative content analysis will not require students to deploy complicated statistical analysis, but it will involve conducting an analysis whose coding outcomes will be numerical values that can be represented in frequency tables or crosstabulation. Quantitative analysis, more generally, attempts to quantify data by numerically measuring the incidences of various themes of ideas in a chosen sample. While you have a broad latitude for the topic you will pursue, your content analysis ought to be a class analysis, race or ethnicity analysis, or gender analysis. It will be based on the analysis of a single or several days of publication of the Globe and Mail, National Post, and Toronto Star. Your ultimate research report you will submit for assessment at the end of March will include sections on your topic/literature review, on your designed research question and its rationale, on your operationalization/coding scheme, on data collection/sampling, on your findings, on your analysis of these findings and offering a tentative answer, on the strengths and weaknesses of your research project. Completing this project you will learn about the key role played by each of these aspects of the research design process, and how together can possibly lead to sociological
knowledge.
Your quantitative newspaper content analysis write-up in the form of a research report (ie NOT an essay) is due on March 30th. Three copies of the assignment must be submitted, through 1) Portal, 2) Turnitin.com AND 3) as a hard copy to Room 225, 725 Spadina Ave, no later than Monday, March 30th at 4pm.
The research report HAS to be between 7 and 8 double-spaced pages (not including title page, bibliography, and appendices), 12pt font (Arial or Times New Roman), with one inch margin all around. Note: turning in a longer report is not an option. Part of the difficulty associated with the research report writing process is summarizing your research in a concise way by making decisions about how to best use the space you have.
The outcome of your work needs to be written following the structure of a research report which should include the following sections, in order, and sticking to the page requirements:
1-Introduction (about 1⁄2 page)
2-Topic/Lit Review (about 1 page)
3-Research Question/Rationale (about 1⁄2 page)
4-Operationalization/Unit of Observation/Coding Scheme (about 2 pages) 5-Sampling/Data Gathering (about 1⁄2 page)
6-Findings/Frequency /Crosstabs (about 1 page)
7-Discussion/Data Analysis/Answer (about 1 page) 8-Reflection/Strengths/Challenges (about 1⁄2 page)
9-Conclusion (about 1⁄2 page)
Introduction (about 1⁄2 page)
In the introduction, you state what your research question is and offer an outline of the content of your research report. Introduction is there to frame your work. It’s there to help set expectations for your reader about what they are about to read.
Topic/Lit Review (about 1 page)
This section is to set the context of your eventual contribution. Discuss what is the broad sociological topic you are interested in, and then the more specific aspect you are proposing to investigate. You should pull from the “literature” which here means the course material. (note: you are allowed to draw from outside of course material, to clarify and add depth, but this is NOT required).
Research Question/Rationale (about 1⁄2 page)
This section is where you re-state the clear and specific research question you propose to investigate along with offering a rationale for the appropriateness of pursuing this particular question. Research questions are the foundations of research project, so here justify your choice. Research question, and the rationale, and variables you use. Explain to your readers why it is a good research question, how is it measurable.
Operationalization/Unit of Observation/Coding Scheme (about 2 pages)
This section is at the center of your research design efforts. Discuss what is your unit of observation. Then clearly lay out how you will measure your key variables looking at these units of observation. Do so by creating a detailed coding scheme that makes it clear how each category is constructed.
Data Gathering/Sampling (about 1⁄2 page)
Describe your data gathering process, including sampling. How many pieces did you gather, how did you choose them, describe the timeline over which this took place.
Findings/Frequency Tables/Crosstabs (about 1 page) Present your frequency table(s) or crosstab and describe it .
Discussion/Data Analysis/Answer (about 1 page)
Discuss the patterns and variations you are observing in your frequency table(s) or crosstab. Make sense of them to see what answer(s) they point to in regards to your research question. Note that since you will not be running the statistical calculations necessary to ascertain whether the patterns/variations you found are statistically significant, your discussion of your findings remain tentative, and not definitive. This is very important to remember.
Reflection/Strengths/Challenges (about 1⁄2 page)
This section is an opportunity for you to reflect on what you believe are the strengths of the research project you completed and what might be its biggest challenges.
Conclusion (about 1⁄2 page)
An opportunity to offer some concluding thoughts about how the research project revisiting the tentative answer to your research question. What contribution would your research project make to the “literature” once its findings have been verified?
This quantitative newspaper content analysis counts for 20% of your final grade. Beyond the “SOC101Y – Content Analysis” document you are reading right now, the principal sources of assistance are tutorials and the TA Corners on Blackboard. Please feel free to post any and all questions.
What is Content Analysis
Content analysis is the the study of recorded human communication. This includes detailed, systematic analysis of “text” to identify patterns or themes. As newspaper readers for example, we are perfectly justified in applying our individual worldviews to texts and enacting our interest in what those texts mean to us; in fact we cannot do otherwise. But as content analysis researchers, we must do our best to explicate what we are doing and describe how we derive our judgments, so that others—especially our critics—can replicate our results.
Content analysis is one of several important research techniques in the social sciences. The content analyst views data as representations not of physical events but of texts, images, and expressions that are created to be seen, read, interpreted, and acted on for their meanings, and must therefore be analyzed with such uses in mind. Analyzing texts in the contexts of their uses distinguishes content analysis from other methods of inquiry. Content analysis is essentially a coding operation. Coding is the process of transforming raw data into a standardized form. Making a judgment about an object according to a set of agreed upon dimensions (some are socially pre-determined (e.g. age) others you must create (e.g. engagement in learning). What we code for depends on our research questions, meaning what we are interested in.
Content analysis is a research technique for making replicable and valid inferences from texts to the contexts of their use. Examples of inferences:
• •
• • • • • • • •
One might date a document from the vocabulary used within it
One might infer the religious affiliations of political leaders from the metaphors used in their speeches
One might infer the readability of an essay from a measure of the complexity of its composition
One might infer the problems of a city from the concerns expressed in letters written to the city’s mayor office
One might infer the prevailing conceptualizations of writers and readers from the proximities of words in frequently used texts
One might infer editorial biases from a comparison of the editorial pages of different newspapers
One might infer the identity of the author of an unsigned document from the document’s statistical similarities to texts whose authors are known
One might infer the political affiliations of citizens from the TV shows they choose to watch
One might infer an individual’s propensity to engage in hate crime from the ethnic categories he uses in ordinary speech
One might infer the likelihood of war from the coverage of international affairs in the elite newspapers of neighboring countries
As a
from
provides new insights, increases a researcher’s understanding of particular phenomena, or informs practical actions. Content analysis is a scientific tool. Scientific research must also yield valid results, in the sense that the research effort is open for careful scrutiny and the resulting claims can be upheld in the face of independently available evidence.
technique, content analysis involves specialized procedures. It is learnable and divorceable just the personal authority of the researcher. As a research technique, content analysis
“Text” has a broad definition here that includes all sorts of social artefacts/human communications:
• Newspapers, magazines, books, journals
• Government or NGO documents/websites
• Sermons, speeches, debates
• Social media sites
• Television, radio/podcasts
• Police reports, laws, legal decisions
• Commercials, advertisements
• Paintings, logos, graffiti, murals
• Many others
Here are six features of texts that are relevant to definition of content analysis
• €Texts have no objective—that is, no reader-independent—qualities
• €Texts do not have single meanings that could be “found”, “identified”, and “described”
for what they are
• €The meanings invoked by texts need to be shared
• €Meanings (contents) speak to something other than the given texts, even where convention suggests that messages “contain” them or texts “have” them
• €Texts have meanings relative to particular contexts, discourses, or purposes. (Once an analyst has chosen a context for a particular body of text and clearly understands that context, certain kinds of questions become answerable and others make no sense. The same body of texts can therefore yield different findings when examined by different analysts and with reference to different groups of readers. For a content analysis to be replicable, the analysts must explicate the context that guides their inferences. Without such explicitness, anything would go)
• €The nature of text demands that content analysts draw specific inferences from a body of texts to their chosen context—from print to what that printed matter means to particular users, from how analysts regard a body of texts to how selected audiences are affected by those texts, from available data to unobserved phenomena (Content analysis infer answers to particular research questions from their texts. Their inferences are merely more systematic, explicitly informed, and ideally verifiable than what ordinary readers do with texts. Whether the analysts’ world makes sense to their scientific peers depends on how compellingly the analysts present that world)
The crucial distinction between text and what other research methods take as their starting point is that a text means something to someone, it is produced by someone to have meanings for someone else, and these meanings therefore must not be ignored and must not violate why the text exists in the first place. Text—the reading of text, the use of text within a social context, and the analysis of text—serves as a convenient metaphor in content analysis. Most content analyses start with data that are not intended to be analyzed to answer specific questions. They are texts in the sense that they are meant to be read, interpreted, and understood by people other than the analysts. Readers may decompose what they read into meaningful units, recognize compelling structures, rearticulate their understandings sequentially or holistically, and act on them sensibly.
So why do we analyze social artefacts?
To gain insight into what is deemed significant and made salient in human communications, understanding that these communications both reflect and help to shape various aspects of the social world. How individuals and groups imbue meaning in social artefacts which gives us an insight into how society is organized through a meticulous study and analysis of social artefacts. With new conceptualizations and an empirical orientation, contemporary content analysts join other researchers in seeking valid knowledge or practical support for actions and critique. However, unlike researchers who employ other empirical techniques, content analysis examine data, printed matter, images, or sounds—texts—in order to understand what they mean to people, what they enable or prevent, and what the information conveyed by them does. Content analysis is particularly well suited to answering the classic question of communications research: “Who says what, to whom, why, and how”. It cannot answer all questions, questions that may require field research, interviews, or experiments.
A form of unobtrusive research, or methods of studying social behavior without affecting it. It makes sense of what is mediated between people—textual matter, symbols, messages, information, mass-media content, and technology-supported social interactions—without perturbing or affecting those who handle textual matter. As Heisenberg’s uncertainly principle
tells us, acts of measurement interfere with the phenomena being assessed and create contaminated observations; the deeper the observer probes, the greater the severity of the contamination. For the social sciences, Webb, Campbell, Schwartz, and Sechrest (1966) have enumerated several ways in which subjects react to being involved in scientific inquiries and how these can introduce errors into the data that are analyzed:
• • • • •
•
Through the subjects’ awareness of being observed or tested
Through the artificiality of the task or the subjects’ lack of experiences with the task Through the expectations that subjects bring to the role of interviewee or respondent Through the influence of the measurement process on the subjects
Through stereotypes held by subjects and the subjects’ preferences for casting certain responses
Through experimenter/interviewer interaction effects on the subjects.
One of
is that it avoids those challenges research techniques involving human beings entails.
the strengths of unobtrusive research, such as the content analysis students will perform,
RESEARCH QUESTION/RATIONALE
Your research question is the foundation of your research project. A research question is a question which you do NOT already have the answer to, a question that is answerable through meticulous research. At its most basic, that is what a research question is. There are two reasons for content analysts to start with research questions, ideally in advance of undertaking any inquiries: efficiency and empirical grounding. Content analysts who start with a research question read texts for a purpose, not for what an author may lead them to think or what they say in the abstract. The pursuit of answers to research questions also grounds content analysis empirically. All answers to research questions entail truth claims that could be supported, if not by direct observation then at least by plausible argumentation from related observations. Formulating research questions so that the answers could be validated in principle protects content analysts from getting lost in mere abstractions or self-serving categorizations. Texts acquire significance (meanings, contents, symbolic qualities, and interpretations) in the contexts of their use. Although data enter a content analysis from outside, they become texts to the analyst within the context that the analyst has chosen to read them—that is, from within the analysis.
A research question is a clear, focused, concise, complex question around which you center your research. You should ask a question about an issue that you are genuinely curious about
Your research question should be focused. Research questions must be specific enough to be well covered in the space available. Your research question should be complex. Research questions should not be answerable with a simple “yes” or “no” or by easily-found facts. They should, instead, require both research and analysis on the part of the writer. A good research question is one that is appropriate to the space (one 7-8 page research report, not a book), time (written for a single course, not for a 4 year PhD), method (content analysis, not interview, fieldwork or experiments) and object of study (newspapers) of your research project. It also has to be a sociological research question, meaning investigating a sociological matter as defined by a research question that displays the sociological imagination. Your research question needs to
fulfill two more conditions, it must posit the variables you are intending on studying, and your research question cannot attempt to establish a causal relationship. No causality. But before we discuss causality, lets go over variables.
Variables
The idea of a system composed of variables may seem rather strange, so let’s look at an
analogy. The subject of a physician’s attention is the patient. If the patient is ill, the physician’s purpose is to help the patient get well. By contrast, a medical researcher’s subject matter is different: the variables that cause a disease, for example. The medical researcher may study the physician’s patient, but for the researcher that patient is relevant only as a carrier of the disease. That is not to say that medical researchers don’t care about real people. They certainly do. Their ultimate purpose in studying diseases is to protect people from them. But in their research, they are less interested in individual patients than they are in the patterns governing the appearance of disease—in essence, the patients are relevant only for what they reveal about the disease under study. In fact, when they can study a disease meaningfully without involving actual patients, they do so.
Social research, then, involves the study of variables and their relationships. Social theories are written in a language of variables, and people get involved only as the “carriers” of those variables. Variables, in turn, have what social researchers call attributes or values Attributes are characteristics or qualities that describe an object—in this case, a person. Examples are female, Asian, alienated, conservative, dishonest, intelligent, and farmer. Anything you might say to describe yourself or someone else involves an attribute. Variables, on the other hand, are logical groupings of attributes. Thus, for example, male and female are attributes, and sex or gender are the variables composed of these two attributes. The variable occupation is composed of attributes such as farmer, professor, and truck driver. Social class is a variable composed of a set of attributes such as upper class, middle class, and lower class. Sometimes it helps to think of attributes as the “categories” that make up a variable. For this research project, students should not seek to investigate more than two separate variables.
Example of difference between variables and attributes:
Social common social concepts: Female, young, gender, upper class, occupation, Asian, age, plumber, race/ethnicity, social class
Variables
Gender
Age Race/ethnicity Social class Occupation
Attributes
Woman, man
Young, middle-aged, old
Caucasian, Aboriginal, Asian, Latino(a) Lower class, middle class, upper class Plumber, carpenter, sociologist, lawyer
The relationship between attributes and variables lies at the heart of both description and explanation in science. For example, we might describe a university class in terms of the variable gender by reporting the observed frequencies of the attributes male and female: “The class is
60 percent men and 40 percent women.” An unemployment rate can be thought of as a description of the variable employment status of a labour force in terms of the attributes
employed and unemployed. Even the report of family income for a city is a summary of attributes composing that variable: $10,980; $35,000; $85,470; and so forth.
When studying more than one variable, you can study the relationship between them. If the researcher can establish an association between two or more variables, whereas the presence, absence or changes in one variable is linked to the presence, absence or changes in another variable, you can establish a correlation. Correlation, however, does not imply causation. Causation requires direction, it requires isolating that variable X caused the changes in variable Y. While students can certainly seek to investigate a correlation between two variables, they should NOT seek to establish a cause and effect relationship. Causality is something very difficult to establish solely through content analysis, and not suitable for this first year research project. So your research question CANNOT attempt to study the impact of one variable on another. It can, however, attempt to establish a relationship between two variables.
Advice: The best kinds of questions are clear, specific and with enough complexity or detail adequate for the assignment. Clear questions use proper grammar. They have been formulated carefully to reflect a great deal of thinking and analysis work. Specific questions go beyond the topic itself (eg. gender) and focus on the way the topic is being operationalized in the analysis of your specific (eg. gender role or stereotypes among a certain group(s) of people). When a question is too broad it says little about what you will do in your analysis. Often student research questions need to be reworded. In many instances wording is highly confusing due to using cryptic language, poor syntax, and over complicated language, it leaves one confused as to what you are asking. A well-worded research question is the best foundation to a good project. Questions appropriate for content analysis are direct and concise. They were not over complicated. These types of questions demonstrate that you understand the nature of the assignment and what exactly a content analysis can accomplish.
Advice: Your research question should not be causal. This means that you implying that X causes Y.
This could be questions such as “What type of family structure leads to more youth deviance?”, “What leads to rural people to being more
conservative than urban people?” or “What caused the Nicaraguan revolution?” These are all questions that are not appropriate to this research project, because they cannot be answered scientifically using the research steps you will undertake.
Advice: Remember that a content analysis is used to explore the content of various media (books, magazines, TV, film, newspaper etc.,) in order to discover how particular issues are depicted (or not depicted). At its most basic, content analysis is an exercise that involves categorizing aspects of the content in order to help us get a picture of patterns, if they exist. It is a good idea to compose your research question in such a way as to focus on representations/depictions in the media (ie your newspapers).
Advice: Lastly, when composing your research question keep thinking about how you will go about measuring/answering this question. If you cannot answer this research question by doing a quantitative content analysis of newspaper, you need to pick a different research question. It will also be useful to think about what parts of the newspaper you intend on analyzing, ie what will
A causal question is one which explores the effect of one thing on another and more
specifically, the effect of one variable on another.
be your unit of observation, since that will also inform which research question is most appropriate. Lets turn to this matter now.
OPERATIONALIZATION/UNIT OF OBSERVATION/CODING SCHEME
Units of observations are here which aspects of newspapers you will be analyzing. Keep in mind newspapers are composed of various types of content including headlines, stories, advertisements, listings, content promotion, comics, graphics, etc. The differences between these categories can be sometimes confusing. You ought to use only one type to be the focus of your content analysis. Conducting content analysis across several units of observation is not a good idea at this stage. It adds a lot of complexity to the research project not appropriate for a first year project.
To help identify what are stories, they are likely meet the following criteria:
1. Longer than two inches in length. On a standard-sized column it can be measured with a ruler. 2. Must be written in complete sentences with a central theme.
3. Must not be part of a paid advertisement.
4. Must be a complete story, not a promotional reference for a full story contained elsewhere. Stories are often confused with listings, which are clearly editorial in content. The reason for differentiating listings from stories is that listings often cover many different topics in one section, can have many authors etc, and most importantly, serve a very different purpose in the newspaper.
Stories Do Not Include
• Content promotion references that lead to stories elsewhere
• Sports tables
• Stand-alone photographs or graphics
• Death, birth, engagement, wedding or anniversary notices that are submitted TV and movie listings
• Stock price listings
• Weather maps
• Crosswords or comics
• Horoscopes
• Community listings or advertisement content, which must be paid for
Each of these can be analyzed separately from stories, but for this research project you should stay away from listings and content promotion as units of observation.
What does listings mean?
The expression listings has been used a few times. Listings are different than stories, they are not a paid advertisement, presented as a collection of numbers and other information and is not in complete sentences, such as:
• Listings are usually in smaller type, “small tables or graphs,” and often are seen in long columns.
• Frequently-seen listings include sports scores, stock prices, entertainment listings, church directories and similar information.
• Listings can include graphics and photographs such as the weather map.
• Statistics and numbers included in the text of a story are not listings. Listings must be a stand- alone feature.
• Be careful of paid-advertisements that might appear to be listings. Examples include: classified ads, personals, paid death notices and other obviously advertisement areas.
You also should not conduct your content analysis about “content promotion”. Content promotion is very broadly defined as anything that:
• Highlights a content item contained elsewhere in the newspaper.
• Highlights a content item that will be published in a future edition or on the website. • Promotes the newspaper overall using the brand name of the paper.
• Promotes any service or feature of the newspaper including circulation sales, classified ad sales etc.
Content promotion includes (but is not limited to):
• The main flag (i.e. the title) of the newspaper.
• Refers, skyboxes, rails and many other front-page bits about inside content. • All indices, front page or otherwise.
• House advertisements of any sort.
• All contact information having to do with the newspaper itself.
• All advertisements for any of the newspaper’s subsidiary publications.
Once you have chosen your unit of observations, you will now need to turn your attention to operationalization, the single most important step of the research design process. Operationalization is the key step that links your research question and your analysis. Operationalization is how will you measure/observe/notice/identify the elements (i.e. variables) in your research question? You need to assign values to your variables. What will count as evidence? What will you be looking for? Be as precise as possible. Indicate whether you will be looking at manifest content or latent content (it is highly recommended to use manifest content for this assignment since you are required to choose a quantitative content analysis).
Manifest vs Latent
Manifest content refers to concrete things that you can point to in the text – for example how many times a certain word appears, or how many different people are quoted. Latent content doesn’t have those same concrete cues and is usually the more subtle result of many different things. Writing style is a good example of latent content. It might be difficult to point to a specific word or sentence that makes a story “positive” or “negative” coverage but it is contained in the story nevertheless. In general, manifest content is much easier to measure than latent content. Latent content analysis is more complex and requires more caution.
Operationalization is the link between your research question and your data collection. It is how you will answer your question. Your operationalization of those variables is what you will actually be looking for, it is the designing of your coding scheme, of creating the categories through which you will classify your data. A researcher will want to ensure consistency so that she is coding for exactly what she wants to code for. Developing a set of rules helps the
researcher insure that she is coding things consistently throughout the process, is the same way every time. This also ensures that the researcher is not engaging in selective observations, noticing only information that fits in the researcher’s pre-conceived ideas of what they expected to find. A researcher must be open to find inconsistent information that does not “fit” their expectations or theories. Good operationalization will help the researcher distinguish between relevant and irrelevant information. Information that ought to be ignored because it does not speak to the research question one way or the other.
Even more difficult than deciding what you want to measure is creating categories (also called operationalization).
Whole books have been devoted to the subject but here are a few guidelines for this research project:
1. Categories should be mutually exclusive. Make sure that they don’t overlap.
2. Make sure there are meaningful differences between categories. Fine distinctions between categories cause confusion and make results unreliable. A good example would be asking people to categorize stories as: not local, somewhat local, primarily local or completely local. You can bet that coders will get confused between “somewhat” and “primarily” and would sit and wonder what a “completely” local story would be.
3. Make sure there’s a category for everything. If there’s no category, coders will try to force stories into inappropriate categories and distort results.
4. Only create separate categories for things you see often. Don’t create a new category specifically for a unique instance that only appears once. The goal is to be able to classify items 95 percent of the time. There will always be 5 percent of items that are odd – create an “other” category and use it judiciously.
These last two points may seem to be at odds. It’s a balancing act between having too few categories and miscoding items and having too many and creating confusion. The best answer is to test categories using a few items (stories, headlines, advertisements, photos, whatever you are analyzing). By coding a few dozen examples it will become clear very quickly if you have too many examples that don’t have a category or categories without any examples in them at all.
Recommended Number of Categories
If your research question has two variables, it is recommended not to use more than four separate categories for each variable (see example Origin/Geographical Focus below, although note that five categories were used for Geographical Focus, which is one more than recommended for this project). If your research question has only one variable, it is recommended not to use more than seven separate categories (see example Group Task Roles below).
In your research report you will communicate your operationalization through the presentation of your coding scheme. Each coding scheme should contain four elements:
1. The name of each category you have broken down your variable(s) into
2. A clear definition of each category
3. Illustrative examples of content that would go in each category.
4. It also presents examples that might seem to go in a given category, but actually belongs
in a different category and explains why
A partial example, involving element 1, 2 & 3 is offered here from researchers Benne and Sheats who coded the various roles discussants enact during group discussion:
Group Task Roles
Roles Typical Behaviors Examples
1. Initiator- Contributor
2. Information Seeker
3. Information Giver
4. Opinion Seeker
5. Opinion Giver
6. Elaborator- Clarifier
7. Coordinator
Contributes ideas and suggestions; proposes solutions and decisions; proposes new ideas or states old ones in a novel fashion.
Asks for clarification of comments in terms of their factual adequacy; asks for information or facts relevant to the problem; suggests information is needed before making decisions.
Offers facts or generalizations that may relate to personal experiences and that are pertinent to the group task.
Asks for clarification of opinions stated by other members of the group and asks how people in the group feel.
States beliefs or opinions having to do with suggestions made; indicates what the group’s attitude should be.
Elaborates ideas and other contributions; offers rationales for suggestions; tries to deduce how an idea or suggestion would work if adopted by the group.
Clarifies the relationships among information, opinions, and ideas or suggests an integration of the information, opinions, and ideas of subgroups.
“How about taking a different approach to this chore. Suppose we . . . “
“Wait a minute. What does that mean?” “Does anyone have any data to support this idea?”
“I asked Doctor Jones, a specialist in this kind of thing. He said . . .” “An essay in the New Yorker reported . . .”
“Does anyone else have an idea on this?” “Can someone clear up what that means?”
“I think we ought to go with the second plan. It fits the conditions we face in the Concord plant best . . .”
“Do you mean he actually said he was guilty? I thought it was merely implied.”
“John’s opinion squares pretty well with the research Mary reported. Why don’t we take that idea and see if . . .”
Benne and Sheats focused on several (this table displays seven) roles that small group discussants enact during discussion. While Benne and Sheats focused on group interaction instead of on media content, this format of coding scheme is a good one to emulate: category name, plus category definition, plus illustrative examples. They didn’t include examples that seem to go in a category but belong elsewhere, which would have been helpful.
Here’s another partial example of a coding scheme with element 1 & 2 this time, this one coding newspaper stories for the variable “origin” and the variable “geographic focus”.
Origin:
______ 1. Wire/News Service ______ 2. Staff
______ 3. Reader
______ 9. Unknown
Geographic Focus:
______ 1. Local
______ 2. State/Regional ______ 3. National ______ 4. International ______ 9. None
Origin/Source of Story
1. Wire/News Service: stories from the AP, Reuters or any other news service. We also include stories credited to another newspaper (not the home newspaper).
2. Staff: stories with or without a byline that are identified as coming from the newspaper. Includes “special to” and correspondents of the newspaper.
3. Reader: use only on either editorial pages where readers write columns or letters to the editor, or in cases where stories are specifically identified as being written by readers.
4. Unknown: use when the source of the story is not stated.
Geographic Focus
Relative to the paper being published, determine the general focus of the story. In case of doubt, think through the following questions to clarify:
1. Is a specific locality, state, region, or nation identified in the story?
2. Is the story significantly more interesting to state, region, national or
international readers?
3. Does the story seem to be tailored for people from a certain locality, region,
state or nation?
By answering these questions, the geographic focus should become clear. Remember that although mention of specific geographic areas is important to note, it can sometimes be misleading. For example, there may be mention of the United Nations in New York, but that doesn’t mean that it’s a New York story. Similarly legislation can be passed in Washington, D.C., which deals directly with a local issue. It’s important to consider what makes the story newsworthy, and, more specifically, why is it in this newspaper?
Good Coding Scheme Construction Requires Practice
It is very good practice to take 10-15 items from your newspapers and practice coding them to continue to tweak your coding scheme. Go through each item and think about how they were categorized – this is extremely helpful in making sure that you have laid out clear definitions. Once you feel like you’ve got a good handle on what you’re going to measure, and what the possible answers for that measurement are, test it. Ask someone else (or a few people) to read
five or 10 typical items and code them. Everyone should come up with the same coding categories 75 percent of the time. If they don’t, you may need to reduce the number of categories to eliminate overlap and make the definitions clearer. If that doesn’t help, you may be measuring something that’s just too unreliable. Note: Testing your coding scheme is an essential part of the research process. It is part of what differentiates conducting a scientific investigation instead of an ad hoc one.
Advice: You must have in mind the word ALIGNMENT. This means that all the stages are related and intertwined. The research process in this assignment and all disciplines follow very similar logic. Overall, it means that each section stems from the one before it, and leads to the next one. Your operationalization/coding scheme must flow out of your research question and what you are attempting to investigate and must prepare the way for data gathering and coding.
Advice: When your question is read, you should ask yourself, “ok, how will I measure this?” If you cannot answer how you would measure it, then it is NOT a good research question. Operationalization is how you will go about doing this. If you operationalize well, it shows that you have an idea of what your question is actually asking and actually capturing. In some ways, this section is to justify that your question is doable. Here you convince your reader that it is something that can be answered in a content analysis. If, for example, you start defining variables or topics that are not related to your question, this is a problem. Another problem is if the question gets restated or rethought in the operationalization. In the research question section if you say you will focus on portrayal of a particular issue, later you cannot now focus on the absence of a different issue. The operationalization needs to reflect the question.
Advice: Good operationalization is when it is in line with your question and tells exactly you are going to count/measure/observe. For example, in a question like, “how does this media source portray men from different class backgrounds?”, your goal of the question is not to prove they are from different class backgrounds. In the way you have asked this question, you are categorizing individuals portrayed from different classes, then classifying how they are portrayed. So, you must say how you will go about classifying characters into different classes. If you will do by focusing on the language used, then you will need to be very clear about what words/language will count as evidence of membership into different classes. Now language can be at the heart of operationalization. If you want to study complicated concepts such as patriarchy or racism, you will need to be very clear on how you will measure or capture those concepts. What will count as examples and what will not. Or, how will you identify patriarchy in the media source you are studying.
Advice: Fundamentally operationalization is about how a particular concept or idea will be measured in your media sources. It’s crucial to the research process. We need to see that you identified what makes something what it is and how you identified or counted that. Its criteria all researchers need to establish before making an attempt to collect data. For example, if you were looking at the term “systemic discrimination” you first would need to define this term using your course book or some other type of academic source. Then you could narrow your focus and describe how you would count this—so really many of your variables had two aspects that were important to explaining your variables (i.e. what does it mean and how did you count or measure it?). Maybe you’re focusing on systemic discrimination in two or three realms like within the
education system and the legal system. But what does that actually look like? What this indicates is that operationalization can have several levels, and each level makes the abstract concept more practical and measurable.
Once you have your research question operationalized, it is time to move to sampling and data gathering.
DATA GATHERING/SAMPLING
The next stage in the research process is data gathering (and sampling). While your quantitative newspaper analysis must be on content from the Globe and Mail, National Post, and/or Toronto Star from August 2014 until January 2015, you cannot possibly analyze all the sections from all the issues of the three newspapers over a six month period. Each student need to analyze a sample of all the available material. Sampling is about selecting a portion (a sub-section) of the whole population. Sampling is necessary only when analyzing the entire population is unfeasible or unreasonable. It is selecting items that will tell us something more general than the specifics items we’ve actually selected. Sampling fits under two broad categories, probability and nonprobability sampling. Probability sampling is when a subset of cases/items is chosen at random from a larger population. Also for it to be probability sampling all items in the population have an equal chance of selection, there is then an excellent chance that the sample so selected will closely represent the population of all elements. Probability sampling avoids researchers’ conscious or unconscious biases in item selection. Under these conditions, results of studies using this probability sampling can then be generalized to the whole population.
Results stemming from nonprobability sampling, such as convenience sampling, where researchers choose which subset of cases/items will be studied based on other criteria, cannot scientifically be generalized to the entire population. Nonprobability sampling is used primarily for reasons of convenience, cost, access, time or because the population cannot be easily identified.
In the case of your research project for this course, once you have delineated your ‘population’, you are free to choose between probability and nonprobability sampling, but you are encouraged to choose a nonprobability sampling due to the additional time and level of difficulty associated with probability sampling. However, whichever you choose, you must be aware of the strengths and challenges associated with your choice.
If you are using probability sampling, you need to identify all the items that fit your unit of observation within your population, you will want to number them. Once you do that, you can select your cases using a random number generator. If you are using nonprobability sampling, you can choose using other methods. In both instances, you need to lay out very clearly and explicitly in your research report how the sampling process was conducted, and what choices
It’s crucial to be able to specify what you mean when you use
particular terms in your projects. You can’t meaningfully answer a question without working out
some type of agreement about the concepts. In another sense, you can think about what you
consider a meaningful reflection of a variable you are trying to study. What exactly are you
looking at in your media sources that represents the concepts you talk about?
were made. Also describe difficulties associated with data gathering, and the time it took to accomplish this task.
Number of Cases to Analyze: Whether using a probability or nonprobability sampling method, it is recommend that you gather and code between 50 and 100 cases depending on which units of observation you are analyzing. These cases can come from only one of the three newspapers (Globe and Mail, Toronto Star, and National Post) or all three. It could come from several days’ worth of coverage, or only a few. That is up to you, based on your research question, your interests, and your sampling strategy.
Note: All the issues of the Globe and Mail, Toronto Star, and National Post from August 2014 onwards have been put aside for this course in the Periodical Room located on the 4th floor of Robarts Library. Of course you are also welcomed to procure your own copies if you prefer.
FINDINGS/FREQUENCY TABLES/CROSSTABS
Once you have gathered your data and coded it using your coding scheme, you will present your findings in a table. The format of this table will depend on your research question. If your research question is investigating one variable, you will need a single frequency table (such as the table 1 just below). On the other hand, if your research question is investigating the relationship between two variables, you will need a crosstab (such as table 2 below). Finally, if your research question is investigating two variables independently, not the relationship between them, you will need two individual frequency tables, one for each variable.
Frequency tables and crosstabs allows you to see how many cases (i.e., stories for example) you have in each category for each variable.
Here is an example where the variable “Geographic focus” was coded based on categories “local”, “state & region”, “national”, “international”, and “none”.
TABLE 1: Single Frequency Table
Geographic focus local
state & region national international none
Total
Frequency Percent
211 35.5 71 12.0 195 32.9 58 9.8 58 9.9 596 100.0
In this table, we see that about 36% of the newspaper’s stories have a local focus, 12% have state/regional focus, about 33% have a national focus, about 10% have an international focus, and about 10% have no geographic focus at all.
The data can address even more sophisticated questions if there are two variables being related, such as, what percent of the politics stories published are local? Or, in what geographic focus do newspaper rely heavily on wire services? You can use a cross-tabulation procedure for that
purpose. This will give you the count and percent of stories for each cell of the two variables, geographic focus and story origin (note: it is NOT a good idea to have any more than two variables in your research question. Research questions with more than two variables become far more complex to analyze which is beyond the scope of this course).
TABLE 2: Crosstab
Geographic focus and Story source crosstabulation
Geofocus
local
state®ion
national
international
none
Total
Source of story
Count
% within geofocus
% within Source of story
Count
% within geofocus
% within Source of story
Count
% within geofocus
% within Source of story
Count
% within geofocus
% within Source of story
Count
% within geofocus
% within Source of story
Count
% within geofocus
% within Source of story
wire staff 11 164 5.21 77.73 4.60 58.78
22 39 30.99 54.93 9.21 13.98
139 41 70.56 20.81 58.16 14.70
38 17 64.41 28.81 15.90 6.09
29 18 48.33 30.00 12.13 6.45
239 279 39.97 46.66 100.00 100.00
reader unknown 9 27
4.27 12.80 31.03 52.94
6 4 8.45 5.63 20.69 7.84
8 9 4.06 4.57 27.59 17.65
3 1 5.08 1.69 10.34 1.96
3 10 5.00 16.67 10.34 19.61
29 51 4.85 8.53 100.00 100.00
Total 211 100.00 35.28
71 100.00 11.87
197 100.00 32.94
59 100.00 9.87
60 100.00 10.03
598 100.00 100.00
The purpose of this section is to lay out your findings in the form of these tables, and describe your table(s) using complete sentences. Any analysis or attempts to make sense of the numbers should be reserved for the following section.
Discussion/Data Analysis/Answer
This section is the data analysis, which means examining your data and attempt to draw out what patterns and variations have been found and whatever conclusions on can offer. You make sense of those occurrences in light of your research question. What answer can you provide to your research question? Is it a straight-forward answer or are your findings pointing to a nuanced conclusion? Were any of the relevant data you accumulated inconsistent with this answer or with your conclusions? Data analysis is NOT a description of your coded observations – it is making
sense of what you found – interpretation is involved in an analysis not just description. Be sure not to overstate your findings – findings should be data-driven, not vice versa.
Note that since you will not be running the statistical calculations necessary to ascertain whether the patterns/variations you found are statistically significant, your discussion of your findings remain tentative, and not definitive. This is very important to remember.
In table 2, for example, some of the patterns are that stories with a national geographic focus rely most heavily on wire services, that most staff stories have a local geographic focus, that reader- generated stories vary in their geographic focus, and that international stories also rely heavily on wire services. Again, however, one would need to conduct a statistical test to see whether the differences in the cells of the table are significant. A good test is the chi-square test, which calculates how different the actual cell distribution is from the expected distribution. This test and other statistical calculations are beyond the scope of a first year course in Sociology.
Advice: Successful sections will be organized well, discussing the patterns and variations most relevant to your research question. Remember this is not simply about describing what you see, you need to “make sense” of what you see in light of your research question. This is the difference between description and analysis. Data analysis is ultimately about recognizing the central themes of your findings and making sense of them. Your job as the researcher is to spend time analyzing the evidence you have collected to try to answer the question you posed. And you need to communicate this effectively, provide evidence that supports your tentative answer (again before definitive answers require statistical calculations) that can be evaluated by your reader. In order for the reader to evaluate your analysis and answer, however, you need to cite evidence to support it so that your reader can decide for themselves if your argument is convincing.
Advice: Dont forget to talk about outliers in your data, the “this was not always the case” type of points. “Although a majority of my data shows…there were still examples of…” Did allows for the complexities and nuances of your data to surface. Simple analyses miss out on the richness of the data available.
Advice: This is the section where a clear operationalization, thought out coding scheme, and well collected data pay off with an interesting and nuance answer to your research question. In almost all exceptional research projects for this course, the answer to the research question is nuanced. You can discuss an overall trend or two or three different patterns. As is the case in social life, there are many exceptions to the rule, and many of the best projects will highlight that.
REFLECTION/STRENGTHS/CHALLENGES
This section is an opportunity to reflect on the months spent on this research project. What do you believe are the strengths of this research project you conducted and what are some of the challenges you faced. You can discuss any steps in the research project, but use this section to highlight for your reader what you might do differently if you knew when you started what you know now.
CONCLUSION
This section is where you re-state your research question and summarize the tentative answer you have achieved through your research. You should discuss how your research “fits” in the literature out of which your question emerged and what you would do next if you go this pursue this research further.
OTHER
Your research report should have a separate title page and should include a bibliography with the various references (ie textbook(s), each newspaper item you analyzed, and any outside sources you have used). You can also include an appendix for any relevant material to your research that does not fit in the sections above. A good idea would be including a handful of the newspaper items you analyzed (three or four) with a caption about how they were coded. Note that the title page, bibliography and appendix do NOT count towards the maximum length requirement of 8 double-spaced pages.
Is this the question you were looking for? If so, place your order here to get started!