California Polytechnic State University, San Luis Obispo
English Department
Dr. Johanna Rubba

LAST UPDATED 9/17/10

ENGL 495 Language and Gender
How to Read Research Articles

• College students who are majoring in the humanities rather than science or social science often find it intimidating to read articles which report empirical, quantitative research. This is understandable, since these articles are structured differently from novels or critical essays that, say, an English major is used to reading. Research articles are designed to be as short as possible; complicated methods and results are compressed into a few pages of summary and graphic displays such as bar graphs and tables. Students also find it contradictory that the researchers are trying to be objective, with all of their numbers and statistics, yet they claim that their results support a particular position or opinion, which does not seem objective. Students often have trouble seeing the forest (the overall structure of the article and its major points) for the trees (each number in a table; each of numerous possible explanations), while the author has in mind the best possible one throughout the discussion. This handout is intended to help you understand the structure of a research article, and to give you strategies for reading them that will make them easier to understand.
  
  
• It is important to understand why research articles appear in publications: It is not just to burden college students with difficult reading assignments. The intended audience of a research article is the writer's professional peers: in our case, other people (usually professors or people who work at research institutes) who study language and gender. Researchers publish their work in order to let their peers know of new discoveries or new ways of looking at issues that the field studies.
  
  
• I will use examples from various of our assigned readings to illustrate the typical features of research articles.
  
  
• Reading this handout will also help you structure your term paper well!
  
  
• Thumbnail overview: 
           You should always take notes while you read, or at least use a highlighter for crucial details and ideas.
  
  
• Research articles, especially in linguistics and psychology, generally follow a typical structure. This is especially true of quantitative/empirical studies (ones which collect and analyze relatively large amounts of data):
  • Introduction/theoretical background; statement of hypothesis, thesis, or research question; brief description of study
  • Description of  method
  • Presentation of results
  • Theoretical analysis of results
  • Often, but not always, a conclusion ends the article. Various things can be accomplished in a conclusion: it might be a summary of the entire article; it may restate the significance of the study to the field; it may suggest substantial revisions in theory; it may suggest further research.
    
    
• Recommended sequence in reading research articles:
  • 1. Read concluding sections at the end of the article for summaries of the hypothesis, method, results, and significance of the study.
  • 2. Read the introductory section, looking out for hypotheses/research questions and a look-forward summary of what is to come.
  • 3. Skim methods and results sections, but postpone close reading until after you ...
  • 4. Read the opening paragraphs of the discussion/analysis, looking for summary of results/method.
  • 5. Read the methods and results sections closely; visualize the study and its results.
  • 6. Return to the analysis/discussion section and read very closely. Keep in mind relevant theoretical concepts as you read; take notes.
    
    
• Strategies for reading research articles
  • The best thing to do when beginning to read a research article is to flip to the end. Look in the last few pages for subheadings or phrases along the lines of Conclusions, In conclusion, To conclude or language that hints of a summary of the preceding article. Read the paragraphs that follow such headings. You may not understand all of what is said, but the conclusion is usually the place where the writer lays out the whole study in brief. The conclusion usually also serves to state how strongly the data support whatever hypothesis/research question the researcher started out to study, so there will usually be some theory in the conclusion as well.  A good conclusion is a thumbnail sketch of the whole article, with emphasis on how well or poorly the data that the researcher collected answers the research question or supports the researcher's hypothesis.
    
    
  • Eisikovits' article ends with a section called 'Conclusions', by which she means, not the end of the article, but what we can conclude from her study—what her study suggests about the reasons for Australian adolescents' language behavior. She writes: "What all this would suggest is that for the girls the norms which they become increasingly aware of are in line with those of external social usage. [...] The boys, however, [...] appear to use non-standard forms to affirm their own masculinity and toughness and their working-class anti-establishment values" (p. 51). Her last paragraph begins "What we have then is very strong evidence for age and sex differences in the variety of Australian adolescent speech—differences which reflect different social and linguistic norms held by the two sex groups" (p. 52). (The ellipses in square brackets [...] indicates that portions of the original article have been cut out for this book. The editors of this anthology have shortened many of the articles.)
    
    
  • Some research articles do not have this kind of conclusion. Some articles just end with the last points of the analysis or discussion of the research results. If you are faced with this kind of study, then, instead of reading the end first, start at the beginning.
    
    
  • The next thing to do is flip back to the beginning of the article and read the first section all the way through. Look for several kinds of information in this section:
    • An introductory section. This places the study in the context of other research in the same subject area or on the same topic, citing others' research pretty heavily. This section states the problem or question the study addresses, and defends the study's significance (why it is worth bothering to do the study). Here one finds the hypothesis or research question of the study. A hypothesis makes a claim and states the expected outcome of the study. A research question does not make a claim or anticipate the results, but simply asks a question about whether the data are thus and so, or whether someone else's claim or "popular wisdom" is correct. Usually there is also what I call a "look-forward": a very short description of the study's methods and results, giving the reader some expectations as to what will follow.
      
    
    
    • Eisikovits begins her article by citing previous research by such scholars as Encell, McKenzie and Tebbutt, Shopen, Horvath, Wolfram and Fasold, etc. She does this in order to establish that there is a good basis in previous research for the question she is examining (Australian society has rigid sex divisions; sex divisions are reflected in language habits), while simultaneously pointing out gaps in the research which her study might fill: "we have little information about whether adolescent females behave like adults ... The identification of when—if at all—these perceptions and hence changes in linguistic behaviour become evident is one of the goals of this chapter" (p. 42).
      
      
    • Eisikovits' study answers a research question rather than proving/disproving a hypothesis. She has two research questions. The first is stated indirectly: "we have little information about whether adolescent females behave like adults" (p. 42). She is suggesting that one thing her research will do is find out how adolescent females behave. Her second research question is clearly stated as such: "The identification of when—if at all—these perceptions and hence changes in linguistic behaviour become evident is one of the goals of this chapter" (p. 42).
      
      
    • O'Barr & Atkins have a clear statement of the claim they wish to support with their research: "It is the thesis of this study that so-called 'women's language' is in large part a language of powerlessness, a condition that can apply to men as well as women" (p.377). In other words, their hypothesis is that what determines some stereotypical features of women's language is not their gender, but their powerless position in society
      .
  • Following the introduction we usually find a thorough but concise (meaning as brief as possible) description of the method used in the study — how many people ("subjects") were  observed, their demographic characteristics (age, sex, etc., but no names, or the researcher uses pseudonyms); what kind of tasks they were asked to perform; and how the data were analyzed — counted, statistically analyzed for significance, etc. Consider the opening of Eisikovits' "Methodology" section:
    
    
    • "The data for this study consist of more than fifty hours of tape-recorded conversation. The sample of informants was made up of twenty males and twenty females, equally divided into two age groups, a younger group in year 8 of secondary school, average age thirteen years eleven months and an older group in year 10, average age sixteen years one month [etc.] ... The three variables to be studied here are: 1. nonstandard past tense forms such as seen and done, as in 'He woke up and seen something.' 2. Multiple negation, for example: 'They don't say nothing.' 3. Invariable don't, for example, 'Mum don't have to do nothing.'  The occurrence of these forms is quantified using the paradigmatic Labovian model, that is: % frequency = number of occurrences of nonstandard form — total number of potential occurrences."
      
      
  • Next, the results of the study are set out. Often, but not always, this section is separate from the next (analysis/discussion). When separate, this section tends also to be very concise — almost compressed, hence very dense. This is where the numerical data is presented and the statistics given. Often there are tables, graphs, or other visuals that depict the quantitative results. If there is a separate results section, the significance of the results to the hypothesis or research question is NOT discussed.
    
    
    • Eisikovits'  Results section has handy summaries. Tables appear, showing the numerical results, including (at first) mysterious notation such as "N = 10" (meaning the number of girls/boys who produced the data) and "p = .001" (a figure that shows whether or not the data reach a level of statistical significance that indicates that they are sufficient to be taken seriously as evidence). It is better to read the prose under these tables before trying to figure out what they show. Eisikovits, for example: "From this table it can be seen that males and females differ considerably in their use of these forms. Among the female speakers there is a significant decline in use with age ... "
      
      
  • The descriptions of method and results are often the two hardest parts of a research article, because they are densely presented and often involve the description of complex tasks/test procedures. The reader has to use the information in these sections to visualize how the study went forward and what the findings were. These sections are usually short, but densely packed with information; the results section will contain numbers, percentages, reports of statistical significance. When you read them, do so slowly and build a clear picture in your mind of each step in the study as you read. Study tables, bar graphs, etc. until you have a clear picture of what they report. However, at least in the results section, authors usually state clearly what appears in tables and graphs, essentially repeating what the visuals show. All the same, it's important to force yourself to understand tables/graphs clearly. This is an essential part of general literacy.
    
    
  • At your current stage of learning, you can assume that the authors' claims about results are accurate, or rely on your teacher's comments evaluating them. If you become expert in a field like this, you will need to develop the skills to make such assessments yourself (which is the reason that psychology and social-science students are often required to take statistics classes).
    
    
• Following the results section, the researcher analyzes the results regarding how they relate to the hypothesis or research question: were the results what was anticipated? Was the claim of others or of "popular wisdom" supported or not? The researcher addresses theoretical issues here: it is the longest and most important part of the article. Often the researcher also writes about how the results might reveal that the issue is more complex than expected, how the results match or contradict other studies, suggests looking at the issue differently, and so on.
  
  
  • Eisikovits calls this section 'Discussion' and opens it with a summary of her results, given indirectly as 'why' questions: "The existence of such different patterns of usage of non-standard forms among the two sex groups poses an obvious problem: why should an increase in age bring about such different patterns of usage among male and female speakers? Why is it only the girls who decline in their use of such forms as they grow older whereas, if anything, the boys increase their usage of such forms?" (p. 47). Eisikovits is moving into the purpose of the discussion section, which is to propose reasons for the results, but notice how, in doing so, she also tells you the general nature of her findings: In her study, girls' usage of nonstandard forms decreased overall while boys' usage of them increased. Notice how well this correlates with the research question and purpose of her article as stated early in the article (see the example under "Research question or hypothesis" above.
    
    
• Many articles end with an explicit conclusion; some just finish with the final statements of the analysis/discussion. There may be a subheading with a title or the title "Conclusion(s)."
  
  
• O'Barr & Atkins give an excellent summary of their whole study on page 385, at the beginning of the section titled "Women's Language or Powerless Language"? They write "In the previous section, we presented data which indicate that the variation in [women's language] features may be related more to social powerlessness than to sex. We have presented both observational data and some statistics ... The speech patterns of three men and three women were examined. For each sex, the individuals varied from social statuses with relatively low power to more power ... " Thus they summarize their method as well as their results.
  
  
• Bibliography:
  • Eisikovits, Edina. 1998. "Girl-talk/Boy-talk: Sex differences in adolescent speech." In Language and gender: A reader, ed. Jennifer Coates, pp. 42-53. Oxford: Blackwell.
  • O'Barr, William M., and Bowman K. Atkins. " 'Women's language' or 'powerless language'?" In Language and gender: A reader, ed. Jennifer Coates, pp. 377-387. Oxford: Blackwell.