ENGL 2105 : Workplace-Based Writing and Research

Home | Introduction | Standard Syllabus | Contribute



Overheard on a street corner in downtown Atlanta: They make the decisions up in there, but they never come down here to see what's going on.

Listening is the key to communication. Effective communicators listen before they talk, think before they speak, and gather evidence before they draw conclusions. You can't craft an effective message if you don't know what you are talking about or who you are talking to or how their needs and goals and assumptions will influence how they interpret your message. In business, if you don't listen carefully to your customers, your suppliers, your partners, and even your competitors, you won't understand anything well enough to communicate effectively.

Listening is actually harder than it sounds.

How Listening Works

Listening seems simple enough, at least physiologically. Sound waves enter our ear canals and vibrate a membrane, ding, sound. But listening to people requires more than just hearing their words. (Listening doesn't necessarily mean agreeing with, by the way.)

Listening consists of the following elements:

  1. Message -- what the person is saying, their words as well as the way they are saying them and the context in which they are being said. If we are talking about speech, then their body language contributes to the message as well. If we are talking about text on a screen, then how the words are displayed and on what kind of screen (phone, tablet, laptop, desktop) contributes to the message as well.
  2. Interpretation> -- The message after the listener decodes it. What the listener hears is not always exactly what the speaker said. Most often what is heard is a variation of the original message caused by the hearer's prior beliefs, assumptions, prior knowledge, and circumstances. This is why you need to seek confirmation that you correctly understood what someone has said and that you accurately infer their intended implications. You can go very badly wrong if you assume you understood what they meant, or worse, anticipate what they are going to say and set out from there.
  3. Interference -- Anything that impedes the transmission or reception of the message. Interference may be environmental or psychological or ideological.
  4. Distortion -- Anything that alters the message so that it says something different from what it would have said absent the distortion. Sometimes distortions are the result of the listeners' interpretive practices, in which case you have systematically flawed (distorted) interpretations that alter the message itself, like when something is quoted out of context or heard in a negative (or positive) way when that wasn't the message's intention (nor the speaker's fault; that would be a broken message). Some distortions are just memory effects, unintentional changes caused by the fact that human memory is nothing like a hard disk.
      Some common causes of distortion:
    • ego -- self-centeredness (egocentricity) wanting to dominate or control the conversation leads people to hear to their partner's disadvantage.
    • current psychological state -- we tend to interpret less charitably when hungry or tired or frustrated
    • status -- you will hear the same thing differently depending on whether you are among your own people or those you perceive to be "other." This could be class, or economic, or even a geographical difference.
    • stereotypes -- assumptions and expectations resulting from generalizations about a group of people applied to a specific individual. You will interpret the very same words differently depending on how you perceive who is saying them and how you relate to the group you are identifying the speaker with. If you separate an individual from a group you will hear their words differently.
  5. Encryption -- The message is hidden in what was said. Often encryption is intentional, when the speaker wants to ensure only the intended recipient will get the message. But sometime we inadvertently encrypt a message, hide it from ourselves in the form of self-delusion, or from others in the form of deception, a lie to avoid unpleasantness, or a fabrication to save face, or a euphemism to avoid vulgarity. There are many ways, intentional and not, to hide what you think in what you are actually saying. Sometimes people resent encrypting but do it anyway, like when one's beliefs run counter to the dominant culture or the imposed will over more powerful people.
  6. Noise -- irrelevant sounds/information. Sometimes what you take for noise is actually the message. Sometimes the message is noise -- disinformation.

    To understand someone you have to eliminate noise, minimize interference, and be conscious of your interpretive activities: Focus on the person speaking, on their body language, on their words, on their tone, be aware of the circumstances and how they might be interfering with your reception of the signal or distorting the signal itself. You also need to pay attention to how you filter what the other person is saying. You can't eliminate the filtration process entirely, but you can minimize distortion. Be selfless, non-judgmental.

    We could extend that list and elaborate the examples, but that's enough to say that we don't just hear what is being said when we listen. We decode and encode and unless we are aware of our biases and assumptions and paying careful attention, we may let our encoding filters distort a message.

    So, ask yourself, "Am I hearing what they are actually saying or am I bending their words to affirm my assumptions, desires, and goals?"

    How to Listen When Conducting an Interview

    A large part of human centered research is talking with people in both structured (interviews) and unstructured (conversations) ways. Regardless of the setting or your goals, you want to listen when someone else is talking, whether you asked the question or just happen to be a part of what's going on.

    To really listen to someone you should do the following:

    1. Pay attention -- focus on the person speaking; cell phone off or down; lean in
    2. Create and maintain goodwill
      • Goodwill is wanting for someone else what they want for themselves or what is best for them. You create goodwill by being nonconfrontational (when you don't have to be), accepting, engratiating, and genial, good natured. Take an interest in the things that seem to interest them. Ask questions. Follow up on some of the answers; let others hang or drop. Find shared interests to broaden the conversation. Focus on them, not on yourself.
      • Demonstrate respect -- people communicate openly only if they trust who they are talking with. Respect builds trust. There is nothing more important than indicating to the person talking that you accept them as equally human, equally valuable. You don't have to agree with them, but you have to show that you understand what they are saying and why they are saying it and that they have as much right to say it as you have to disagree with it.
      • Validate -- validation is the feeling that your contributions to the world are appreciated, valued. What everyone wants more than any thing is to feel like they contribute something valuable and to have others recognize and appreciate them. If you want someone to communicate freely and openly, you need to make them feel like their contribution matters to something imporant. To make anyone feel good, appreciate what they do; if they don't do anything (or what they do isn't relevant), appreciate what they value or, if all else fails, what they have.
      • Suspend judgment -- don't analyze or infer or evaluate until you are sure you heard what they meant to say and keep in mind that we don't always say what we mean or mean what we say, sometimes through lack of skill or nervousness or lack of forethought, but also of course sometimes through guile. Remember that why a person says something can inform what they mean but you can't discount what they said merely because you suspect their motives.
      • Seek clarification -- Don't assume agreement on the definition of key terms. "When you say 'expensive' do you mean $100 or $10,000?" When you say "This is hopeless", do you mean "success is impossible" or "I'm tired and frustrated and I want to stop"? Often, to understand someone you need them to keep talking. Try to draw your interlocutor out by saying things like, "That's interesting. Tell me more." "Do you remember a time when..." "What would it be like if..."
      • Seek elaboration when informant's words inspire further (relevant) questions
      • Reflect -- when you think you know what their message is, say it back to them in the least annoying way possible, either by literally repeating their words or by drawing a clear and relevant inference. "So, what I'm hearing you say is..." "Am I right, then, in thinking..." "It seem like you are feeling..."
      • Summarize -- if the conversation is lengthy and includes different sections, summarize before you transition. Seek confirmation of agreement at the end of each section and at the end of the conversation.
      • Share -- don't just nod and say, "If I hear what you are saying..." Offer something of your own experience in return. Mirror, commiserate, re-assure, share. Don't play "I can top that" in an effort to share, and don't inadvertently redirect the conversation. Let the interlocutor lead but be visibly and audibly present.
      • Talk less; Listen more.

    How to disagree

    As I've said, listening to someone doesn't necessarily mean agreeing with them. When it comes to workplace-based research, most of they time your job is to identify what people think, not change their mind or argue with them. So in that sense disagreeing is irrelevant to workplace-based research.

    Still, disagreements can be important learning moments and when it comes to identify research topics or garnering resources to do research, you may find yourself disagreeing with others.

    Here are some guidelines and recommendations worth following when you find yourself disagreeing with someone:

    1. Check your emotions by acknowleding them to yourself--denial and supression don't work.
      Don't take ideas personally.
    2. Make certain you heard them correctly.
      Paraphrase what they said, and ask if you got it right. Or ask a question that indirectly asks them to confirm what you thought you heard.
    3. Identify the key words and make sure you agree on the definitions.
    4. Decide whether or not disagreeing openly will be worthwhile.
      Locate the shared goal this disagreement impedes.
      Once you know there is a shared goal at stake, decide if now is the right time to have this disagreement.
    5. Locate the issue, the specific point of contention.
      If they say, "It is cold!" and you say, "It's sunny!"; you haven't agreed, but you haven't actually disagreed with each other either because one of you is talking temperature and the other is talking atmosphere. Different topics, so no direct disagreement. If one of you says, "It's cold," and the other says, "It's 55°", that may be either an agreement or a disagreement. If both of you think 55° is cold, then you have a shared understanding and a specific goal. But if you have different cold thresholds, then "It's cold" and "It's 55°" is a disagreement about what "cold" means. If that's the case, there's no point in disagreeing because the sensation of cold is entirely subjective. Arguing about whether or not it is 55° is also a waste of time because 55° is an objective measure. A thermometer will settle the disagreement.
      To have a useful disagreement about the temperature, you need to be able to control the temperature and there needs to be a shared goal for the control. Then it's just a matter of finding evidence to identify "the best" temperature. If you can't find "the best temperature," then you need to either find different places to work or you need to negotiate, find a temperature that's least objectionable for both of you. You both need to give something up in order to gain something in common.
      Learn to distinguish misunderstandings from disagreements, differences of perspective from actual issues.
      If you have a common goal, understand each other, share the same definitions and recognize the same kinds of evidence, and have identified the issue, then a useful argument is possible. If not, change the subject.
      Arguments fail if both participants leave believing more strongly what they believed at the beginning and are none the wiser about either the issue or the other person's point of view.
      If you are wrong, then losing an argument is the best possible outcome.
      If there is no right or wrong (or one person is unwilling to accept evidence), arguing is a waste of time.

Good Question


The foundation of workplace-based writing and research is a question that can only be answered by input from real people. They don't have the answer, but they do have the data upon which you can build an answer. How you frame the research question is critical to your success. To conduct workplace-based research you need a testible hypothesis.


A statement the validity of which can be tested by a reliable measure, and the test independently replicated. In a deterministic setting, where the variables can be precisely measured and controlled, a proven hypothesis enables one to accurately predict an outcome.

The more uncertain the future, the more bearish the markets: if you can define "uncertain" in such a way as to identify moments of uncertainty in the historical record and then observe market fluctuations at that time, then you would have evidence to verify your hypothesis. At that point, when an moment of uncertainty arrives, you could make stock market decisions based on the expectation of sizable declines coming soon.

In a stochastic setting, where, control is imperfect at best, a proven hypothesis enables a probabilistic decision. We can't predict an outcome, but we can have a level of confidence consistent with the likelihood of a range of possible outcomes.

Let's say the principle that uncertainty leads to bearish markets is basically sound in the sense that the past has more examples of it happening than examples of it not happening and the number of events can't be explained by chance alone. Good so far, but how do we measure "uncertainty"? Uncertainty isn't purely subjective, like pain, but it isn't objective like the density of H2O as a variable of ambient temperature. Since we can't measure "uncertainty", the best we can do is look for occurrences in the past that were followed by drops in the markets. Then if an event similar to the our past examples were to happen, we could hypothesize a drop in the markets. If the market dropped, we would have more evidence to support our hypothesis, but we couldn't create uncertainty in order to test our hypothesis, and since we can't be certain of how uncertain one uncertain situation is compared to another uncertain situation, because we can't measure uncertainty, we can't predict accurately. We might even be mistaken about the similarity of two events, one being uncertain, the other only look uncertain, for example.

In an world where the variables are multiple and dynamic and nothing can be controlled, the best decision is one based on the best research available at the time. We can't predict an outcome or even rely on a probability or even a range of probable outcomes. We can at best understand the situation and decide based on what we know, experience, intuition, and insight.

Hypothesis: The electric fuel cell car will replace the combustion engine car by 2025. How could we test the validity of that hypothesis? Notice the time frame. If we didn't specify a date, the hypothesis would be purely speculative. Some day isn't helpful. We had electric cars before we had gas cars. What does that past fact tell us, if anything, and how are the relevant conditions today different from those in the past? What related technological, political, customer preference, changes seem immanent?

What makes a good hypothesis?

  1. Identifiable and measurable variables.
  2. An experiment you can control sufficiently to determine that a change in the independent variable will lead to a measurable change in the dependent one.
  3. A quantifiable relationship between variables.
  4. Knowing the relationship between independent and dependent variable helps you do something important.
    (The infamous "So what?" test.)

I have an economist friend who said he once had a student propose the following omnibus hypothesis: the easier the day's SUDOKU puzzle, the more optimistic the investors. The more optimistic the investors, the more bullish the market. Since the puzzle's get progressively harder during the week, the market will drop at the end of each week. So, he hypothesized, as the SUDOKU gets harder, the market declines, and vice versa. Could this student get the necessary data? Given what data he could get, could he prove the hypothesis one way or the other? (You can compare Monday closing numbers to Friday closing numbers, and get a simple graph. Sure. But can you measure the "hardness" of given SUDOKU? Can you identify a weekly trend towards harder? Can you prove that all investors do SUDOKU, that they invest after they do the puzzle (if they invest before then no influence is possible for that puzzle on that day), that they do the puzzle and invest every day, that people who don't play SUDOKU invest differently? Etc..) As you can see, some of the relevant data is available, but a key element, the "hardness" of a SUDOKU puzzle would be difficult to establish and even if that problem could be overcome, the number of variables between hardness of puzzle and optimism of investor are too many to handle, to say nothing of establishing the optimism :: investing mechanism or accounting for the presence of institutional computerized investing.

So, this student came up with an untestable hypothesis. That's fine. Good even. You need to generate many, many hypotheses so you can practice anticipating how you might succeed or fail at proving any given hypothesis. Eventually you will get good at creating testable hypotheses. It will take more experience with analysis still to get to the point of creating both testable and valuable (so what?) hypotheses.

Let's say you suspect that there is a direct correlation between attendance and grades. How would you test this hypothesis? Compare a bunch (how many exactly?) of attendance sheets to final grades. You would need to get a representative sample of classes, grade levels, departments, professors, and any other variables you can think of (maybe spring semester for graduating seniors provides outlier data) to make sure that it's relationship grades and attendance and not something else you are observing.

Research Question

A research question is a hypothesis expressed as a question that can be (but has not yet been) answered by systematic research. If the answer already exists, then you simply need to look it up. You don't need to rediscover the known. If the answer is unknowable, because there are too many variables or there isn't enough time or money or will or whatever, then move on. Rather than predicting an outcome and then testing the prediction, you ask the question and then set about figuring out how to answer it. Once you have sound method for testing possible answers, then you pilot the method to verify it will work, and then you gather the data in full to discover the answer. Once you have all the data, ideally, you can use prediction to verify the answer.

Will the electric fuel cell car replace the combustion engine car?
Does attendance correlate with grades?
Do frequent readers write better?
Is there a maximum income past which no greater happiness is achieved?
Do good looking people experience greater success in life?

Qualitative and Quantitative Question

  • Qualitative: experience, feeling, subjective impression, belief, attitude,
  • Quantitative: anything measurable

    What was once only available as qualitative data might one day become measurable. Before hear disease was understood and orthoscopic procedures were available, a person who expressed to his doctor a persistent experience of exhaustion would be offering qualitative data and the Dr. would have to diagnose from there. Now that we can look inside the primary arteries and measure occlusion, Drs. have quantitative data to work from.

    Typically, organizations prefer quantitative data because it seems to offer certainty and precision. The squishy feelings and beliefs and attitudes evidence is dubious and hard to draw inferences from.





    Relevant (realistic)


  • Questionnaires

    advantages -- potential for large sample, even a random sample -- aggregately answers -- asking the same questions over time enables capture of change (and so improvement) -- easy to administer using online tools and email lists

    disadvantages -- people don't like to fill out surveys -- harder to frame questions accurately than it looks, compelling numbers can be misleading.

    1. anonymous or not
    2. open ended - close ended - closed question open response
      direct, indirect
      misleading, leading
      connotations leading to a preference of answers
      positive expressions vs. negative --- allowed to X prohibited from -x
    3. sample size
    4. validity -- observes a verifiable truth
    5. reliability -- achieves the same result more often than not. margin of error
    6. accuracy vs. precision -- precision is hitting the same mark repeatedly, reliably; accuracy is hitting the mark you wanted to hit reliably. A rifle that hits the outer ring of a bulls eye 10 time in a row is precise but inaccurate. A rifle that hits the bulls eye repeatedly is precise and accurate.
    7. order of questions matters -- context changes meaning -- what was your worst dating experience, do you like dating
    8. must be unambiguous, so each respondent answers exactly the same question, no interpretation required or enabled
      be specific -- how can we improve the product is not as good a question as how can we improve the look of the product. You need to make sure the question includes the answer to "what"? in the expression.
    9. yes no must be truly yes or no -- exclusive of each other. If an item can have both descriptions then all yes answers are suspect
      rock or mineral? answer: yes
    10. people will choose an answer even if they don't know, so a don't know is a good idea

    A Worked Example of an Attitude Survey

    Let's say you have been asked to find out what students at a large urban research institution of higher education think about their learning management system. The system they use is locally branded but is more widely known as Desire 2 Learn. Where do you start? Well, pretty much every student has some familiarity with the software, so identifying the relevant population won't be hard. Do you need to distinguish among levels of experience or can you get by with just flattening that distinction?

    Could you just ask each person you see for the next 60 minutes: How do you feel about D2L?

    Stated in that way, you are asking an open-ended question. You would need to allow for as many different answers as people you ask. And then you would need to figure out how to code the answers so you can group them, so you can generalize about feelings. Also, what does "feel" mean and might it not mean something different to each user? And how would you know? You could ask each informant to define feel, but add a lot of time to the questionnaire and many of not most would have no idea how to respond helpfully.

    You could offer what's known as a Likert Scale, thus enabling quantification and reducing the amount of time a person would need to do the questionnaire.

    How consistent is the following statement with your experience of D2L?

    I love D2L!

    1) Not at all consistent -- 2) Somewhat inconsistent -- 3) Somewhat consistent -- 4) Totally consistent

    Now if you asked 10 people, you would have 10 numbers to work with. Is 10 people enough? Well, if your your population is this class, 10 would be roughly a third. What if your intended population where the whole school?

    Representative sample

    What is an N and how do you know if it's representative?

    Refining our previous effort

  • How often do you use it -- give a range so you can quantify, otherwise you get uncountable answers
  • When do you use it? (too vague -- time of day, time of semester, time as defined by due dates?),
  • What features do you use?
  • Do you like it when profs don't use it?
  • Do you like when profs use tools in addition to D2L?
  • Which features do you like?
  • Which features would you like to see?
  • Which features don't work the way you wish they did?
  • Have you ever used a different learning management system?
  • If you took a class where the prof didn't use it, did you wish they had?
    Asking intervening question can disrupt a train of thought. That can be good or bad.
    Asking contradictory questions can cross-check accuracy of belief.

    Questions to ask of all survey data

  • What was the N
  • Was it the relevant population
  • How representative of the relevant population
  • Were the questions unambiguous, unbiased
  • Will the answers change over time
  • How were respondents motivated to answer
  • How well informed were they