This exercise will give you some experience with taking field notes and setting up a filing system. Take notes on whatever interaction you find interesting, but you should observe for at least 15 minutes to get enough material. Remember to make your notes concrete and exhaustive—everything you see, hear and smell! Document what’s happening as it occurs—processes, interaction between individuals, and anything else of interest.

 

You will be creating two sets of notes: mundane and analytic. Mundane files contain information on the different individuals involved, the setting, and the organization (or couple or group). Analytic files, on the other hand, reflect concepts or processes that you believe can explain an individual’s actions or interactions between people. This set of files usually grows as you obtain more material and new ideas arise.

 

Make multiple copies of your notes, so that you can file them in the appropriate mundane or analytic files. For example, let’s say you’re interested in a group of people who seem to get a lot less done working in a group than when working alone. You name one of your analytic files “Social loafing,” and place copies of notes describing this phenomenon in this file. If Mary, John, and Bob are part of the group, put notes referring to each of them in mundane files named after them.  

 

There are many places you can do this exercise: at a bar, at lunch, in class, in a store, and so on. Just be sure that you can observe the individuals involved for at least 15 minutes without being too intrusive. Good luck!

 

Mundane files:     

 

 

 

 

 

Analytic files:

 

 

 

 

 

 
Exercise #2: Reliability and Validity

 

 

For this exercise, grab two of your classmates and head to the student center. Select an unobtrusive spot where you can observe at least one interaction between your fellow students. Take field notes on the interaction presented and compare yours with those of your two fellow observers. Are there differences in what each of you noticed, or the ideas you have generated? Why do you think this has happened? It’s important to remember that while everyone’s observations are valid, given their point of view, background and focus of attention, they are not necessarily reliable.

 

Different people may advance different hypotheses, so we should not use this method to test hypotheses, ONLY to explore and generate ideas. Can you think of some examples of how people commonly use observation in the field to test and confirm hypotheses instead of exploring and generating ideas?

 

Just for Fun: Take a peek at the movie, “Never Been Kissed,” for some interesting and fun examples of what problems and consequences might result, when deception is used in research.

 

 

 

 

 

 

 

Researchers must make sure that the behaviors they select as indicators of the phenomenon of interest are valid; that is, they indicate what you say they are supposed to indicate.

 

Let’s examine the phenomenon of “flirting.” Flirting is often considered an indicator of sexual interest in another person. But how do we define flirting, as it’s a complex behavior? How can we find out whether it indicates sexual interest or not? Can we say with certainty that the behaviors that comprise flirting mean someone is flirting?

What behaviors are involved? List them here:

 

 

 

 

 

 

What other explanations are there for these behaviors, besides flirting? List them here:

 

 

 

 

 

 

How can we establish that these behaviors are flirting, and involve sexual interest in another person? Get the evidence! Ask people who seem to have an interest in another person, evidenced by “flirting” behavior, if this is really the case (and get ready for some surprised looks!). Build up a case that behaviors X, Y, and Z (prolonged eye contact, leaning forward, and licking lips might serve) are part of flirting and do indicate interest.

 

Now, think of another phenomenon that interests you and decide what behavioral indicators you would use. Are they valid? How will you establish this? Discuss how you might do so with at least one other classmate. Do you agree? If not, why not?

 

 

 

 

 
 
Exercise #4: Reactivity

 

 

The presence of an observer may change the behavior of those being observed. This can be due to evaluation apprehension, or concern with how other will judge them, which often increases the occurrence of socially desirable behaviors and decreases the occurrence of socially undesirable behaviors.

 

Behavior may also change when the attention-feedback-regulation cycle is set into motion; that is, the presence of an observer calls attention to certain behaviors, which the person being observed notices, and then changes. Here are some examples of reactivity:

 

1)      You are an experimenter interested in the effects of caffeine on time needed to solve puzzles. You begin to notice that whenever a subject seems to become aware of the video camera mounted on the wall above them, their performance drastically improves. You have not been videotaping their performance, only timing it, and have not made reference to the camera at any time. What can you do to correct this problem?

 

2)      You are researching honesty out in the “real world” by dropping money on the street, then observing whether people make an effort to return it to you or not. You have noticed that when people believe someone has seen them pick up the money, they are less likely to pocket it and more likely to return it. Is this a problem, can you change it, or have you inadvertently made an important discovery?

 

3)      You are interested in finding out what behaviors contribute to successful therapy conducted in a group setting. While the group you have chosen is open to anyone, they are aware that you are observing them for research purposes, and that you are not there for the same reasons they are. You find that the participants rarely express negative emotions; everyone reports that they are doing “fine,” and members seem reluctant to talk about their problems. What can you do to avoid reactivity in this situation?

 

Which example(s) reflect(s) evaluation apprehension? Attention-feedback-regulation cycle? What other examples can you come up with? What might happen? How can you avoid this problem? How can you be sure this is even a problem? Discuss as a class.

 

 

 
 
Exercise #5: Behavior Code

 

 

Researchers use behavior codes to guide observations of behaviors of particular interest to them. For example, therapists-in-training often watch videotapes of sessions with clients to evaluate their own performance. Interventions can be coded to see if therapists are overusing some skills, neglecting others, or could have used a different intervention. Here’s a sample behavioral code (Hill & O’Brien, 1999):  

 

        Category                                                             Description

 

1) Approval and reassurance              Provides emotional support, reassurance, encouragement

                                                            and reinforcement.

2) Closed question                              Requests a one- or two-word answer (“yes,” “no,” or a

                                                            confirmation) and is used to gather information.

3) Open question                                 Requests more than a one- or two-word answer; asks

                                                            clients to clarify or explore thoughts or feelings. Can be

                                                            phrased as directives or queries.

4) Restatement                                    Repeating or rephrasing of the content or meaning of

                                                            client statements that typically contains fewer but

                                                            similar words, and usually is more concrete and clear than

                                                            the client’s statement.

5) Reflection of feelings                       Repeating or rephrasing of client’s statements, including

                                                            an explicit identification of client’s feelings.

6) Challenge                                        Points out discrepancies or contradictions, defenses, or

                                                            irrational beliefs of which the client is unaware or is

                                                            unwilling or unable to change.

7) Interpretation                                  A statement that goes beyond what the client has overtly

                                                            stated or recognized and gives a new meaning, reason or

                                                            explanation for behaviors, thoughts or feelings so that

                                                            the client can see problems in a new way.

8) Self-disclosure                                Reveals personal non-immediate information about the

                                                            helper’s history or credentials or about feelings.

9) Immediacy                                      Disclosing immediate feelings about self in relation to

                                                            the client, the client, or the therapeutic relationship.

 

Now select a behavior of interest to you, whether it’s how you instructor teaches or interactions among children at a local playground. Develop a behavioral code that you think is exhaustive, relatively simple and reflects your initial hypothesis. As you observe, you may expand or change your code to refine the process.

 

Just for Fun: For an amusing example of someone using a behavioral code to observe events, watch “Addicted to Love,” with Meg Ryan and Matthew Broderick. Do you think his coding scheme was useful, simple, or exhaustive? Was he right to use his data to predict events? What do you think?

 

 

         

 

 

 

This can be done in class or outside in the “real world.” At least two students will need to pair up and observe behavior using the behavioral code they developed in the previous exercise. You are not to compare coding with another while doing it! This will prevent observer drift from happening. After coding for a sufficient period (10 minutes is a good start), compare how you both coded. You will probably find that you are not in 100% agreement. How can you improve this, to be in complete agreement? Remember, this method of establishing reliability is not without problems—it does not correct for chance, observer drift may occur, and the more complex the categorization system, the lower the percentage of agreement. It is important to correct for this where possible, and accept a lesser percentage of agreement when necessary. How might you address some of these problems? Discuss as a class.

 

 

 

 

 

 

 

As you will recall from the text, archival research uses records to answer questions about behavior. Archival research is relatively non-reactive, as the data was collected for purposes other than yours.

 

Your first step will be to define your research question. You must operationalize your research question so that the measure from the archives reflects the concept of interest and as little else as possible. Think of a research question you think might use archival data to help answer it:

 

 

 

 

Now, what archives will you use? It is important to remember that explaining behavior is much more difficult than describing it with archival records. Additional information may be necessary to rule out plausible alternative explanations (PAEs), and make a case for your explanation.

 

What is your hypothesis?

 

 

 

 

 

How will you rule out PAEs? What are some PAEs you might encounter, given your hypothesis? What problems might you encounter in this process? Discuss as a class.

 

 

 

 

 

 

 

Below is a categorization scheme that presents a number of problems. What are they, and what can you do to correct them? Remember the five principles of developing categories:

 

1)      they must be exhaustive

2)      they should reflect the research purpose

3)      they should be mutually exclusive

4)      they should be derived from a single classification principle (different levels of analysis should not be mixed)

5)      assignment of one item to a category should not affect assignment of other items

 

Let’s say you’re interested in examining the nature of violent acts perpetrated on women vs. men in TV shows. Here’s your categorization scheme so far:

 

                                                            Women              Men

     Rape                                              ______             ______

     Murder                                           ______             ______

     Robbery                                         ______             ______

     Assault/Battery                               ______             ______

     Domestic violence                          ______             ______

 

 

As you watch TV with your big bag of chips, you start to notice you’re having trouble with your scheme. What problems do you see at this point? How can you fix them? Discuss as a class.

 

Now, think of your own research question and devise your own categorization scheme. Keep the five principles in mind as you work. Pair up with a classmate and discuss what problems you see in each other’s schemes (if any), and how they could be improved.

 

My research question:

 

 

 

 

My categorization scheme:

 

 

 

 

 

 

 

As you’ve learned from our text, qualitative and standardized interviews are quite different, in purpose, structure and flexibility. It is important to decide which subtype will suit your research purposes best. Below are some situations in which you might want to use one or the other—decide which you think is better suited to the task at hand, and why. Discuss as a class.

 

1) You are interested in differences between men and women in their attitudes toward capital punishment, if any.

 

2) You are examining whether counseling psychologists-in-training are willing to receive therapy while in school, and under what circumstances they would seek therapy.

 

3) You want to find out whether 5-year-olds have developed gender-based ideas of which jobs are “appropriate” for them.

 

4) You are trying to find out if gender, age, or socio-economic status has an effect on attitudes toward sexual minority members.

 

 

 

 

 

 

 

Standardized interviews mostly make use of closed questions—all response options are provided and the respondent chooses one of these options. It is important to word questions carefully, so that they mean the same thing to all respondents AND you! Below are some questions that may or may not be poorly worded or poorly constructed. Try to select items that contain poor questions or poor response sets, and discuss how you might correct them.

 

1)      Do you believe in abortion?  ____Yes  ____No  ____Not sure

 

2)      Do you think that the need of the South American people justifies the decimation of the rainforest?  ____Yes  ____No              ____Not sure

 

3)      How often do you watch prime-time TV?  ____Never  ____Less than once a week

         ____Once a week  ____2 to 3 times a week  ____4 to 6 times a week  ____Every day

 

4)      How do you feel about programs designed to integrate the disabled into the work force?

         ____Against  ____Neutral  ____For

 

5)      How many TV programs do you watch each week?  ____None  ____1 to 2  ____3 to 5

         ____More than 5

 

6)      Are you in favor of bigger tax cuts for smaller, rural towns?  ____Yes  ____No

         ____Undecided

 

7)      How many times a week do you exercise?  ____I don’t exercise  ____Once  ____Twice

         ____3 to 4 times  ____5 to 6 times  ____Every day

 

8)      Do you agree with the recent movement toward greater tolerance and legal rights for

         individuals who identify as gay, lesbian, or bisexual?  ____Yes  ____No  ____Not sure

 

9)      How frequently do you engage in sexual intercourse?  ____Seldom  ____Occasionally

         ____Frequently

 

10)    On a scale of 1 to 10 (with 10 meaning completely successful), how successful has the

         welfare limit been in getting recipients to find gainful employment?  ____

 

11)    On a scale from 1 to 10 (with 10 meaning very positive), how would you rate your

         attitudes toward people of other races?  ____

 

 

 

 

 

 

 

Using what you have learned about what makes a good question/response set, it’s your turn to make up an interview script. Think of a topic you might like to quiz your fellow students on and generate a short list of questions (10 or more). You’ll also need to create a written set of instructions to introduce the interview, the topic and so forth.

 

Next, pretest your script by trying it out on people. See if any problems come up in the process—questions, requests for clarification, questions to which everyone responds in the same way, etc. If you have any open-ended questions, think about probes you can use to get more information from respondents, if answers are unclear or not detailed enough. For example, you might repeat the question or say, “Tell me more.” If you run into problems, how will you adjust your script? What kinds of problems do you foresee? Discuss potential problems and solutions as a class.

 

 

 

 

 

 

 

Select a topic or question you would like to explore using a survey format. Construct a short survey (about 10 questions) and administer it to a few fellow students to see what you may need to change. What problems can you envision encountering with your survey? How will you solve these problems? Your classmates can serve as a good resource during the pretesting phase by deliberately trying to misunderstand and misinterpret questions, giving answers that avoid the real purpose of the question, and generally trying to be as clueless as possible; this will force you to be simple, straightforward and clear in your survey construction. Discuss the results of your efforts as a class.

 

 

 

 

 

 

 

Let’s use this class as a sample from the population of interest to us, namely, the psychology department. We are interested in revamping the department using the information we generate from our survey. Please answer the following questions:

 

1) Do you find the research courses to have been helpful in preparing you for graduate school?  ____Yes                                     ____No  ____Neutral

 

2) Do you feel additional courses in more specialized topic areas should be made available?  ____Yes  ____No  ____Not sure

 

3) On a scale from 1 to 10 (with 10 meaning very high quality), how would you rate:

 

    ____  The quality of instruction?

    ____  The quality of materials used?

    ____  The overall quality of your education in psychology?

    ____  The quality of equipment?

    ____  The quality/availability of research experience?

 

4) Please rank order the following in order of importance, with 1 being most important:

 

    ____Quality of instruction

    ____Research experience availability

    ____Accessibility of faculty

    ____Non-competitive atmosphere

    ____Quality of educational materials selected

    ____Variety of special-topic courses

    ____Prestige of department

    ____Quality of research courses

    ____Other:_________________________________________________________

                      

 

Now, let’s discuss our results. Do you think students in other classes will respond in the same way? Why or why not? How might this change if, for example, our sample was a graduate class of counseling students? Experimental psychology students? How does the composition of the class affect our conclusions?

 

What we are looking at here is sampling error—this class may or may not be representative of the population as a whole, and may therefore give us different results. Because we know our sampling frame (the population we wish to examine), we can use a number of ways to attempt to make our sample as representative as possible, and avoid sampling error as much as possible.

 

 

 

 

 

 

 

To do this, we need a short set of questions—it can be fun stuff, so don’t sweat it! Ask about love lives, foods people like, whatever! The point of this exercise is to demonstrate how the method we use to sample can affect results.

 

1) Systematic random sample—we choose every fourth name using the class roster. What are the results?

 

2) Simple random sample—we’ll just draw some names out of a hat, or just point to a few people randomly! Are the results          different now?

 

3) Multistage cluster sampling—we’ll divide the class into quadrants, and pick a few from each quadrant. How are things            different now?

 

4) Stratified—we’ll divide into men and women, or by class ranking, and then select equal percentages of each. What’s                different this time?

 

It is important to use the sampling method that fits your research question. If gender is a theoretically important variable, use stratified sampling. If you have a large population, like a big city, that you want to sample, multistage cluster sampling may be best. The point is to tailor your method to your research question.

 

 

 

 

 

 

 

Causal inferences are statements that an intervention really produced the differences we observe. It’s important to rule out plausible alternative explanations (PAEs) that might also explain the difference, because it increases support for the causal inferences we are making.

 

Let’s say we’re curious about whether being addicted to smoking is more psychological or physiological, and which hampers kicking the habit more. Our research participants come in every week so we can assess how they’re coping with having quit smoking. Half of our participants receive a nicotine patch and the other half a patch containing no active ingredients (a placebo), although none of them knows who got which. The staff are warm and encouraging, and the lab has attained a nearly 90% success rate. The researchers suspect that addiction to smoking is largely psychological because only half the patches are nicotine, but the success rate is much higher. Unfortunately, they have forgotten to check potential PAEs. Can you figure out what they might be?

 

Possible answers: social reinforcement and support, evaluation apprehension, or even demand characteristics. Any of these could explain the high success rate, which may have nothing to do with the patch at all! This is why it’s so important to pay attention to possible PAEs, especially when we’re conducting research out in the real world. Can you think of any real-life examples? This focus on ruling out PAEs is called internal validity—whether the intervention we introduced or the change we observed has really caused the difference.