EDF 5481 READINGS AND ASSIGNMENTS

ASSIGNMENT FOUR SPECS ASSIGNMENT FOUR EXAMPLES ASSIGNMENT FIVE SPECS OVERVIEW NOW POSTED: EXAM 3 STUDY GUIDE

IF I HAVE REQUESTED A COPY OF THIS, OR A PREVIOUS, ASSIGNMENT ON DISKETTE, I WOULD LIKE TO RECEIVE THEM BY OUR EXAM, DECEMBER 12, 2001.  THANKS VERY MUCH!

 

It has been very enjoyable to watch your study designs evolve. Everyone is much more sophisticated about study design than they were three months ago.

At this point, nearly everyone is able to write a comprehensive, yet succinct, research statement. When you embark on your research, this paragraph should be placed somewhere on the first page of your paper, to alert your readers to the topic, its importance, and what you plan to do.

I hope that you can now fully appreciate my statement at the beginning of the semester: that your research problem statement will be the most difficult task of the semester! Even while wading through the mountains of terminology, and the strengths and weaknesses of different designs, it is surprisingly difficult to clearly and briefly state what your research is about, why the topic matters, and what, specifically, you will do.

Congratulations! You will now find that writing the rest of your research paper is  much, much easier, because you truly know what your research topic is and what you plan for it. This will facilitate your catching redundancies, to only address issues that directly relate to your research problem, and to systematically plan your data collection and subsequent analysis.

Here are some points to note:

Match your study design to what you want to know. You generally cannot infer the internal states of individuals (e.g., beliefs and attitudes) through field observations because you typically observe external behaviors. Ethnographies typically examine a group culture so they will not be much help in assessing attitudes or beliefs either. Both ethnographies and field observations are useful to study interaction among people, however.

If you want more information about internal states, you may gain more insight through in-depth interviews or an oral history.

An oral history is different from a structured survey! It is much longer and uses virtually all open-ended questions. It is very unlikely that you would do oral histories with 60 students in an online course or 40 team members on a sports team. Each oral history typically takes hours at a minimum, and some take weeks, even months. Remember, you are basically doing a type of biography and that's not a form someone can fill out in 10 minutes.

Consider what is the best sample for YOUR study purpose. Even if you take a probability sample, you will not be able to generalize from "a sample of one" (e.g., a single school, hospital, or company). A purposive sample may be your best choice in this instance, or perhaps some type of quota sample. Avoid a grab sample, however, if this is at all possible.

Examine the answer to Jorge's proposed ethnography, question 2 BELOW.

There are many different kinds of secondary analysis. Secondary analysis is a generic term that refers to analyzing data that were originally collected for a different purpose. Thus, my current statistical analysis of how gender and education influence cognitive complexity in Internet use with the General Social Survey data is secondary analysis, because that was not the original purpose of this GSS module. Your analysis of interaction patterns in television situation comedies is also secondary analysis because these comedies were originally made to entertain (and attract advertising dollars for revenue.)

Note the differences however. I am not really content-analyzing the General Social Survey data, which remains a structured survey. I am building some new indices and then analyzing those but the original categories remain the same. Anyone who works with this subset of te GSS will have to take the precoded responses as given. There is very little depth to the answers because this survey almost entirely utilized closed questions.

Contrast my use of the GSS with your study of interaction patterns on television. You are content analyzing. You are taking originally unstructured data, creating your own coding scheme, and creating an entirely new and unqiue set of data. Someone else might do a content analysis of the same comedies and create a totally different schema, thereby constructing a second original dataset.

We both did secondary analysis. But most practitioners would argue that you did a content analysis and I did not.
 

HAD TO DO ALL OF THE FOLLOWING:

• WAS COMPLETE. This means that every section of the assignment was completed.

• Wrote a clear AND CONCISE (no long, long, long paragraphs, especially on paragraph one) description of their research topic, why this topic is important to study and what they planned to do.

• Accurately designated their choice of research design and described why it was appropriate. For example, participant or field observation is not generally very good at diagnosing internal states such as beliefs or attitudes. A survey design would be more appropriate here. Oral histories are lengthier and go into much more detail than a short, structured survey.

• Gave enough accurate detail so that it was clear that what the student wanted to do in fact matched the design designation they chose. This typically meant giving an example or two of what would be coded or observed, providing a small coding grid, or describing the type of field notes that would be taken.

• Described their sample or census thoroughly. Described why a probability or non-probability sample would be better to take in this instance.

• Accurately diagnosed Jorge's research design problems. Used less structured design terminology accurately. Recognized why a non-probability "sample of one" may be adequate for design purposes. Understood which of Jorge's measures were more or less subjective. At least partly addressed the issue of "what is scientific" in part 6.

• You STILL have problems with different types of samples.
• You confused a grab sample with a purposive (judgment) sample.
• The operational variable you provided for Jorge was difficult or even impossible to operationalize.
• You weren't sure WHAT kind of design you had!
• You had so little detail that it wasn't clear what kind of design you would use. For example, your design appeared to describe a structured survey which you designated as an oral history.
• You thought a less structured design wasn't scientific because it didn't produce numeric variables to analyze.

As a professional, you will spend considerable time reading research materials. Nearly all of them will draw conclusions and many of them will recommend interventions and/or policy continuance or changes. Therefore it is essential that you are able to accurately assess the methodological soundness of the studies that you read and the kinds of conclusions that you can draw.

This becomes particularly important with less structured study designs. Because hypotheses often are generated well into the data collection phase and because many important insights come from naturalistic field observations, internal validity is typically lower for less structured designs than for designs such as experiments, quasi-experiments and surveys, which more often employ a priori hypotheses to test.

Look for similar problem-solver questions on Exam Three.

Here is Jorge's planned ethnography:
 
 

1. Jorge's prospectus states: I am interested in how aspects of teacher-student interaction may relate to "school climate" at a local high school.

• What is the technical term for this type of statement?

Because the naturalistic field is so rich, one must have some focus to begin. Another term also used in this instance is foreshadowed problem.

2.  Is Jorge's sample a probability sample? Please answer YES or NO, then explain in a sentence or two:

• WHY or WHY NOT?

• Was the type of sample Jorge employed generally a good sample or a bad sample for the purposes of his study? Explain your reasoning briefly.

When you only have ONE school, generalizing is not going to be an issue, whether you take a probability or a non-probability sample. Obviously, you can't establish a valid standard error on a sample of one! This "sample of one" may be the best that Jorge (or any other researcher) will be able to do.

Jorge is attempting to select a school with a mix of students, a "so-called typical" school. In fact, were he to take a probability sample, he might not have such a mix. For example, he might end up with a nearly all-White or nearly all-Black school, or a school from a very poor or very wealthy district.

However, if Jorge is planning this study as the first in studying several schools, he may want to take a probability sample because with a larger sample of schools, a probability sample will allow better generalization to similar districts.

If you may have an interest in doing research at a Leon County school at some point in the future, check out the Tallahassee Democrat's synopsis of each school and the Leon County district as a whole at this Web site:
 
 

3. What do you see as the major difference between the "keeps to self" category in the coding grid and the other three behaviors that are shown in Jorge's proposed example grid?

• Which of these behaviors do you think would be the easiest to code? Why is that?

• What kind of concept do you think the behavioral entries in the grid are trying to measure? Give one additional SHORT CLEAR-CUT OPERATIONAL EXAMPLE that you think would reflect this concept.

In contrast, a wave or a smile is objective and could be coded by more than one observer. Are these "molecular variables" more valuable? Hard to tell. The concept Jorge appears to want to measure was called, for example:

• "friendliness"
• social interaction
• isolation versus inclusion

• frowns (reversal)
• hugs
• pats on shoulder ("skinship")
• eats lunch with a group
• estimate physical distance between individuals as an indicator of closeness
• mutual laughter

4. Here are some techniques that Jorge COULD use in his study:

Content analysis
Gleaning
Oral histories
Participant observation

Jorge's MAJOR technique is participant observation because he plans to spend considerable time "shadowing" students and faculty, sharing meals, and even going to gym class!

While he will also do content analysis, gather some oral histories and gleaning (graffiti in bathrooms, perhaps?), it is clear that his participation in student and faculty activities will take the major amount of time. Most ethnographies will combine all four techniques, but may vary the emphasis.

Here are several useful suggestions:

• A student assistant will be less obtrusive in observations of student activities and can probably participate in activities that Jorge practically cannot.
• This will be especially true if he has at least one FEMALE assistant (girl's gym, lunch table, lavatory).
• An assistant will provide someone for Jorge to interact with who can provide feedback to help keep Jorge objective.
• An assistant can help with the reliability of observations. No one person can observe it all, let alone reliably.
• An assistant can analyze Jorge's field notes for major themes.

• What is your BRIEF rejoinder to Jorge's colleague's comment?

EDF 5481 READINGS AND ASSIGNMENTS

OVERVIEW

Happy Holidays!

Susan Carol Losh. December  4, 2001.
This page was built with Netscape Composer
and is best viewed with Netscape Navigator
600 X 800 display resolution.