Criteria for good research

advertisement
1
Criteria for Good Research with Emphasis on Qualitative Research
This document is based on an exploration of the question, “What makes research
good?” by students in a similar research methods course than you are taking. I have
added some criteria that were missing from your notes and filled in details. You can use
the document for:


Evaluating an example of research
Designing your own research (i.e., research proposal)
1) Problem statement
This is extremely important. The first question about any piece of research is: What is
the problem? The problem logically comes before the research questions. A good
paper/proposal has a problem statement that does the following:



Explain what the problem is
Explain why the problem is important to research. This includes the novelty of
the problem (has it been investigated before? Is there previous research, but
there still are important aspects that need research?)
Explain how the study contributes to knowledge about the problem
All this should be in very simple language. Try explaining your problem statement to
someone in the canteen, or a family member who is not an expert in education. If you
can’t get them interested in your research, the problem statement is not good enough.
The problem statement should help to develop a focus, so that the research does not try
to cover a general topic.
2) Research Questions
The research question is also extremely important and must be researchable. A
question is researchable when it is possible to design a study that can answer the
research question; some questions cannot be answered empirically, or do not require
empirical research. You also mentioned that studies have hypotheses, but this is less
common for qualitative research. Here are some other important issues with research
questions:



Have a small number of research questions, or else the paper/proposal
becomes too unfocused and it becomes difficult to see what the new contribution
of the research is
The research questions should be related to the research problem, and
suggest specific ways to study the problem
The research methods must be capable of answering the research questions.
Can you really come to a conclusion about the conceptual change of a student
based on a 200-word transcript of an interview?
2

The research questions must be answered during the study, at least to some
extent.
3) Objectives
Any one study will not be able to fully solve a difficult problem, so it important that
researchers spell out what they hope to accomplish with the study via objectives. For,
example, “The objective of this study is to gain a better understanding of how successful
teachers encourage student to student talk, with a view to designing professional
development.” Objectives can clarify:




What it will be possible to do with the study results?
The kind of knowledge expected (e.g., practical knowledge in action research
versus more generalizable knowledge)
The boundaries of the study. What must be studied to achieve the objectives
and what is better left to a future study? This is especially important for case
studies.
Feasibility: Are the resources and time available sufficient for achieving the
objectives? A project can be too ambitious for the resources available, but it also
is important to be aware that a study can be so small-scale that nothing
worthwhile can be learned from the study.
4) Literature review
The literature review is important for framing the study and demonstrating how the
study adds to existing knowledge. The literature review develops the argument for
doing the study—why it would be useful. Note that just the fact that research has not
been done before is not necessarily a reason for doing it now. The literature review
should:



Discuss what is known about the problem already. If research already exists it
helps to demonstrate that the field is interested in the field
Discuss what is not known about the problem. There may be gaps in the
literature—aspects that have not been studied at all or have not been studied
well—that can help you to identify a “niche,” some research questions that would
be worth studying. For example, “Almost all studies in this literature examine
teacher learning, but no study was found that also examines the impact on student
learning. Some authors (references) argue that professional development should
have impact on classroom learning.”
Introduce the major concepts that will be used to analyze the data—the
conceptual framework. For example, if you are interviewing teachers about
classroom management, what kinds of theories and concepts will be used to
analyze the teacher’s responses? Here it is impossible to define terms well.
There are different ways to think about a problem, so it is important that researchers
anticipate alternate ideas than the ones that underlie a conceptual framework (i.e.
3
criticisms). It is particularly to note that researchers may be biased or make unstated
assumptions that can be challenged.
Literature reviews vary greatly. Some journals require short articles, so the literature
review can be very focused; other journals require a broader exploration of the
problem. Without an adequate literature review it is very difficult to evaluate if a
study makes a contribution to knowledge.
It is important that literature reviews are based on good literature (e.g., not
conference papers when journal articles or books on the subject are available), and are
reasonably comprehensive (e.g., also works from before 2000, so that major areas of
work are not missed, or also social psychology of learning in groups and not just
cognitive psychology). And a good literature does not just summarize studies (“Chan
said xxx, Postman said yy”) but provides analysis (your commentary), so that the case for
doing the new piece of research gradually becomes clear. There is a tendency for
readers to lose their way in the details of a literature review, missing the main points.
5) Description of data collection methods
The methods section of a paper/proposal is very important. Readers need to be able to
understand what the researchers did, and need to be able to evaluate whether the
methods used are of sufficient quality to make answers to the research questions
possible.
The first job of the methods section is to justify the overall approach chosen. Why
action research and not a case study? Why not an experiment? Researchers need to
demonstrate that they are aware that the research questions can be studied in
different ways, and that there are debates in the field about how certain kinds of
research questions are best investigated.
There also are many issues with sampling and recruitment:





Is the “case” chosen the best one to study the problem or research questions?
E.g., if you wan to study the difficulties of implementing assessment for learning,
why would Queens College make a good school to study?
Are the phenomena sampled enough? Is solving problems in one area of the
curriculum (e.g., trigonometry) enough for meeting the objectives of the study?
Do you have enough participants? What kinds of participants are necessary
(e.g., teacher and his/her students, but not friends or the principal)? Sample size
also is important for inferential statistics (t-test, ANOVA, etc., in which you are
trying to generalize to a population.
Besides sample size, representativeness of the sample is also important,
especially in quantitative research. For example, does the proportion of
participants in your sample who are women match that in the population or
society?
How are the participants recruited? Are there any criteria for excluding
participants from the study? How were the participants informed about what
would be involved in participating (ethics)
4
Description of procedures:





What was the nature of the intervention, if there was one? E.g., in action
research, what was the action undertaken, why, and how?
How long did the intervention last?
How well was the intervention carried out? For example, if it was important
to have mixed-ability groups, were these used, or did students choose their
friends to be in their group?
What special materials were used, and how? If students did online discussions,
did they do them in a computer lab or after school? What instructions were
students given? Were these discussions assessed?
Ethics: Did the procedures follow or fail to follow ethical principles (e.g., a long
experiment that gives a treatment to one class but not another class, the use of
deception, payment of participants).
The basic idea here is to provide as full a description of what happened, what the
students and teacher did, as possible.
Description of data collection:








Was a pilot study completed to clarify the research questions or various data
collection methods? E.g., did the researchers try out their interviewing
procedures before doing the main study?
Do the interview questions lead the interviewee in a certain direction or are
they more neutral? Is the interview carried out in a way that produces the best
information possible? Were interviews carried out in English or Cantonese?
How often did interviews take place? What kinds? How long did they last? Were
they recorded?
How often did observations take place? What was their focus? How were they
recorded?
What kinds of materials were collected? Why?
What tests were used? What do we know about their validity? What process
was used to develop the test (if a new one)?
Was more than one data source collected? How does each data source contribute
to answering the research questions?
Is the data collection plan logically adequate for answering the research
questions? For example, if a study lacks a pre-test can you still evaluate learning?
Is a comparison group needed? The researcher needs to make decisions
regarding these issues before the data collection. Was the study long enough to
answer the research questions?
Description of data analysis:

Are procedures for processing the data described? E.g., levels of transcription,
checking of transcription by participant.
5














Are questionnaires and materials in that were developed in English translated
into Chinese and back translated (to check accuracy of the translation)?
Are the procedures by which a coding scheme is developed explained in
detail (e.g., different kinds of coding). What was the source of ideas for coding?
Are any judgments of a subjective kind (e.g., ratings, coding) repeated by an
independent rater? Or would it make sense not to expect close agreement?
Are the theories or perspectives that underlie the analysis described (e.g.,
grounded theory, conversation analysis, hermeneutics)?
Are inferences made from some of the data tested on the remaining data?
This can mean different things. E.g., testing inferences from interviews in the
observation data (triangulation), or testing inferences from 6 cases on the
remaining 2 cases (Grounded theory should be tested)
Are questionnaire data cleaned? E.g., removal of incomplete questionnaires or
obviously bogus responses.
Is the coding audited, so that biases could be detected?
Is both data that confirm and disconfirm reported, if both exist?
Is an effect that happened just once over-emphasized?
Is it explained how the quotes shown in a paper were chosen? Why do the
researchers focus on this quote and not many other possible ones that would
lead to a different conclusion?
Have the researchers done anything special to investigate questions along the
way? E.g., if 4-5 students were absent from a pre-test and were excluded from
the study, did the researchers investigate if the sample was still representative?
Is the data analysis superficial or in-depth? Are the analytical methods
appropriate for the data—i.e., are the assumptions about the data that
underlie the analysis valid?
Ethics: are the results “too good to be true”. Are results likely to be fabricated
or altered?
Are learning assessments sufficiently distant from the intervention?
6) Reporting of findings, discussion, and conclusion
You won’t have to deal with this in the proposal, but it is important in any research
report. Some issues:
The Results section



Are the research questions answered? Do the researchers explain what the
data are revealing about the research questions? Or is the presentation datadriven, in which the researchers summarize the results in an unfocused way,
without returning to the research questions?
Are assertions based on the data analysis, or do they seem like opinions on the
topic that are not related to the evidence presented?
Was a participant check carried out to see how the participants reacted to the
interpretations and descriptions in the paper.
6




Are alternate explanations considered? For example, rather than the
intervention maturation of the participants could explain the results
Do the researchers clearly delineate the boundaries around their claims?
“The results show that xx but not that yyy.”
Does the Results section stay focused on the results, or does it wonder away and
go into long discussions.
Does the study include clear comparisons that facilitate interpretation? Does
the study have clear findings that are likely to have impact? Or are the findings
so commonplace that they do not seem to require research?
The Discussion/conclusions section:







This should not report additional findings. That’s what the results section is for!
Do the researchers review whether their objectives were met?
Do the researchers discuss the implications of the study by returning to the
problem statement literature?
Do the researchers point out the limitations of this study to qualify their
conclusions?
Do they suggest next steps? Are these steps based on the findings of the study?
E.g., if the researchers recommend more attention to professional development,
is that a recommendation based on the study, or is it their favorite
recommendation, which they make after every study of theirs?
Does the paper have a conclusion? Are the conclusions based on analysis of
results/data?
Is the focus of the study preserved in the discussion or does the framing wonder
away from earlier sections?
Download