d i g i t a l   a d v i s o r . . .

Using Theories, Hypotheses and Frameworks

You may wonder what our obsession with theory is all about. This is best explained in a song by George Schultz, which goes like this:

A fact without a theory

Is like a ship without a sail,

Is like a boat without a rudder,

Is like a kite without a tail.

A fact without a figure

Is a tragic final act,

But one thing worse

In this universe

Is a theory without a fact

Theories are the medium in which facts reside. They give meaning to facts, help us see why the facts are important. But, as the last line of the poem also reminds us, theories can become fantasies, dreams of how he world is that lack any basis in reality.

The relationship between theory and fact, then, can be difficult. If you have a theory about, say,how how classroom life affects teachers, then it is easy for you to "see" evidence for your theory all around you. You can interpret everything you see as supporting your theory. This is why we do research, to test the merits of theories, force ourselves to examine evidence more closely.

This is also why we often focus on articulating the details of a theory. All of us hole numerous informal theories that we use to interpret everyday events. These are called beliefs, prejudices, values, assumptions, etc. If you don't have a formal theory guiding your research, you will likely base your findings on these un-articulated prior ideas.

One of the most difficult aspects of research to come to grips with is your own prior ideas.  It is not possible to enter the study with a blank-slate mind. You have some expectations for what you will see, some suspicions about who is influencing whom, some assumptions about how things work. These ideas can't help but influence your study design and they can't help but influence what you actually do see. They influence your design in the sense that you select particular kinds of schools, teachers, or classrooms to study because you expect to see certain things there, and you phrase your interview questions or survey questions in certain ways because you expect these questions to yield certain kinds of responses.

Often, the researchers' prior ideas and assumptions are harmful because they encourage you to "see" some kinds of findings and blind you to other kinds of findings. Attached here is a chart showing all the ways your prior ideas can influence your research design, data collection, and data analysis.

This is one of the reasons why we tend to obsess about theory. One way to protect yourself from these problems is to use a conceptual framework, hypothesis or theory.  All of these things are forms of "starting ideas," just as are your own unspoken beliefs. But they differ from your unspoken beliefs in that they are explicit. All the assumptions and predictions are laid out, so that other people can see them and can know where you are coming from.

These starting ideas can be used in many different ways, some more powerful than others. Here are a few of the ways people use them.

1. to interpret to events

Starting ideas are often used to interpret observed events or data you have collected. For instance, suppose your starting idea is that teachers' beliefs about the nature of subject matter influence how they teach that subject. You could do a study of a particular teacher that involves watching her teach this subject and interviewing her about the subject, and then trying to sort out the places where her beliefs appear to have influenced her practice.

Or suppose you start with the idea that learning communities are a good format for promoting teacher learning. You could study a learning community and you examine all of the interactions among teachers, looking for evidence of teacher learning.

Notice that, when you use starting ideas in this way, you are not learning anything really new. You already believe that the teachers' beliefs influence her practice, or that learning communities are a good thing. Your study does not change your mind on this issue.  Instead, the study serves to reinforce this starting idea.

2. to raise new questions

Starting ideas can also help you formulate new questions. For instance, suppose you take constructivism as your starting idea. This is a theory of learning that has also led to a set of pedagogical proposals, but most of these proposals are still speculative. You could use this state of the art to raise new questions, perhaps about the relationship between the constructivist pedagogies and the constructivist learning theories. Or you could ask about the different ways people interpret this term.

3. to generate competing hypotheses

Starting ideas can also help you develop a set of competing hypotheses that your study can address.  Larry Cuban did this in his study, How Teachers Taught.  He started with several hypotheses for why teaching practices have remained relatively stable over time and then uses the historical record to search for evidence supporting or refuting each hypothesis.  My caretaker, Mary Kennedy, also used competing hypotheses in her study of what teachers learn from research. You could use this strategy to evaluate competing hypotheses about classroom practices, competing hypotheses about school district hiring practices, or competing hypotheses about about nearly anything else.

Policy research is ready-made for competing hypotheses studies, in the sense that every policy is, in a sense, a hypothesis.  For instance, when a state introduces a high-stakes assessment, it is hypothesizing that this assessment will increase student achievement. Nearly every policy has both advocates and critics, however, so a research project could be designed to search for evidence favoring or refuting each position.

4. to test the ideas themselves

The three uses of starting ideas described above all assume the idea itself is true, or at least plausible, but what if it is wrong?  You can also develop a study that challenges your starting idea by posing some alternative premises. For instance, hypotheses about constructivist teaching often are based on the assumption that children must be actively engaged in learning new material. But what if it is possible for them to learn passively, with nothing more than a textbook or a movie to help them learn? The technology that is presently available would allow you to create different kinds of learning activities that vary in how much they require the student to actively interact with the stimuli, so you could actually test this premise rather than assuming it to be true.

5. to develop your research methods

Starting ideas are also helpful in develop research designs, interview guides and so forth. For example, if you were testing a policy hypothesis about high stakes testing, you would likely want to include teachers who taught in the grade level where the test existed and teachers who did not, because you might suspect that the influence of the test would differ depending on how far, in grade levels, the teacher was from the high-stakes test.

Some researchers also use Argyris and Schön's distinction between espoused theories ( the ideas about human nature that you claim to hold) and theories in use (the ideas you appear to believe, given your behavior) to develop research designs. One set of authors, for instance, suggests that researchers should not trust interviews with teachers about their practices because these interviews provide only their espoused theories, and it is important to actually observe them to see what their practices look like to an outsider.

Similarly, if you were interested in children's misconceptions in science, you would likely design your study to include children at particular grade levels where particular misconceptions have been known to appear, and perhaps, for comparison, some other grade levels that have not been investigated.

© Mary Kennedy, 2006


Evaluating Research

Reasoning with Evidence

Doing Your Own Research