READING ASSIGNMENT
Fraenkel & Wallen Text: Chapters 1 and 2
WRITTEN ASSIGNMENT
Your assignments for this unit are to select three subjects of interest to you and then pose each of them in three different ways of questioning as follows:
1. State the subject in general terms, such as the relation of class size and student achievement, together with the reasons for your interest in it.
2. State the subject in a more specific but still non-researchable way, as with our modified question from the Study Notes.
3. Restate the subject in a researchable way with its terms and behaviors appropriately defined, specified, and measurable.
Prioritize the three according to your degree of interest (Healthcare Educator Degree). The 3 subjects are Promoting Wellness, Health Concerns, and how to cope with or manage existing health conditions(One page each should suffice.)

STUDY NOTES
SCIENTIFIC THEORIES AND METHODS
A course in educational research design should begin with a brief consideration of the nature of scientific methods and theories and what sets them apart from other means of inquiring and of reaching conclusions. Basically, the aim of the scientific method is reliable replicability: similar procedures in similar circumstances, similarly controlling all variables, should always yield similar results, no matter who is performing the procedure. The aim of replicability requires that questions being researched be sharply defined rather than vague, specifically focused rather than broadly general, and that the phenomena being questioned be both physically observable and measurable?or at least be reliable representations of something not physically measurable. (The latter leads to some of the most bitter controversies in educational research, as we shall see.)
The aim of scientific theory is verifiable explanation of observed phenomena, with the grander the theory, the larger number of phenomena being explained. Theory is based on the results of the scientific method and suggests further observations or experiments, the results of which would either reinforce or undermine the stimulating theory. A good theory will provide accurate predictions of the existence of previously unrecognized facts.
The central core of all academic disciplines, however, from textual criticism, through contemporary education, to quantum physics, is not a body of known facts but a body of unanswered questions and unresolved controversies. Theories are generated in part to supply tentative answers to those questions. So, theories stimulate investigations and experiments of many different kinds to see if the theory is valid or not. This course will mainly be about the opportunities, requirements, and pitfalls of designing those investigating researches. But before we examine the requirements of ?researchability? in this unit, several aspects of the underlying
theories must be clear. They are commonly known as ?falsifiability,? ?convergence,? and ?synergism.?
?Falsifiability? is what distinguishes theories from beliefs. All theories must be testable, and they can be invalidated as well as demonstrated by facts discovered through scientific methods. All supporters of a theory, though they be as faithful to it as to their spouses, can always specify facts which, if discovered, would invalidate the theory. When asked what would invalidate evolution, for example, the English biologist J. B. S. Haldane unhesitatingly responded ?mammal fossils in the pre-Cambrian layer.? All real theories are eternally tentative, always subject to modification or abandonment.
However, most academic disciplines?most assuredly including education among them?feature a large number of competing alternative theories. This competition stimulates a large number of different kinds of research investigations, all intending to generate evidence in support of one theory or another. ?Convergence? refers to the condition when a number of different kinds of research?your text devotes a dozen chapters to as many different kinds?consistently provide more support for one theory than for its competitors. A number of different kinds of research, by different people, into different facts, and by different methods, is usually required for a theory to be strongly supported, if only because the history of scholarship shows all too well how research can be artfully designed to support a predetermined conclusion.
Some research is purely descriptive in nature?call it the ?what is that there?? type?while other research is aimed at hypothesis testing?call that the ?what would happen if?? variety. The most common form of the latter type is to measure something, then change something about it or its environment, predict the effect, then re-measure and see if the results are as the hypothesis predicted. ?Synergism? refers to the interaction of descriptive and hypothesis-testing research designs. The former produces baseline measurements of something, e.g., a group of students? reading performances, a group of patients? recovery times, etc. The latter suggests that changing something, say materials or methods, would improve something desirable. The change is implemented and the results measured. Those results potentially become new descriptive data. Further hypotheses can thereby be generated, and so forth.
All such conclusions, however, must ordinarily be seen as tentative since we can seldom be entirely certain when studying people that the change (or lack of it) which we observe has been the result of what we did, or the result of some other extraneous, unaccounted-for factor. The study of people and the attempt to assign causes to their behaviors is replete with ?extraneous variable? problems, of which there is more in Unit 2.
Scientific research can be divided into essentially two kinds: that having to do with the behavior of human beings as its subject matter, and that dealing with other animals, plants, and inanimate matter. Some human research that deals with our involuntary responses, such as much medical investigation, has similarities with the second kind, but most human research involves attempts to measure and/or manipulate our conscious, ?voluntary? capabilities. On the other side of that coin, studies of animal behavior are often used in attempts to explicate human behaviors?think of Ivan Pavlov and classical conditioning?but such arguments from analogy are always to some
degree suspect. The concerns of educational research are predominantly with peoples? behaviors which are subject to a mixture of voluntary efforts and involuntary constraints.
In studying people, it can be useful to begin with the distinction between observation and manipulation: the former is present in most educational research designs, the latter in many but not all of them. Similarly, with observations, a useful initial distinction is between those observations known and unknown to their subjects. Some profoundly important sciences are purely observational, such as astronomy and taxonomy, where observation is necessarily unobtrusive. It is also frequently the means that is preferred in descriptive research on people, but such is often impossible.
The major problem with studies based on observation that is known to its subjects is commonly called the ?Hawthorne effect,? from Mayo and Rothlichsberger?s classic factory studies in the 1920s: groups of people who know they are being observed for some unknown purpose may (or may not) change their behavior in apparently unpredictable ways. It is the human equivalent of Heisenberg?s uncertainty principle in subatomic physics where the observed is changed by the act of observation. Much ingenuity has gone into attempts to minimize this effect, but its invalidating dangers are always present.
Particularly in educational research, unobtrusive data collection is often desirable but impossible. Often the best we can do is to at least initially disguise the purpose for which we are collecting our data, as in the use of ?double blind? designs. Fortunately, such anonymity is often unnecessary since many research designs can be improved by enlisting the active cooperation of their subjects. It is incumbent on the research practitioner to make good decisions on when and how research purposes will be made known to the research subjects.
?You can see a lot by observing,? said the great American philosopher Berra, and he was certainly correct. But probably the most common reason for observing in educational research is to prepare for, or evaluate the effects of, some kind of deliberate change?some kind of intervention. Much educational research has the pragmatic purpose of improving some kind of learning, often through a teacher or other practitioner deliberately changing something. To be able to tell what, if any, improvement has taken place, pre-intervention baseline data must be gathered, if possible, without the subjects? knowledge of the eventual purpose to which it will be put. Such preliminary data-gathering is surprisingly easy to overlook, but it is absolutely essential for any valid conclusions to be drawn from the work.
One of the most common means of measuring improvement is by the pretest-posttest method: gather baseline data, implement some innovation, gather similar data afterward, and see if a hypothesized change has taken place. To assure validity, however, a control group is greatly desired for comparisons: a very similar group of people to those who experienced the change, but who themselves did not experience it, on whom similar data will be gathered. Without such a comparison, the significance of the measured change in the first group could be very difficult to judge.
RESEARCHABLE QUESTIONS
Before proceeding further with research designs, however, we need to first nail down the nature of scientifically researchable questions. Research always begins with the asking of questions, but those questions must have certain characteristics. ?Where does my lap go when I stand up?? is obviously not scientifically researchable, but neither is the question: ?Are smaller classes more conducive to student achievement than larger ones??
The requirements of researchable questions are feasibility, significance, clarity, and morality. The latter will be addressed later in the course. Of the first three, feasibility (can I do this) and significance (does this matter, especially to me) are usually relatively easy to establish. Clarity, on the other hand, is all too often another matter. Note that the second question posed above has a common meaning and embodies a commonly accepted idea that smaller classes are somehow generally better than bigger ones. But, as presented, it is still not a researchable question. It obviously lacks a definition of words, and it also wants for specific observational and measurable terms as well. It poses a relationship between class size and student performance which might well be true, but which as stated cannot be verified.
?Defined, specific, observable, and measurable?: these words should be a regular mantra of educational researchers. In our question, ?smaller? and ?larger? need more than a dictionary definition; they need a numerical one that will not mean different things to different people. Similarly, ?student achievement? must be defined operationally: achievement based on their having done what observable behavior specific enough to be accurately measured? (But even operational definitions can be quite tricky, especially when they involve abstract constructs such as ?intelligence.? They can be structured in such a way as to determine the outcome of the study. Many cost-benefit analyses demonstrate this: ?costs? and ?benefits? can be defined and given numerical values so as to determine the conclusions of a supposedly impartial study.) In our example here, grades might be a possibility, although today student performances on various state- or national-level tests are more commonly used. So our question could be restated as ?Will students in classes of 15 or fewer perform better on statewide tests than students in classes of 16 or more?? This is improved in definition, specificity, and measurability, but it is still seriously inadequate.
At this point, this is because so many variables have not been included in the question or considered in its context?the type and nature of the subject matter and learning materials; the age or grade of the students; the ethnic and socioeconomic demography of the students? families; the size, location, and per capita budget of the schools; the experience and methods of the teachers; and on and on. All these and more could have as substantial an impact on ?student achievement? as the sizes of classes. They are all potential extraneous variables, the presence of which, if uncontrolled, would invalidate the researcher?s conclusions.
Some of them, such as the subject matter and the students? grade level, can be easily accommodated in a one-sentence research question: ?Will eighth-grade students in English classes perform better on statewide reading tests in classes of 15 or fewer than in classes of 16 or more.? Others, however, such as those dealing with student-family demographics and the
characteristics of the school(s) and qualities of the teacher(s), strongly imply a requirement of separate research efforts to establish comparison data for adequate control.
Note here that the question as presented also implies that student performance on high-stakes testing is an accurate ?surrogate indicator? of students? mastery levels of a body of knowledge or a set of capabilities, which is itself a wide-open issue currently undergoing a vigorous debate. In fact, the most serious contemporary conflicts in educational research are found mainly in design issues involving surrogate indicators and intervening variables. We will consider both in Unit 2. But before we contend with those shoals and reefs, we must be sure that we have a seaworthy boat, the SS Researchable Question.