• "The act of searching closely or carefully for or after a specified thing or person"
• "An investigation directed to discovery of some fact by careful study of a subject"
• "A course of scientific enquiry" (where scientific = "producing demonstrative knowledge")

What particular cognitive and affective skills should we have aquired already?

Bloom's taxonomy of educational objectives can be taken as an approximation for the expansion and development of knowledge (Figure 1) and thus also of the objectives of research, whose purpose (other than fun and entertainment) is to expand and develop knowledge:

This taxonomy gives us some strong indications of the sorts of activities associated with research. Research is most often associated with the last four of the Cognitive objectives (taking the first two for granted as necessary pre-conditions for effective research), and with the last two of the Affective objectives (again taking the first three for granted or excluded as activities belonging to management or organisation rather than research).  However, it is seldom wise to take at least the first two of Bloom's Cognitive objectives for granted in any research effort, and part of this course (module) will be concerned with meeting these objectives.  Furthermore, the affective objectives are also important in much social research.

In short, the process of research consists of deciding why we want to research what we want to research, and how we think we are going to do it (the planning phase).  Subsequently, we decide how to actually do the research and then do it (what data we will collect and how, and then what we will do with these data when we have got them).  Finally, we present the results of our efforts.  Now, as a reuslt of our research, we think we know something which was not known before, and so are in a position to ask and answer the question: so what?

Notice: data are plural - the singular is datum.  Note, too, that data are generic, and include all sorts of information, from more or less highly specified and defined numbers (quantitative data), which are supposed to actually measure some quantity or some continuous or discrete quality or characteristic of the things we are interested in, to qualitative data, such as peoples' ideas and perceptions, and such as the rules and institutions which people use in carrying on their everyday lives.

Note, too, that Data collection (assembly) and analysis - interpretation involves verification (confirm the accuracy and internal consistency) and validation (confirm the coherence and applicability) of the evidence or data we are using.

However, the process of research is typically not linear, as should be obvious from the above - surely the possible answers to "so what" are important in deciding "what" and "why".  The process is circular.  The conventional (academic, scientific) cycle is typically represented as follows (following Frankfort-Nachmias, C. and Nachmias, D., (1992), Research Methods in the Social Sciences, Edward Arnold, p 22)

It is the paradigm which detemines the answer to the key question within a research process - how can we tell whether or not what we think we see is consistent or not with what we think we understand about the way the world works.

Our theory tells us a story about how data is generated in the systems we are looking at - how bits of it are related to and interact with other bits.  Our methods tell us how to look at and interpret these data and observations systematically - how to test whether they make sense (or not) with our story or theory.  The current literature and measurement are the places we go and look to find what the current understandings and stories and evidence (the paradigm) are telling us.  This literature and associated measurements (evaluations of observations) reveal the gaps and inconsistencies in the present paradigm. The objective of new research is then to fill these gaps and resolve these inconsistencies.  Easy.  Well, perhaps not easy, otherwise we would know more than we do.  But easy enough that we have made remarkable progress in understanding how our physical, chemical and biological worlds fit together and work.

Not quite so obvious that we have made the same sort of progress with our social sciences, though.  Perhaps we just haven't been doing it for long enough yet.  The process of social science research is seriously hampered by the fact that there is no commonly accepted paradigm which covers human behaviour.  Rather there are a number of partial, incomplete, and substantially separate social science paradigms:  psychological, sociological, economic, political, anthropological; all of which offer some insight into the human condition, but none of which relate closely with each other, and many of which (since there are even separate paradigms within each of the social science disciplines) are inconsistent with each other.

For example, are income distributions between people the natural consequence of economic flows and interactions, or are they the consequence of political engineering, power and control?  Obviously both - but how do the two paradigms interact?  We must suppose, I suppose, that there exists a new paradigm which might include both the economic and the political interactions - but we do not have this new paradigm yet (although there are attempts being made to develop a more general 'model' of political economy and public choice).  But even then, how does sociology or psychology fit with political economy, or even with each other?  We don't know, yet.  If we ever can hope to.

But we have to make decisions and live in this human world. To make more sensible decisions, we need to research our understandings of the ways in which our human worlds work.  Whatever we believe about the fundamental natures and processes of human behaviour patterns, we do each have some notion of how the bits of the world we live in and are interested in work.  We have certain expectations of how other people will behave in particular circumstances and contexts.  Not that these expectations are always met - we are more or less continually surprised.  But we the modify our expectations in the light of experience.  We each build up a conception, a 'model', a pattern of the way people behave and the way the world works.  So, social science might then consist of understanding how each of these patterns or world views is established and how they interact with each other - the systematics of the ways in which the world works.

So, what does the cycle of social science research look like, then?  The answer is that it tends to be customer driven to a much greater extent than biophysical science.  If you like, applied social scientists are equivalent to the engineers and doctors of the biophysical world - trying to develop better techniques and practices for living and for resolving practical difficulties, and hence responding to the demands of life - the customer.   Of course, if we are rich enough or lucky, we may also be able to simply follow our own curiosity, and not care who wants to know and why.  But we are unlikely to able to earn a living doing this - we will have to do it as a hobby. Otherwise, to do what we want to do, we will have to persuade someone to pay us while we do it, and thus persuade them that they really need to know what it is we think we can tell them.

There are three major differences in this picture of the research cycle in social science, compared with the previous illustration of the scientific cycle.

• Who wants to (or should be made to) know and why tends to substitute for the theory and methodology of the 'pure' sciences.  As a consequence, persuasion of these people and their organisations tends to become more important than scientific proofs and incontrovertible evidence accumulation - the development of a scientific paradigm.  How can we tell whether or not what we think we have found makes any real sense and fits with the ways in which the world works?  The answer, at root, involves telling a  story (as a substitute for formal theory) which convinces or persuades those who want to know (or should want to know) that we have found something useful, interesting, and novel, which can be defended against critics.

Notice that much academic social science research is aimed primarily at the peer group responsible for judging the quality and importance of the output.  These peers are the established and recognised practitioners of the disciplines (sociology, economics, etc.) in which we practice - they are the customers for academic (or 'pure') social science.  So social science has a tendency to fragment and disintegrate round rather separate disciplinary paradigms, each developing their own rules and best practices about how to do research.

The relevance of our social research to the lay public, or to the paymasters of our research, or our target audience (business managers, policy advisors etc.) then also depends very much on the extent to which they buy (are willing to accept) the presumptions and approaches of our core disciplines - our economics, our sociology, our psychology etc.  Much of our social science literature is then devoted to making the underlying (and rather different) stories we tell of the way the world works more convincing and salable.  Our social sciences are in competition with each other - which might not seem a very sensible way to try and understand the world.
 

• Our difficulty in doing social science research is seriously (perhaps even critically) hampered by the lack of a coherent, consistent and more or less complete paradigm underpining what we are trying to do.  We do not have a simple set of rules and procedures we can follow to answer the major research question of How can we tell? The methods we use are critically dependent on the particular paradigm we choose to use and believe in - economics, political science, sociology etc.  The answer to how can we tell typically resolves itself into who do we have to convince?

 
• The question of so what - what will we know, when we have done our research, that do not already know - becomes much more important for social scientists than for natural science.  Indeed, it usually helps very considerably to ask the right question - the question for which the answer is likely to prove useful, helpful, interesting and novel; the question whose answer provides a new insight into the way our human worlds seem to work; the question whose answer will prove sufficient to convince the users and judges of our research.

Key questions worthy of preliminary answers right at the outset are: at the end of this project, who will want to know the answers and why, and what use will they be able to make of them?  In other words - what is the answer to the question So What? when you have completed your research project. If you cannot think of a satisfactory answer to this question, then the chances are that you are not asking the right research questions - those which are capable of being answered beyond reasonable doubt..

The criteria against which research is conventionally judged are now well-established (the scientific method):

• ‘objectivity’, as defined by the conclusions rationally, reasonably and rigorously reached by the scientific method (or verified by peer review);
• ‘universality’ or generality of the knowledge and understanding (which are not the same things);
• ‘parsimony’, where the most useful explanation is the one which produces both the most and most reliable predictions (has largest scope) from the fewest argumentative or unprovable assertions or axioms (from the ‘simplest’ theory);
• ‘falsifiability’,  through which the explanation can be genuinely tested and shown to be something other than simple tautology;
• ‘reproducibility’, so that the results and arguments can be replicated and thus tested for veracity and validity.

• Hypothesising a specific cause-effect relationship (as an implication derived or deduced from a theory as a general law, or a consequence of the connections between demonstrated truths, or simply a guess or conjecture about how the world works);
• designing an experiment or making observations of the world (collecting data) which will allow us to test (prove) this hypothetical relationship against the facts;
• analysing the results of the experiment or the data to confirm or reject the initial hypothesis.

Because the results of our science will seldom be clear cut, we need to be able to replicate our experiments or our data collection and analysis to be sure that what we thought we had observed as either conformation (rather than confirmation) or refutation (rejection) of our hypothesis were not simply accidents but were genuine reflections of the way the world is.  Not that every research report will contain replication of evidence or experiment - but every report should contain sufficient information to allow other researchers to duplicate or replicate the research.  The greater the number of independent studies finding the same things, the more credence we can give to the results.

Furthermore, we might still be mistaken in comparing the hypothesis with the evidence, since it is extremely difficult to be sure that our hypothesis is exactly commensurate with the evidence - the hypothesis might refer to different sorts of data, in different contexts and circumstances, than we are able to collect or observe, even as a result of a more or less controlled experiment.  There are essentially two sorts of mistakes or errors we can make in confronting (testing) hypotheses with evidence:

1. accepting the evidence as refuting the hypothesis, when the hypothesis is a 'true' statement of the state of nature or the world (mistakenly rejecting a true hypothesis - a Type I error).
2. accepting the evidence as conforming to the hypothesis, when in fact it does not  (mistakenly accepting the hypothesis as an (always provisional) true reflection of the state or character of the world - a Type II error)

Clearly this objective will be frustrated if we fill our pool of scientific knowledge with Type II errors - mistakenly accepted false hypotheses as if they were (provisionally) true.  Hence, the scientific approach is inherently cautious - it uses and designs tests to minimise the chances of Type II errors (which typically means that the chances of Type I errors are increased).  Scientists are trained to be cautious, and to hedge their findings and conclusions with caveats - 'please do not rely on this experiment, evidence, as the basis of what you should do next.'  In short, rigorous science always runs the danger of being irrelevant to the problems of the practical world.

The lower arm of this figure - the research approaches which do not suppose or seek to find evidence for cause-effect relationships - might be classed as pre-scientific or a-scientific according to Popper's falsification criterion.  Yet there is little doubt that such investigations contribute to the development of science and of its associated paradigms.  For instance, much of the astronomical evidence and observation which were used to test the Copernican theory were collected without any conscious thought of testing a theory.  These data were collected as correlational or exploratory/interpretive.  In short, it is these pre-scientific investigations which are one major source of new conjectures and hypotheses about the way the world works.

Scientists are trained and bred to continually demand that more tests and experiments be done to 'confirm' that the theory - the present understanding - really does match up and stand up to experience of the world, while always recognising that the present theory may be exactly wrong, rather than roughly right.  The classic example of being exactly wrong is the theory that the sun goes round the earth.

How did we discover that it was wrong?  The answer is that science does not and cannot rely solely on empirical (observational or experimental) evidence.  It has also to rely on reason and logic - the genesis of the theory or general law.  Copernicus reasoned that an alternative hypothesis - that the earth goes round the sun - would be just as consistent with our everyday observations - the sun rising in the east and setting in the west.  Furthermore, development of this reason and logic also better explained observations of the movements of the other planets.  Thus the "theoretical - conceptual - analytical" box in the diagram above is a critical part of the scientific process, as, too, is careful use of inferrential reason and inductive logic (the lower arm of the figure).

So, pursuit of scientific methods results in a body of knowledge ( a paradigm - a coherent theory and a supporting collection of consistent empirical evidence), which is continually updated, revised and modified in the light of both theoretical developments and additional evidence, conformations and refutations, but which is always and inevitably provisional.

Perhaps the best way of characterising "the scientific method" is to think of it as making the best possible case for a particlular understanding of what is going on, or what has happened in the past and, thus what is likely to happen again in the future, or what has happened in one place and therefore can be expected to happen in other places.  This best possible case will consist of a story (theory) about why certain things happen, and evidence (data) that the story makes sense with what we see and hear.  The jury we seek to convince consists of our fellow researchers and informed participants in the actual events and behaviours we seek to examine. The judge, ultimately, will be our future experiences.

• there is a conventional best practice to the pursuit of social science research - the scientific method as applied to social sciences.  The key modification is to keep the target audience of your research (and thus the questions which most trouble and interest this audience) firmly in mind when planning and organising your research.
• But there are also very considerable problems with this conventional best practice:  the several different paradigms of social science are fragmented and largely inconsistent with one another ( in contrast to much of the natural and biophysical sciences); as a consequence, persuasion is not equivalent to proof in the same way as it might be in natural science.

Back to DRH Lecture note index.

Comments and questions?