Course:KIN570/TOPICS/Research questions

THE PURPOSE OF RESEARCH QUESTIONS

Scientific knowledge is distinguished from other types of knowledge, such as religious knowledge and common sense, by being characterized as coherent, consistent, original, objective and verifiable.

However, much like what occurs in other systems in which the whole exceeds the sum of its parts, here too there is an x-factor that is responsible for having scientific research’s different elements come together and work as a team that is driven by the same common purpose. The x-factor that influences scientific research in such a way is researchers’ ability, creativity and competence in developing research questions.

By properly managing the development of research questions, scientific research is given direction, in the form of a study goal. Furthermore, researchers are also capable of assuring that study goals are meaningful.

On this issue, Salzinger argues that research can easily be made meaningful by “Investigating the behavioural mechanism behind an experiment’s surprising results ...”, which the author characterizes as “... generally more useful than trying to replicate the experiment.”

CHALLENGES IN DEVELOPING RESEARCH QUESTIONS

Unfortunately, the task of developing research questions is frequently affected negatively by some elements, such as researchers’ motivations. Nowadays, societies are mostly centred on extrinsic motivational elements known as rewards, with researchers becoming mostly focused on elements such as number of papers produced. Despite allowing researchers to progress in their careers, such criteria seems to be highly ineffective in fuelling pertinent and meaningful research questions, with research accumulating much like a pile of random and useless bricks.

GUIDELINES FOR DEVELOPING RESEARCH QUESTIONS

Historically, mankind’s inquisitive instincts have manifested themselves in different ways. From the early times of (unsupported) ruling theories, people developed into classifying possible explanations as mere hypothesis until proven otherwise. This brought about the search for arguments to support working hypotheses. Nevertheless, by having just one hypotheses being established, the emotional attachment that researchers developed towards their “pet” hypotheses made some researches a bit biased. Consequently, this elicited the adoption of a new methodology which consists of establishing multiple working hypotheses. Such procedure allows for a less biased research environment but, additionally, it also stimulates complex thinking, in the sense of looking to see the big picture and how variables relate amongst themselves.

Finally, scientific research looks to gather arguments that lend support to hypothetical explanations of the scenarios which are under scrutiny (tentative explanations). However, research results that correspond to the predicted outcome can be a result of both a correct hypotheses and an incorrect hypotheses. This means that hypotheses can never be fully proven (not candidate facts) and the most researchers can aspire to is to test hypotheses by looking to disprove them. Through the accumulation of failed attempts to disprove hypotheses, researchers are able to generate greater credibility to support tentative explanations of the studied phenomenon, which is designated as strong inference.

Developing the habit of systematizing each field’s knowledge into inductive trees of knowledge allows for the continual analysis and improvement of scientific knowledge, under a multi-factorial overview perspective.

HYPOTHESES AND IF-THEN STATEMENTS IN RESEARCH QUESTIONS

The formal if/then statements (along with modus tollens) are the only logical way to discount a hypothesis, which is central to null hypothesis testing. Even if people don't think they are using these principles of formal logic, they are (they have just not formally learned how to write it out). Being able to write out formal if/then statements and understanding how they are used to disprove the precedent ( the part in the 'if' portion) is crucial in ensuring that the experiment is testing the actual question that is being asked. All hypotheses need not be formally written out in such a manner, but researchers using the null hypothesis method should be able to formally write out their hypotheses as if/then statements and understand how this relates back to disproving the null.

If {the null hypothesis} is true, then {the intervention that was used in this experiment has no effect on the dependant variable} is true.

Written another way:

{the null hypothesis}$\to${the intervention that was used in this experiment has no effect on the dependant variable}

By Modus Tollens:

$p\to q$, therefore $\neg q\to \neg p$

{the null hypothesis}$\to${the intervention that was used in this experiment has no effect on the dependant variable}, therefore

$\neg$ {the intervention that was used in this experiment has no effect on the dependant variable} $\to \neg${the null hypothesis}

So,

If {the intervention that was used in this experiment has no effect on the dependant variable} is false, then {the null hypothesis} is false.

If probabilities weren't involved, if/then and modus tollens would be all there is to null hypothesis testing. Unfortunately probabilities are involved, so it's not so simple (the probabilities are where induction vs. deduction get involved). Some excellent examples can be found here(the 'incorrect premise' example gets at why it's crucial that an experiment's if/then statement be valid, and the probabilistic examples get at the inductive vs deductive issues).

BIBLIOGRAPHY

• Platt, J.R. (1964). Strong inference: Certain systematic methods of scientific thinking may produce more rapid progress than others. Science, 146

• Kinraide, T.B., & Denison, R.F (2003). Strong inference: The way of science. (3642), 347-353. The American Biology Teacher, 65

• Forscher, B.K. (1963). Chaos in the brickyard. , 419-424. Science Letters, 142

• Salzinger, K. (2001). Scientists should look for basic causes, not just effects. , 3590. Chronicle of Higher Education, 47( 23), B14

• Chamberlin, T.C. (1897; 2004). The method of multiple working hypotheses. Houston Geological Society Bulletin, 47, 68-69.

• Newell, A. (1973). You can’t play 20 questions with nature and win…In W.G. Chase (Ed.), Visual information processing. NY: Academic Press