Testable hypotheses (by Dan Reisberg, ch. 1)
Research in cognitive psychology can yield results that are both important and
useful. These results have value, however, only if they are based on sound
methods and good science. If not, we may be offering practical suggestions that
do more harm than good, and theoretical claims that lead us away from the truth,
not toward it.
For students learning about cognitive psychology, therefore, it is important to
understand the methods that make our science possible. This will allow them to
see why our results must be taken seriously, and why we can, with confidence,
draw the conclusions that we do. For this reason, I have written a brief methods
essay, highlighting a methodological issue or focusing on a research example,
for each of the chapters in the text. I hope these essays will help convey why
our methods are as they are, and help students understand more fully how our
science proceeds.
First, though, we might ask: What is science, and what is it about
cognitive psychology that makes it count as a science? One key lies in the
testability of our claims: In order for psychology to be a science, our
claims must rest on a basis that's more solid than mere opinion, or
interpretation, or argument. We need our claims, instead, to rest on the facts,
and that means we need a way to check each assertion against the facts, to find
out with certainty whether the assertion is correct or not. Then, if we learn
that an assertion does not fit with the facts, we are obligated to set it
aside, to make sure that we offer only claims that are supported by good
evidence.
How does this testing of hypotheses proceed? The answer has many parts, but one
idea is crucial: Scientific hypotheses need to be framed from the start in terms
that make the testing possible. That means we must avoid vague or ambiguous
terms, so that it is obvious what facts fit with the claims and what facts do
not. Otherwise, we'll have no way of making certain that our claims are well
founded.
Even when we have a relatively specific hypothesis, we need to define our terms
with care. For example, consider the claim: "No matter what day of the year you
pick, someone famous was born on that day." Is this hypothesis correct? Imagine
that the most prominent person you can think of, born on December 19, is Daniel
Reisberg. Does this confirm the claim, because Reisberg is famous? (After all,
many thousands of students have read his books.) Or does it contradict the
claim, because Reisberg isn't famous? (After all, most people have never heard
of him.) Both of these positions seem sensible, and so it will be a matter of
opinion whether our observation about December 19 fits with the claim about
birthdays or not. Hence the truth of the claim about birthdays depends on
opinion, not fact, and so the claim is not testable. To make it testable, we'd
need to find a suitable definition of "famous." Only then could we say if our
observation about Reisberg fits with the hypothesis or not.
This example illustrates, therefore, why a scientific hypothesis must be framed
precisely-so that we can check the facts, and then say with certainty whether
the hypothesis is correct. But how do we "check the facts"? We'll explore this
in upcoming methods essays.