Business Research Methodologies
ws20180928
Research_Reasoning.pptDeductive, inductive and abductive approaches
Deductive reasoning
You can do deductive reasoning while sitting in your armchair.
Deductive reasoning involves inferring that if propositions A and B are both true, then this implies that C is also true. The classic example is over two thousand years old and involves the question of whether the ancient Greek philosopher Socrates was mortal.
Statement A: All men are mortal.
Statement B: Socrates is a man.
Inference: Since Socrates is a man, and all men are mortal,
Socrates must therefore be mortal.
Deductive reasoning can be used very effectively in combination with a powerful new theory (e.g. deductions that followed from Newton's theory of gravity). It can also be used to make inferences from the research literature, where different parts of the literature contain the initial statements for the deduction.
Although deductive reasoning can be very powerful, its inferences need to be checked against reality, because it can easily produce results which look completely sound but which are based on subtle errors.
For instance, the Socrates example depends on statement A and statement B using exactly the same definitions of 'mortal' and 'man' as each other.
It also assumes that no other factors are involved. However, other factors are often involved in cases of deductive reasoning. Consider the following example.
All eagles can fly
Tweety is an eagle
Therefore Tweety can fly
What if Tweety has a broken wing? Tweety would then be unable to fly.
We could handle this by changing the initial statement to say 'All healthy eagles can fly'. However, there are many other reasons why an eagle might not be able to fly.
For example, Tweety might have had some wing feathers removed by the vet. If we list all the possible restrictions, then the statement can become very long or so restricted that it is of little use.
Another problem involves the definition of 'eagle'. This is a classification and classification is often complex, debatable or fuzzy.
For instance, the usual classification of human adults into 'male' and 'female' assumes that everyone either has XX or XY chromosomes but this isn't the case. There are people with XXY or XYY chromosomes, as well as numerous other conditions which make this classification fuzzy.
These issues limit the usefulness of deductive reasoning in research.
You may also be wondering where the initial statements come from, and the answer is usually that they come from observation of the world, which involves a further set of challenges for the researcher.
These challenges are clear in the case of inductive reasoning, described next.
Inductive reasoning
This involves working from observations towards an inference.
A common example of inductive reasoning is:
Observation: All the crows that I have ever seen were black
Inference: All crows everywhere (including ones I have never seen) are black.
A lot of research involves induction from observation of unexpected regularities.
For example, you might collect a large sample of jokes to study and make the observation that most jokes about stupidity involve a social group who speak the same language as the person telling the joke but who live at the geographical border of the joke-teller's region. You might then make the inductive inference that the same regularity might occur in other parts of the world that you haven't yet studied.
Inductive reasoning can provide extremely valuable insights but has obvious limitations.
It's usually based on a set of observations which is not complete, so you can't be sure whether you haven't seen a white crow because there are no white crows in the world, or whether there are some white crows in the world but you simply happen not to have seen any of them yet.
Abductive reasoning
Abductive reasoning involves deciding what the most likely inference is that can be made
from a set of observations.
A classic example is the wet grass example.
Observation: The grass outside my window was wet when I woke up this morning.
Known fact: Rain in the night can make grass wet.
Abductive inference: There was probably rain in the night.
Abductive reasoning is important because there is often many or an infinite number of
possible explanations for a phenomenon, so you need some way to decide which
possible explanations to look at first.
In the case of the wet grass, for example, it's just about possible that the grass outside your window was wet because a lot of people walking by just happened to decide to empty their water bottles onto your lawn, but this is not very likely.
Usually when you're researching you start with the most likely explanation and see whether that's true. If it isn't, you move on to the next most likely explanation (for example, that your lawn is wet because your neighbour was using a hose and accidentally sprayed your lawn).
There's a substantial literature about ways of assessing probabilities and assessing the strength of evidence.
