Handout: Inductive Reasoning

MAN 5245 / MDarnell

Inductive Reasoning

Inductive reasoning is the reasoning we engage in to establish empirical facts. A necessary condition for
inductive reasoning is empirical data – we need to experience something. Because we rely on our bodily
experiences to establish empirical facts, it is important to remember that facts are not necessarily true.
“The sun will rise tomorrow morning from the east” is a fact, and we accept it as true, but is it
necessarily true that the sun will rise tomorrow from the east? Absolutely not – there is a possibility that
some event will alter the relationship between the earth and sun, though this possibility is so slim, we
accept with something like 99.9% probability that, indeed, the sun will rise tomorrow…. Notice, the we
often accept that the evidence we have to support a particular belief as “good enough” to simply accept
it and move on – this closely aligns with the concept of ‘satisficing’ that is discussed in relationship to
bounded rationality decision making models.

Of course, we distinguish between some claims as having evidence that is ‘good enough’ for me to
accept and live my life as though it is true, and other claims that I accept but with reservation or perhaps
will personally accept but do not expect others to accept, and still other claims that I reject because of
insufficient evidence or even evidence to the contrary. This is the distinction, for example, between a
fact and an opinion. Opinions are beliefs we hold despite a certain weakness in the evidence we have for
holding them. Note that the evidence we have for believing a claim (this is what we call an
epistemological issue – an issue relating to our knowledge of something) is independent of whether the
claim is actually true or false (that is a metaphysical question – a question of reality). Thus, just because
a belief is an opinion, doesn’t mean it isn’t true, or helpful. Still, to be good decision makers, we want to
make decisions that come with a high degree of probability of being the ‘correct’ decision. We want to
deal with, and establish, facts.

A fact is a belief which is supported by a strong, cogent argument. Strength is a matter of how well the
evidence you are relying on (the premises) can establish the conclusion, if those premises are true (note
that strength is for inductive arguments what validity is for deductive arguments). Strength ultimately
boils down to 4 characteristics of evidence: Relevance, Reliability, Objectivity, and Quantity. If you have
a lot (quantity) of objective evidence, that is relevant to the conclusion and either establishes a reliable
correlation or is from a reliable source (or both), then there is good reason to believe the conclusion. As
any one of these criterion begins to lack, the strength of the argument decreases. If an argument is
strong, then it can be evaluated for cogency – a strong argument with premises that are actually true is
cogent (all weak arguments are automatically uncogent). To determine if a premise is (empirically) true,
you will once again need to engage in inductive reasoning, and evaluate this new argument for
soundness and cogency (and yes, this process of reasoning might go on indefinitely, but at some point
we simply must accept certain basic things as true, and get on with our lives!).

There are named patterns for commonly structured inductive arguments, which in turn enable us to
focus on specific elements of strength. Four of the most common are:

Statistical Syllogism: this is when we reason from empirical knowledge about a group or class of things
to a conclusion about a single member of the group (“75% of resumes contain misleading information. I
Handout: Inductive Reasoning
MAN 5245 / MDarnell

am currently reviewing a resume for a position I want filled. This resume probably contains misleading
information.”) The % indicated in the first premise is a key indicator of strength (but it can never be
100%, or the argument ceases to be inductive).

Enumerative Induction / Inductive Generalization: this is when we reason from individual members of a
group to conclusions about the group as a whole (from particular to general) (“75% of surveyed clients
express satisfaction with our services. Therefore, 75% of all of our clients are satisfied with our
services.”) Strength for this pattern of argument is largely determined by the sample size (how many
people were surveyed?) and how representative the sample is of the total population (did we only
survey people from a specific segment of our clientele? Is there a bias inherent in the sample?). This
form of argument also raises issues about how information is collected in surveys ( e.g. were posed
questions biased?) and also statistical measurements of data that is collected (what is the margin of
error? Confidence level? …).

Analogical Induction/ Argument by Analogy: in this form of argument, we reason from an awareness
that two or more things are alike in some respect, that they must be similar in some further respect
(“My last intern was a student at UF, majored in general business administration, had a GPA higher than
3.5, and performed exceedingly well in all assigned tasks. This new person I’m considering bringing on as
an intern is also a student at UF, is majoring in business administration, has a 3.6 GPA. So, this new
person will also likely perform well in all assigned tasks.”). Strength for this type of argument is
dependent on: relevance of known similarities between the things being compared to the property we
are concluding about (is college/major/GPA relevant to completing tasks in a work environment?) ,
relevance of known dissimilarities between the things being compared to the property we are
concluding about (if my last intern was male, and the new person I’m considering is female, is that
relevant to being able to complete tasks in a work environment?), the number of instances being
compared (if I’ve had 20 interns from UF, majoring in business admin., with high GPAs, then I have even
more reason to believe that this next person will be able to complete tasks…), and diversity among cases
(has the work environment or type of tasks assigned to the last intern different from the
environment/tasks that will be assigned the new intern? Has there been a significant change to the
quality of education at UF between the time of my last intern and now? …).

Inference to the best explanation: this is when we experience a certain phenomenon, recognize a
theory that is the best explanation for that phenomenon, and then accept the truth of the theory (“Sales
for product X are low. The best explanation for the low sales is poor marketing. So, it is probably true
that our marketing department is not doing a good job.”). We generally accept that any argument that
follows this pattern is strong (simply because of what we mean by a best explanation), but here we can
see the importance of cogency. If the explanation we cite isn’t actually the best explanation, the entire
argument falls apart. So how do you know if some explanation is the best explanation? Well that too
requires inductive reasoning. Explanations are generally evaluated on the following criteria, which are
presented in 2 distinct groups:

1. Logical consistency (it isn’t self-contradictory)and external consistency (it fully accounts for the
phenomenon – not just part of it). These two criteria are necessary conditions for an
Handout: Inductive Reasoning
MAN 5245 / MDarnell

explanation – if an explanation is inconsistent in either of the two ways mentioned, it

We each engage in the above forms of reasoning on a regular basis. The important thing is to be able to
recognize when you are relying on the above (or other) patterns of reasoning, so that you can be more
aware of potential errors in your reasoning, and be able to build confidence in the conclusions you
reach/decisions you make.

If you are interested in exercising your decision making skills, work through the following problems.
Answers are provided on the last page of this handout.

Exercises:

1. Evaluate the following argument: “Most Americans are happy with their jobs and derive a great
deal of satisfaction from them. A survey of 1500 adults with an annual income of $48,000 –
$60,000, employed in various occupations, supports this assertion. When these subjects were
asked if they were happy and satisfied with their jobs, 82% said yes.” (taken from Vaughn, The
Power of Critical Thinking)
2. Evaluate the following argument: “Character is the foundation stone upon which one must build
to win respect. Just as no worthy building can be erected on a weak foundation, so no lasting
reputation worthy of respect can be built on a weak character.” (R.C. Samuel)
3. This morning, one of your employees was late to a very important meeting. When you speak to
your employee about this, she says “I’ve worked here for 6 months, and most mornings I leave
for work at 7:00am, and I’ve always been on time. This morning, I left at 7:00am, so you really
can’t blame me for being late.” Is your employee right? Is it reasonable to blame her for being
late?
4. Bloomberg Businessweek is always printing stories about successful business people engaging in
very physical, often high risk activities (e.g. extreme mountain climbing) as hobbies. So, if you
want to be successful in business, you too should start engaging in physical, high risk activities.
5. One of your employees, who for the past several years has been an ideal employee, has recently
been slowing in productivity, is often late coming to work or leaves early, seems to be more
reclusive, looks as though his physical health is suddenly deteriorating, and you recently
overheard him asking one of his co-workers to lend him some money. You believe the best
explanation for this change in behavior is drug abuse, and are thinking about reaching out to
your company’s legal counsel about requiring a drug test, but also don’t want to damage this
young man’s reputation with false accusations. Is the explanation you’ve arrived at (drug abuse)
the best explanation?
Handout: Inductive Reasoning
MAN 5245 / MDarnell

Answers/comments on sample exercises:

1. This is an enumerative induction. While there are some ‘good’ aspects to this argument,
ultimately it is weak, because the sample size is low, we have little information about who
actually constituted the sample (Americans?), and the sample is likely biased, since the income
range of the sample is very specific. This evidence alone doesn’t give us good reason for
believing the conclusion (“most Americans are happy with their jobs…”).
2. This is an argument by analogy, and is a somewhat strong argument. Character refers to an
unchanging demeanor that allows us to predict behavior in the future and what the individual
will be able to successfully endure, just as the foundation of a building allows us to predict the
likelihood that a building will be able to endure certain stresses.
3. The argument made by the employee follows the pattern of a statistical syllogism, and her
argument is actually somewhat strong. If for 6 months, leaving at 7:00am results in being on
time for work, it is reasonable to believe that she would be on time this morning, assuming that
today was very much like every other day in the past 6 months (e.g. no new major road
construction projects). Nonetheless, in this scenario, the manager needs to recognize that there
might be some other issues that the manager her or himself might be partly responsible for –
like making sure employees know they are expected to be at the office early on those days that
a meeting takes place. To this point, the employee wasn’t recognizing that her own reasoning
pattern was inductive, and that being on time in the past doesn’t guarantee being on time
today; if she had good critical thinking skills, and took the meeting seriously, she might have
taken the initiative to get to work early!
4. This argument can be re-constructed as an argument by analogy (you are like the people in the
stories in the sense that you desire to be successful in business, the people in the stories engage
in physical activities for their hobbies, so you should engage in similar activities), or an
enumerative induction (a % of successful business people who are written about in
BusinessWeek engage in physical hobbies, so all successful business people engage in physical
hobbies). Either way, the argument is weak – simply engaging in certain hobbies by itself is not
particularly relevant to whether you will be successful in business, and the % of people written
about in Business Week is very small. Still, there might be an interesting relationship here to
explore – perhaps tolerance for risk taking and high need for achievement explains both the
physical hobbies and the success in business.
5. The explanation for the changes in behavior is reasonably strong: it is testable, has a wide scope
(can explain all changes in behavior), is simple, and is conservative (it fits with established signs
of drug abuse).