Assumptions, opinions, and beliefs act like knowledge until you convert them into a testable hypothesis. Daniel J. Boorstin has said, “The greatest obstacle to discovery is not ignorance – it is the illusion of knowledge.” To challenge assumptions, you need to be able to write a testable hypothesis.
Being able to form and test a hypothesis might be a skill you’d expect in a science lab, but not in a business environment or even in your everyday life. However, the world is changing.
Leaders are using data to inform decisions. To use data, you need to be able to understand what makes a testable hypothesis. Getting big data to speak to you requires the formation first of a testable hypothesis.
One leader lamented, “I know I am supposed to delegate this, but I can’t let go.” Just because many other leaders in similar businesses delegate that task, it doesn’t mean it’s right for him or his business. Forget “supposed to” and test if it’s true for yourself.
Unpacking why he can’t delegate that task, he articulates beliefs such as no one will do it as well; it will tarnish his branding, among others.
Components of a Testable Hypothesis
Once you have identified any beliefs, assumptions, and opinions, they can be tested to find the truth. Any ambiguity can resolve clarity with a testable hypothesis. The facts help you make a confident decision and move on. To get there, we need to write a testable hypothesis.
Expectation of an Outcome
In short, a hypothesis is a statement that could reside on a true or false exam.
It can take the form of an if/then statement, such as “If fines for texting while driving were increased, then fewer people would engage in that activity.” You expect a certain effect, and your experiment is all about testing to see if that effect you expect is accurate.
Based on Knowledge and Information
The point of a testable hypothesis is to build on the current knowledge base to move it forward and should reflect that edge of knowledge.
Unfortunately, publishing bias may cause you to repeat an experiment. That is results which don’t turn out as the experimenter hoped to end up in a drawer, not in a journal. Studies that do wind up in journals may not yet have been replicated. A lack of replication throws any conclusion into question.
Continuing the texting example, we know that fines are a traditional tactic to change behavior. We also know we want to drive change in that direction because of the truth of the danger, and the drivers (lack of) appreciation of the truth.
Simple and Concise
Keep it simple.
One statement, one concept, and one opposite. It should be written as a definite statement, not as a question. For example, “Snails prefer wet soil to dry soil.” It will involve only one test variable at a time. In this case, the moisture condition of the soil.
A good hypothesis can be proven false.
It’s critical to know that the absence of evidence is not evidence of absence.
That is, you can find evidence to prove something does exist, but you cannot prove that it does not exist. You only need to find one black swan to prove that all swans are not white, but you can’t prove that purple swans don’t exist.
A good testable hypothesis will state what will happen in a clear form, such that when anything else happens, the experimenter will say, “that’s funny.” More than Eureka moments, it’s the oddities we want to explain away that have the power to open new insights.
“Ambition makes people work harder” might sound like a good hypothesis. However, you are going to have to answer the questions what is ambition, and what is harder?
If you can debate a word, you need to use measurable terms. An operational definition adds measurability.
Operational definitions are used to clarify terms. A dictionary definition may not be precise enough, or you may need to introduce organizational or industry specific words.
For instance, if you decide that harder means work days that are 10 hours long, what of the work day that involves reading the paper for four hours? Stick to concrete, measurable concepts for provable hypotheses.
To challenge your limiting beliefs, you need to form testable hypotheses. With a good hypothesis in mind, a straightforward experiment is much easier to design.