Think about a couple of scenarios.

When you ask your agency for a status update, your account manager replies, “We are likely to submit the work to you today.” Do you believe you will actually receive something from them today?

When an equity research analyst states in a report, “Company X is likely to meet its earnings target,” would you consider investing in the company?

Qualitative terms such as “likely”, “unlikely”, “probably” and “good chance” can be ambiguous and misleading.

In the equity research example I mentioned earlier, when an equity research analyst states, “Company X is likely to meet its earnings target,” they may actually be suggesting that there is a 60% chance of this happening. However, investors might interpret the term “likely” as a strong indication of confidence, implying a probability of over 80%. As a result, they could become overly optimistic about the company’s performance.

Qualitative terms can be ambiguous because people may interpret them differently.

Prof. Wade Fagen-Ulmschneider, Teaching Professor of Computer Science at the University of Illinois Urbana-Champaign, published an article “Perception of Probability Words” in April 2019. He created an intriguing survey to examine how individuals interpret probabilistic words and invited his social media connections to participate.

What he found out from n=123 respondents:

• “Likely” and “probably” refer to an approximate 70% chance.
• “Probably not” means approximately 25%.
• “Almost no chance” means approximately 2%.

This pretty aligns with “Words Of Estimative Probability” written by Sherman Kent in 1964 while he was working for the CIA. According to Kent, “probably” indicates a probability of approximately 75%, while “probably not” suggests about 30%.

What do all these discussions suggest?

First, we should avoid using the qualitative terms such as “likely,” “probably,” or “unlikely” without careful consideration. Before you speak or write, always ask yourself: “Am I correct to say so?” or “Will people really get what I say?”

The PHIA Probability Yardstick published by the UK Government has provided a good guidance. If you are about 80% certain about an event, you should say “highly likely” instead of “certain”. If your certainty is only about 70%, use the term “likely”.

Second, if you can have the exact data to support your argument, speak it out rather than using qualitative terms.

The weather forecast reports that “there is a 30% probability of precipitation,” rather than “it is unlikely to rain.”

If your market research findings indicate that 62% of respondents prefer the new formulation, write “62%” in your executive summary instead of “likely prefer.”

I look forward to your thoughts.

Vincent

I am Vincent Kwong, a seasoned consumer planner with nearly two decades of market research and business planning experience on both the agency and client sides. Currently, I am Director & Partner of Intuit Research, a research agency specializing in multi-country market research projects. Companies that I have served in my career include McDonald’s, Diageo, PepsiCo, Estee Lauder, Heineken, GSK, Pizza Hut, Lenovo and HSBC. I am also a part-time instructor at Schulich School of Business, York University in Canada.

Read My Other Viewpoints

<< Back Home