A colleague of mine wrote a sentence of the following sort. I'd like to know if more likely to have died is used properly. The use of likely here does not seem to indicate uncertainty. How many of the respondents died should be verifiable by the researchers as a black-and-white matter. Also. from the way the sentence is written, it does not seem to be a general statement referring to drug takers in general. So why is "likely" used here at all?

Following up twenty years later, the team learned that among the original 5,500 interviewees, those who had said they took drugs in their teens were 25 percent more likely to have died. Many of them died of cancer.

  • The sentence is fine. Likely refers to probability not certainty. You are more likely to find that a person in that group had died, compared to some other implicit control group, and that fact is being expressed with them as the subject. People subjected to ads are ___% more likely to purchase an item. – Tᴚoɯɐuo Nov 5 '18 at 13:22
  • But there is not anything probable. How many of the respondents died is a clear matter. And the sentence is not an assertion about drug takers' life expectancy in general. – Apollyon Nov 5 '18 at 13:26
  • Probability can be assessed after-the-fact. There's no requirement that it be a prognostication. I'm not sure how you're coming up with the requirement that likely has to "indicate (prospective) uncertainty". What it means is that individual members of that group had an increased probability of dying within the implicit time period (compared to the member of a control group). – Tᴚoɯɐuo Nov 5 '18 at 13:32
  • more likely to [past infinitive] =the probability was greater that. – Lambie Nov 5 '18 at 14:14
  • @Apollyon It is a comparison. How likely was group A to have died? Suppose 30 out of 100 in group A died. How likely was group B to have died? Suppose 55 out of 100 in group B died. So, group B was 25 percent more likely to have died than group B. – Brandin Nov 5 '18 at 16:30

The use of likely does indicate uncertainty. In this context, there was a interview 20 years ago that presumably asked about drug use in teens. We can group the people who underwent the interview into two groups. Let's say that group A reported that they used drugs as teens and that group B reported that they did not use drugs as teens.

Twenty years later, the researchers determine which of the interviewees have died. They can see that the death rate of group A is 25% higher than group B. The group percentages are a known quantity, and the fact of whether any specific individual died is a known quantity. However, the point of the sentence is that if you select a random individual from group A and another random individual from group B, it is 25% more likely that the person from A has died than that the person from B has died. On an individual level, the researchers could just check to see whether the people died, but the point of statistics is to illustrate the overall probability of certain events.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.