In discussions of probability we often find ourselves saying that we can make an event more likely or less likely. It feels wordy, like there should be a single word for that. I don't mean "preclude" or "necessitate." I mean a change of probability that doesn't reach 0% or 100%. For example, we might say:

If we increase the sample size, we [decrease the probability of] certain kinds of error. If we decrease the sample size, we [increase the probability of] certain kinds of error.

What goes in that bracketed space? Not "rarify," which means something else; and not "preclude," which we can never seem to accomplish.

  • 1
    So you want a word for "the alteration of probability"? Or do you want one that specifies whether it is a decrease or it is an increase in probability?
    – Hank
    Mar 2, 2017 at 20:04
  • 2
    Based on what you say in your question, this is simply "altering the probability", and many synonyms to "alter" would work as well. Mar 2, 2017 at 20:08
  • 2
    Perhaps "inhibit". Mar 2, 2017 at 20:09
  • 2
    If we increase the sample size, we reduce statistical error through smoothening. Mar 2, 2017 at 20:10
  • 1
    @Canis Lupus Can you help me understand why your remark is not silly? It's unclear what I'm asking? Are there not boatloads of requests on this site exactly like this one in all important respects? Your first suggestion "Altering the probability of" is not one word; it's the same number of words I used in the original bracketed expression. And "If we increase the sample size we ALTER certain kinds of errors" doesn't even mean the same thing. Do you sometimes worry about your quickness to vote to close as it's unclear that you even read twice the question you're condemning?
    – Chaim
    Mar 3, 2017 at 12:50

9 Answers 9


The OED actually contains words for this:

disprobabilize, v.

trans. To deprive of probability, render improbable.

probabilize, v.

trans. To make probable or likely.

probabilify, v.

trans. To make probable, give probability to.

("disprobabilize, v.", "probabilize, v.", & "probabilify, v." OED Online. Oxford University Press, December 2016.)

Only the first is marked as rare(!), and the latter two (along with alternate spelling probablize) have all been updated with attestations into the 21st century.

So, for your example, you theoretically could say:

If we increase the sample size, we disprobabilize certain kinds of error. If we decrease the sample size, we probabilify/probabilize certain kinds of error.

To my ear, however, they all sound goofy enough to probablize you sounding like you made them up on the spot and disprobabilize a serious reception. The fact that such unwieldy words needed to be coined also probabilifies the absence of better alternatives.

  • 2
    I must say that this is my favorite response so far, for (and despite) all the reasons you give. But I'm afraid these words would probabilify a lot of giggling and confusion in class.
    – Chaim
    Mar 3, 2017 at 18:41
  • @Chaim I completely agree. I hadn't heard of any of these until I went looking for an answer to your question, but now I am thinking I might occasionally sprinkle these in conversation and in my own classes, just for the pleasure of their "mouth feel"...and being able to follow up with "It's in the OED! Look it up!"
    – 1006a
    Mar 3, 2017 at 18:59

A possibility is mitigate

Make (something bad) less severe, serious, or painful.

This doesn't apply in the general case of reducing probability, but when you're talking specifically about an adverse condition, reducing its probability also tends to reduce its severity, and I think I've seen the word used in contexts like your example.


If we increase the sample size, we disfavor certain kinds of error. If we decrease the sample size, we favor certain kinds of error.

For example, in the case of the Ising model, if we consider two neighboring variables, X, Y, the local function will favor (higher probabilities) configurations in which X = Y and will disfavor (lower probabilities) configurations in which X ≠ Y.
Luis Sucar; Probabilistic Graphical Models (2020)

With this information, they could see which sorts of triadic configurations seemed to be favored (present above chance) or disfavored (present below chance).
John Levi Martin; Thinking Through Statistics (2018)

The universe does not carry a memory of past results that will favor or disfavor future outcomes.
S. Martinez-Conde et al.; Sleights of Mind (2011)

This fact disfavors the possibility that genetic personality differences are maintained by a mutation selection hypothesis.
David Buss and Patricia Hawley; The Evolution of Personality and Individual Differences (2011)

They see the pattern as a "historical marker," something that disfavors chance resemblance, is persistent in language families, not implied by typological universals, and so on.
Lyle Campbell; American Indian Languages (1997)

Suppose intelligence is asked to estimate the comparative merits of two hypotheses—one of imminent war, the other of no imminent war. The estimate is to be expressed in terms of the odds favoring or disfavoring the war hypothesis.
H. Bradford Westerfield; Inside CIA's Private World (1997)

Both parts of the dichotomy disfavor her success in school and society. One part of the dichotomy provides genetic reasons for her failure; the other provides sociocultural ones.
Richard Meyer; Official Portrais and Unofficial Conterportraits of 'At Risk' Students (2010)

Domestic conditions favoring or disfavoring successful operation of incentives in the recipient state are likewise critically important considerations for policymakers.
William Long; Economic Incentives and Bilateral Cooperation (1996)

  • Do any dictionaries consider these senses idiomatic enough to list? Oct 13, 2023 at 13:49
  • 1
    @EdwinAshworth I haven't found even one, including online. I'll look for some examples of this use outside the sciences.
    – DjinTonic
    Oct 13, 2023 at 14:03
  • 1
    @EdwinAshworth I forgot about Sven Yargs' dictionary find in Do dictionaries disfavor "disfavor"?
    – DjinTonic
    Oct 14, 2023 at 8:34

If we increase the sample size, we militate against certain kinds of error.

According to Cambridge Dictionary:

militate against sthg: to make something less likely to happen or succeed


"If we increase the sample size, we minimise (reduce to the smallest possible amount, extent, or degree, OED) certain kinds of error.


What about reduce? The Oxford Dictionary defines it as:

  1. Make smaller or less in amount, degree, or size.

If we want to be technical (of course we do) it doesn't explicitly carry the implication that it's less probable, just that it happens less, but this distinction is commonly glossed over.


If we increase the sample size, we avert the occurrence certain kinds of error.

avert the occurrence: to keep from happening; prevent TFD

  • This would normally be understood that to imply that the probability is reduced to zero, while the OP is looking for a word that stands for reducing the probability somewhat, but not all the way to zero.
    – jsw29
    Oct 13, 2023 at 15:05

conducive – making a certain situation or outcome likely or possible.

From Wiktionary:

conducive (comparative more conducive, superlative most conducive)

  1. Tending to contribute to, encourage, or bring about some result.

    A small, dark kitchen is not conducive to elaborate cooking.

  • 1
    You should include the source of the definition. Aug 28, 2021 at 15:07

From the industrial food safety sector, regulatory agencies in the domestic US favor vernacular terms such as follows in policy and guidance documents:

  • Control
  • Reduce
  • Mitigate

These are additionally associated with the following mechanisms, and prefixed by "critical" where compliance is mandatory or otherwise vital:

  • Limits
  • Parameters
  • Control Points

Lastly, there is an explicit expectation for quantification of all analyses conducted when determining the necessity for controls to be implemented at any particular process flow point for the mitigation of potential hazards. As such, the industry typically defines hazard risks as the arithmetic product of the likelihood (of occurrence) and the severity (of ramifications); assigning values that are plotted into a two-dimensional chart then enables the qualitative assignment of risk categories utilizing terminology such as "high risk", "low risk", etc., given the expectation that these are clearly defined in policy documentation.

Due to the nature of the industry, there's an obvious skew in favor of only one direction of influence for risk-based — i.e., preventative — controls. For data collection purposes, I've seen the use of terms such as innoculate when referencing the intentional introduction of, e.g., microbiological agents or contaminants.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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