Risk versus Chance – What are the Odds? — Health and Communications

Roughly 5-to-1. But we’ll get to that in a minute.

First look at this common phrase…

______ of contracting [ disease name here ]

You might fill in the leading blank with any of these words, all of which are generally considered synonyms: risk, chance (or chances), likelihood, odds, possibility, probability. But even if synonyms, there are subtle –and significant– differences between them. And because words matter when discussing health issues like infectious diseases and cancer, that leads to three questions:

When considering health communications,

What words are favored by different global health organizations?
What words tend to be linked or associated with which diseases?
What word actually should be used?

1. What Words are Favored?

For a dozen organizations, Table 1 shows the number of times each word appears in the phrase “_______ of contracting” on their website. [ 1 ] What’s interesting here is reading across the rows. [ 2 ] Each organization shows a clear preference for using the word risk. It’s roughly five times as prevalent as the next most-used word, chance or chances (3425 vs. 708). And with the exception of WHO and CDC, the four remaining words –likelihood, odds, possibility, probability– are all in single digits; i.e., barely used at all.

2. What Words are Paired With Each Diseases?

Table 2 looks at these words from a different angle. [ 3 ] Here we see how they’re used across the internet, independent of any particular organization. So reading down the columns, the numbers show the results for “risk of contracting HIV” then “risk of contracting AIDS,” etc. Some observations:

Though, again, the word risk is most prominent, chance, chances, and possibility are far more often used on the internet generally than when tied to one of the selected organizations.
Only polio and zika are more frequently paired with chance/chances than with risk.
Odds and probability are, as before, used far less, with likelihood showing a little more usage.
And while TB and tuberculosis, not surprisingly, mirror each other’s numbers, the flu tracks very differently from influenza and flu.
Perhaps most interestingly, there are stark differences between the word usages applied to HIV and AIDS. Could it be that acquiring HIV is seen as the result of risk-related behavior while developing AIDS from HIV is viewed almost as a random or chance occurrence?

3. What Should We Use?

Of all these terms, risk is the only one whose definition makes it inherently negative. For example, you could speak of the chance, chances, odds, likelihood, possibility, or probability of wining the lottery, but you’d never speak of the risk of winning. As Merriam-Webster puts it, risk is “the possibility that something bad or unpleasant (such as an injury or a loss) will happen.” [ 4 ]

Chance frequently has the connotation of “small chance,” as in There’s a chance it might snow tonight. So it has an embedded, minimal quality to it.

Chances often has the opposite, maximal expectation, as in Chances are the ice cream has already melted.

Possibility is many times used (like chance) to suggest a “small possibility,” as in There’s a possibility they were mistaken.

Likelihood, much like its cognate likely, is skewed far to the end of the scale, implying that something will almost certainly happen. He was likely already gone / In all likelihood, he had already left.

Both odds and probability have a mathematics connection so you might think that they would be fairly neutral terms, but given the common phenomenon of “optimism bias” [ 5 ] in estimating likely outcomes, it’s more plausible that they actually convey less of a sense of danger than they should.

Given all of those connotations and common usages, risk seems like the most appropriate word to use in describing whether someone will acquire an infectious disease.

A Few Caveats

This brief look is done only in English. There are nuances and differences in other languages that might produce very different results.

Only Google’s search tool was used. Using Bing’s or Yahoo’s tools might make a difference, though we should all hope not.

The web is so dynamic, with so much new content being added, that the numbers in these two tables will almost certainly be different next week, and the week after that, and so on. (One estimate is that there are an additional 571 websites created every 60 seconds.) [ 6 ]

Using acquiring instead of contracting might similarly have changed the findings.

Selecting different organizations and other diseases might also shift the results, in which case the prominence of risk might be replaced with odds or chances, etc.

Conclusions

With those caveats in mind, risk is clearly the word most frequently used. For the organizations shown here, risk is almost five times more prevalent than chance or chances (–and the answer to the title question). For these diseases, across the entire internet, the disparity is only three times greater (312,940 to 102,759), though still in favor of risk. Given the connotations and common usage of the other terms, risk is also probably the best choice for describing the prospect of contracting an infectious disease. Given risk’s negative meaning it is also the most honest choice. [ 7 ]

There might be some interesting results if we looked at a closer level of detail; producing cross-tabs showing each organization crossed against each disease crossed against each of the keywords (12 organizations x 16 diseases x 6 keywords = 1,152 combinations!), which is too much for this site though certainly any organization could look at its own results.

Zika is much in the news lately but it will be interesting to see if over time the relative frequency of words used to describe it begins to align with those used for malaria, the other mosquito-borne disease in this list.

[ 1 ] It quickly became apparent that the search functions for each website were behaving quite differently. Some returned only a set number of hits (say, 30), some tried to be helpful by inferring interests and returning results based on synonyms, and some didn’t provide a total count of the returns. There was, in short, no way to compare across organizations by using their websites’ native search functions. So instead, I used the Google search formulation:

site:[website URL] [“term”]

as for example

site:www.who.int "risk of contracting"

[ 2 ] Reading down the columns and comparing organization to organization could be misleading unless each number were divided by the total number of pages on each site. Otherwise, you’d be left wondering if, for example, fhi360 used risk three times as often as MSF because they perhaps had three times as many pages on their site.

[ 3 ] Diarrhea admittedly is not a disease but a symptom of it and in this case is a stand-in for an entire class of water-borne and similar diseases that cause diarrhea. And unlike the others in this table, cancer is not infectious. What matters is that for both diarrhea and cancer, we do talk about contracting them and this post is looking at how we communicate that.

[ 4 ] http://www.merriam-webster.com/dictionary/risk

[ 5 ] See, for example, Dan Ariely’s discussion of “optimism bias” here: http://danariely.com/2009/09/05/the-curious-paradox-of-optimism-bias/

[ 6 ] See: http://blog.qmee.com/online-in-60-seconds-infographic-a-year-later/

[ 7 ] All that said, we're still pretty bad at estimating relative risk. See "10 Ways We Get The Odds Wrong" in Psychology Today: https://www.psychologytoday.com/articles/200801/10-ways-we-get-the-odds-wrong