Negativity Dominance

What it is

The term comes from Rozin and Royzman's paper "Negativity Bias, Negativity Dominance, and Contagion", published in Personality and Social Psychology Review, volume 5, issue 4, pages 296 to 320. The paper treats negativity bias as a family of related asymmetries and names four of them. Negative potency is the claim that a negative event, object or trait has more impact than a positive one of equal magnitude. Steeper negative gradients is the claim that negative things feel increasingly bad faster as a person gets closer to them in space or time than positive things feel increasingly good. Negative differentiation is the claim that negative entities are more varied, support more complex concepts and provoke a wider range of responses. Negativity dominance is the fourth, and it concerns combination.

When a positive element and a negative element are put together, there are two things a judge could do. The judge could weigh each by its size and add the results, which would give a mixture of equal good and bad a neutral value. Or the judge could let the negative element count for more, which would pull the mixture below neutral. Negativity dominance is the claim that people do the second. The overall evaluation of the combination is more negative than the simple sum of the individual values would predict.

The same paper also suggests that negative events tend to dominate partly because they are more contagious than positive ones, which is the contagion in its title. That part of the argument is outside the scope of this page and of the article that uses the term.

Rozin and Royzman published in the same year as the broader review by Roy Baumeister and three colleagues, "Bad Is Stronger Than Good", in Review of General Psychology. The two papers are usually read together. The Baumeister review gathers evidence for the greater power of bad events across a wide range of everyday and major life events, and ends its abstract by saying that hardly any exceptions can be found. The Rozin and Royzman paper is the one that supplies the vocabulary of separate elements, which is useful because "bad is stronger than good" can mean several different things.

In effect

The place where negativity dominance matters for ordinary readers is the averaging of opinions. A product with a run of good reviews and one bad review is a mixture. If people judged mixtures by adding values, one bad review among five good ones would barely move the overall impression, because the good ones would carry it. The saying that one bad review outweighs five good ones is a way of claiming the opposite, and negativity dominance is the part of the psychology that gives the claim a footing.

The nearest evidence from commerce is a 2006 study by Judith Chevalier and Dina Mayzlin in the Journal of Marketing Research. They compared reviews and sales of books at Amazon.com and BarnesandNoble.com. Their abstract reports that reviews at both sites were overwhelmingly positive, that an improvement in a book's reviews raised its relative sales at that site, and that one-star reviews had a larger impact than five-star reviews. That is the direction negativity dominance predicts. A bad rating moved sales further than a good rating did.

The study is evidence for a direction and nothing more. It compares the effect of two ends of the star scale on sales of books at two retailers. The abstract does not report a ratio between them, and nothing in it turns "larger" into "five times larger". The popular phrasing, one bad review against five good ones, is a figure of speech that borrows the authority of the finding without the number.

A further reading of why the asymmetry might appear in reviews comes from the accessibility-diagnosticity account that Paul Herr, Frank Kardes and John Kim used in 1991 for word of mouth. On that account, information counts when it comes easily to mind and seems to help the judgement. A specific complaint may read as a better guide to how a product will fail than general praise reads as a guide to how it will succeed. That step is this site's own reading and is not a result from Rozin and Royzman or from Chevalier and Mayzlin. It is offered as one candidate explanation of the outcome that negativity dominance describes.

What it does not say

It does not give a ratio. Neither the Rozin and Royzman paper as summarised here nor the Chevalier and Mayzlin abstract reports a number of good reviews that one bad review cancels. Anyone quoting five to one, or any other figure, for reviews is not quoting these sources.

It does not say the asymmetry holds in every consumer setting. Rohini Ahluwalia's 2002 paper in the Journal of Consumer Research asks how prevalent the negativity effect is in consumer environments, and its abstract says that the effect's robustness there may be overstated. Her point is that earlier studies mostly asked people to judge unknown or hypothetical targets with the aim of forming an accurate impression, while shoppers often know the brand and may be trying to manage an impression or defend a view they already hold. Herr and colleagues found in a second experiment that the word-of-mouth effect weakened or disappeared when people already held an impression of the brand. Familiarity is the variable both consumer papers point to.

It does not say that good information is inert. The claim is about relative weight when the two are combined.

It does not settle why the asymmetry exists. The paper names the elements and the Baumeister review collects the evidence. Diagnosticity, salience, contagion and plain emotional impact are all candidate explanations, and this page does not choose between them.

---