Accessibility-Diagnosticity Model

What it is

The model rests on two properties of a piece of information. Accessibility is how readily it comes to mind when a judgement is made. Diagnosticity is how far it seems to help tell one answer from another. A fact can be easy to recall and useless for the question at hand, or highly useful and out of reach. On the model, information carries weight in a judgement when both conditions are met, and carries little when either is missing.

The reason the model is attractive for consumer research is that it gives a single mechanism for effects that are usually listed separately. In outline, the argument runs as follows. A vivid account is more accessible than a pallid one, so it can outweigh larger bodies of dull information. Information formed earlier can persist, because it is already in memory when the question arises. Negative information can count for more if people treat it as better evidence about which of two outcomes to expect. The accessibility-diagnosticity perspective puts all three under the same heading.

Herr, Kardes and Kim applied it in a paper in the Journal of Consumer Research, volume 17, issue 4, pages 454 to 462, titled "Effects of Word-of-Mouth and Product-Attribute Information on Persuasion: An Accessibility-Diagnosticity Perspective". The authors argue in the paper that the vividness, perseverance and negativity effects can all be explained by the model. That is the part of their argument that matters for this site, and it is stated here as their argument, not as an established result.

In effect

The paper reports two experiments, as summarised from its abstract. In the first, word-of-mouth communication delivered face to face was more persuasive than the same kind of information in print. In the second, the word-of-mouth effect weakened or disappeared when people already had an impression of the brand in memory. Both results fit the model. A face-to-face account is likely to be more accessible than a printed line, and an existing brand impression changes how much a single new report seems to add.

The second result matters beyond the paper. It places the strength of word of mouth in the hands of the listener's prior knowledge, which is the same variable that Rohini Ahluwalia identified in 2002 when she asked how prevalent the negativity effect is in consumer environments. Her abstract notes that earlier studies mostly had people judge unknown or hypothetical targets, while shoppers often know the brands they are judging. Read together, the two papers say that the effects measured in the laboratory on unfamiliar targets may shrink where the buyer already holds a view.

The model also offers a candidate reading of the finding that one-star reviews moved book sales further than five-star reviews, reported in 2006 by Judith Chevalier and Dina Mayzlin. Their abstract does not give a mechanism. If negative information is treated as more diagnostic, a specific complaint about how a product fails could be more useful to a reader than general praise is. This site presents that as a reading of the evidence and not a result from either paper. It has not been tested in the studies cited here, and other mechanisms, including plain emotional impact, remain possible.

What it does not say

It does not say negative information is always more diagnostic. In the model, diagnosticity is a property of how information is perceived in a context, and the context can change it. The paper's own second experiment shows word-of-mouth effects fading when a prior impression exists.

It does not explain by itself why bad should be stronger than good. The Baumeister review of 2001, "Bad Is Stronger Than Good", collects the evidence for the asymmetry across many areas, and the Rozin and Royzman paper of the same year names its elements. The model is one proposed route to the asymmetry in consumer judgement, not the only one.

It does not give numbers. No ratio of good to bad information comes out of the account, and the popular claim that one bad review outweighs five good ones is not derived from it.

It does not settle its own history. This page names no originator because none was confirmed. Herr, Kardes and Kim apply the model to word of mouth, and that is as far as the sourcing here goes. Anyone who needs a credit line for the general framework should trace it before using one.

It is not the same as selective accessibility, the account Mauricio Palmeira and Joydeep Srivastava give of why a free offer holds its value, though both lean on the idea that what comes to mind first shapes judgement. It is also not the same as the vividness effect described elsewhere on this site, in which a single vivid story outweighs statistical evidence, although Herr and colleagues list vividness among the effects the model explains.

---