Conditional Persuasion:
Uncertainty, information seeking, and persuasion online

Food for Thought

No current approach to persuasion explains the sudden and often huge shift in public sentiment that often accompany significant unexpected events. Recent examples of such events include both natural disasters like Hurricane Katrina, economic events like the 2008 stock market crash (more on this below), and acts of aggression like the 9/11 attacks on the World Trade Center. Less recent events include the assassinations of John F. Kennedy , the bombing of Pearl Harbor, and the onset of the Great Depression, The reader can doubtless identify other similar events. For the moment is is enough to note that each of these events provided the scenic backdrop for highly successful persuasive messages that substantially changed prevailing beliefs. All have, to at least some extent, acted as a pretext for initiating, with strong public support, wars and/or substantial changes in the structure of social systems.

The extreme effectiveness of persuasion in the wake of these kinds of events, especially in Germany during the run up to WWII, inspired much of the classic persuasion research of the 1950's and 1960's. While the collapse of the German economy and the burning of the Riechstag in the 1930's are clearly unexpected events which, like those noted above, led to a huge shift in public sentiment that that enabled Hitler's rise, the classic persuasion research of the 1950's and 1960's (including the work of Hovland, Festinger, Rokeach and many others) paid no real attention to such events, focusing instead on the psychology of persuasion and experimentally manipulable mechanics of the persuasive process. If, however, there is one clear outcome of the experimental research tradition in persuasion, it is that persuasion isn't generally terribly effective. In the laboratory, the effects of persuasive messages, even messages that are well constructed by any reasonable measure, fail to persuade most people, and the effects that can be measured are usually small, limited in duration, and difficult to convert into action in the real world. This should not be entirely unexpected. Most debates (arguments) in te real world end with neither party changing their position. Most real world persuasion is relatively ineffective.

This essay examines this contrast - the effectiveness of persuasion in the wake of unexpected events and the ineffectiveness of persuasion in general - based on a (so far) sixteen month ethnographic study of an online discussion group where persuasive messages occur on a daily basis. Most of this persuasion has been directly relate to the focus of the discussion group, which is focused on a single small cap stock that pays a healthy dividend. Some of the "persuasive" messages (many are poorly crafted) have involved other stocks, online products like stock market videos, and (of particular note recently) the Presidential elections, but most surrounds the stock itself. The persuasive messages are intermixed with informational and exploratory discussions of the stock, the market sectors the stock is located in, the broader economy, and the mechanics of investing and trading in the stock market. The discussion board was several years old when the observation started, and for the first several years of the boards existence, discussion was intermittent at best. Days and sometimes weeks would pass between new posts, in large part because there really wasn't all that much to discuss. There was no critical mass. If there was any kind of “community” associated with the board , it was chiefly composed of lurkers.

That changed in 2006 when the company made a controversial strategic acquisition. The value of the stock and the dividend both surged, creating interest in the stock. Posts to the online discussion group surged as well. Much of this volume was simple information seeking on the part of prospective investors and concerned investors. Some were looking for a shortcut past reading the prospectus, annual report, and other SEC filings that describe a companies condition and prospects. Others were for tactical and strategic insight from current investors and traders. The motivations of a number of other participants, including at least some of those who provided answers, was more murky. There were informational answers from satisfied investors and questions raised by investors who weren't happy about the acquisition. Some felt the stock and the dividend were safe. Others were concerned about developing economic issues that might undermine the value of the acquisition and the prospects of the company. Among the latter group, some reported selling their stakes. Less informative posts increased as well, including posts from “day traders”, "spammers", and "trolls". The traders were generally less interested in the long term prospects of the stock than the short term gains they could pick up by shorting the stock down or playing the volatility (buying low and selling high on price variations during the day. These traders often had a great deal more to gain from manipulating a stocks price up or down. Finally, there were a set of "trolls” who didn't seem to have any kind of stake in the company (for investment or trade), but had strong opinions about the stock and the factors that might drive its price.

Use of the terms troll and spammer isn't judgmental. These and other terms used here that could be regarded as judgmental (long, short, basher, and pumper, among them) are standard terms used on online fiancial discussion boards. They do imply judgment in those domains, but the judgments, especially of long and short, can vary considerably based on who is using them. The intent here simply to describe a range of players, some of whom are seeking information, some of whom are providing information, some of whom are providing informed opinion, some of whom are actively trying to persuade people to buy or sell for profit, some of whom are actively trying to manipulate people into reacting to their comments (trolls), and some of whom are trying to persuade people to visit other (generally commercial) sites or buy other stocks (spammers).

The online forum found critical mass in this mix of players. Discussion was a daily event and a considerable mass of lurkers emerged. Conventional wisdom suggests that online discussions don't have much effect on the stock market, but that “wisdom” has been cracking of late, and as the “pumping” (persuasive messages aimed at pushing up the price of the stock) gave way to “bashing” (persuasive messages aimed at pushing the price of a stock down), the price of the stock went down by over 50%, which brought even more people to the board looking for information and more traders and trolls to the board seeking to persuade an manipulate. It should be noted that the company was correctly reporting excellent results during this price fall. Hence the change was arguably prospective (based on speculation of future events) rather than reactive (based on actual company results. All of these activities expanded exponentially when the company had an unexpected failure in in a part of its business in the mid-2007. The participant observation starts with this failure and huge and sudden drop in stock price that accompanied it.

What we see, in these tactics, is a mixture of central path and peripheral path persuasive tactics. Those seeking and providing information were operating on the central path, largely attempting to make rational investment decisions in the face of the, depending on point of view, disasterous fall in price or major investment opportunity. Information seeking is a central path activity, and the answers provided will generally involve a mix of information and generally rational advocacy. Pumpers and bashers, by contract, are focused on peripheral path influence, and often focus on emotional appeals oriented around fear, greed, self-image, and credibility.

Its difficult to the full flavor of this in a short essay, but the comments of "bashers" often consist of very short fear-inducing messages along the lines of “If you are holding this POS you are stupid”, “Get out while you can”, “Timber!!!”, and "Company going Bankrupt". They are often combined with pointers to stocks that will do better (a distraction tactic that appeals to greed), attacks on company management and on individuals who take a position that they disagree with (credibility), and talking up their success (credibility). Information seeking is primarily a central path (deliberative) activity, but readers can as easily get caught up in this emotional rhetoric, which is often produced in large quantity, even if the reaction is limited to "why take a chance on this". The emotional persuasive tactics of the peripheral path appear to work online as well as they do in any other medium, and the same factors that drive information seeking may very well leave leave participants more open to this kind of persuasion.

Other elements of the stock market, including electronic elements, can be manipualted for persuasive effect. The most important of these is the price of the stock itself, which can change substantially on relatively small trading volumes. Most people don't appreciate the extent to which the price of a stock can be manipulated through market buy and sell tactics like short selling (which tends to drive the price of a stock down), especially when they are combined with media coverage that explains why the price has changed. The media coverage doesn't need to be accurate. Indeed, a great deal of the financial news that is published is little more than the opinion of credible stock market analysts, company executives, and fund maangers. The price of a stock is a message, and one this is often noticed by stock holders, and one that can be framed by media coverage in ways that amplify price movement.

Television financial analyst and former hedge fund manager Jim Cramer has described the mechanics of this amplification. A hedge fund managers or other credible analyst who wants to move the price of a stock in a particular direction will call up a reporter, tell them that there is a problem (or an unknown strength) at a company, provide enough compelling information to get a story published, and then follow the news with a big buy or sell order that appears to confirm the news. Buyers or sellers often follow, and the fund makes its money by either selling into the higher price or buying into the lower price. This isn't a pretend scenario. Cramer describes this as a tactic that he's executed successfully himself.

It appears that a variant of this tactic can be successfully executed using online discussion groups. The only difference is the direction of the activity. When a stocks price falls sharply (or rises suddenly) investors (and prospective investors) want to know why. A large manipulated drop in the price (or a large manipulated rise) will set off huge amounts of information seeking which, for good legal reasons (the risk of shareholder lawsuits), most companies will be unable to respond to in a timely fashion (or at all), frequently because they have no idea why the manipulated drop or rise has happened. If companies can't or don't provide useful information, however, people keep looking, and will generally have no problem finding it on online bulletin boards, as other investors, traders and trolls will be happy to supply them with explanations, including scary ones. If these explanations are accepted (even at the who wants to take a chance level), they help to push the price of the stock down further. This can be a vicious cycle, as each decline in price is likely to lead to more information seeking and the exposure of more people to these various forms of influence.

This usually doesn't happen all at once, but an unexpected good or bad news announcement can put these processes into overdrive, and levels of information seeking and peripheral path persuasive messages suddenly spiking at the same time that the price of the stock drops by a huge amount all at once (what is referred to as a "falling knife") in accelerated trading. When this happened to the company discussed on this board, bashers were quick call the company's management is incompetent, to suggest that the dividend was about to be cut to zero, and predict that the company would soon declare bankruptcy. Investors were urged not to "catch a falling knife". It didn't matter that the company had no way to predict or control the negative event, that the loss represented a very small part of their portfolio and earnings, and that the companies cash flow was secure. Truth is peripheral when a stock is being "bashed". Similarly outlandish positives will be suggested when a stock is being "pumped".

Axiom 3 of Charles Berger's uncertainty reduction theory, which may be that theories most general statement, seems to apply here: “High levels of uncertainty cause increases in information seeking behavior. As uncertainty levels decline, information-seeking behavior decreases.” This relationship is a readily observed on stock market discussion boards. When stock prices change significantly, information seeking is readily visible as board volume suddenly increases and people seek information so they can decide whether to take action. This axiom appears to be extremely valid for this and other stock market discussion boards, where large and sudden changes in a stocks price (which can be reasonably be presumed to create uncertainty) seem to substantially increase traffic, including both information seeking and persuasive messages..

The persistence of this increase in persuasive messages across financial boards suggests that the people who post them regard these messages as effective, and additional changes in stock prices suggest that they probably are effective. This suggests a parallel hypothesis: “High levels of uncertainty make people more persuasible. As uncertainty levels decline, persuasibility declines” that will be treated as as second "axiom" for the remainder of this discussion. Uncertainty not only leads us to seek information that will help resolve uncertainty, but opens us to being persuaded by those who provide that information. This axiom/hypothesis appears to be extremely valid for this and other stock market discussion boards. Large changes in price seem to substantially increase persuasive activity, including central path discussion (the provision of answers and pointers to useful information) and peripheral path discussion, including strong emotional appeals.

Taken together, these principles provide a the basis for a different view of the persuasive process which, for the purposes of this essay, will be called “conditional persuasion”. It pairs a "punctuated equilibrium" model of persuasion with a mechanism through which that persuasion occurs. Specifically

  1. when we are faced with uncertainty about something we care about,
  2. we generally seek information to resolve that uncertainty, thus
  3. increasing the power of those who control that information to persuade us.

It may well be that primary reason why we can't demonstrate strong and enduring persuasive effects in the laboratory is that persuasion depends on the occurrence of destabilizing events that provoke uncertainty and information seeking; that persuasion depends on our willingness to actively seek information that resolves the uncertainly associated with unexpected events. All of the large scale events cited in the at the top of this essay are large scale destabilizing events that caused people to actively seek information in ways that opened them to persuasion. They demonstrate that strong and enduring persuasive effects are possible when destabilizing events occur.

The observations of this discussion board and the market activity that has accompanied discussion strongly suggests that it is possible to observe this kind of event on much smaller, and much more routine, scale. All that is required is an event that creates substantial uncertainty about a specific stock and a willing trader or troll with a cogent explanation. It is possible to identify any number of stocks that have had their prices driven down to zero by such tactics, even when the companies were profitable and/or had products that were likely to be profitable in the near future. Hence it is may be possible to observe real world persuasion happening on these discussion boards and to observe the real effects of that persuasion on stock price and the viability of the company as a continuing entity. Doing that will require a more complex level of content analysis than we normally do. Transcripts, and analyses of the contents of those analyses, will be have to be tied to price of the stock and (as a control) the movement of related stock market indexes.

It remains, however, that the conditions that lead to information seeking and persuasion in the real world can be detected as they happen. That sets up timely observation of online discussions that are likely to be the target of subsequent information seeking and persuasive messages. Its difficult to do this kind of research when large scale events occur. It is much easier to do when the events are frequent and detectable, as they currently are for the stock market. Its a direction of research worth exploring, if only because there are important implications to the conditional persuasion model presented in the two axioms above.

I will select just a small set of implications that are worth considering as I close this essay:

  1. There is a single major precondition that is most likely to make a persuasive message effective: an audience that is willing to both consume the message and be persuaded. That willingness is signaled by information seeking behavior.
  2. Once that condition is satisfied, almost any conventional approach to persuasion is likely to work. What matters more than anything is having a way to get that message to information seekers.
  3. In a major crisis, that role belongs to the mass media and those with access to the mass media. As an example, in the current financial crisis, four people had primary access to that Microphone: the President, the two Presidential candiates, and the head of the Federal Reserve. The person who used that Microphone most effectively (with, as it turns out, a combination of central path and peripheral path appeals) subsequently won this years Presidential election.
  4. For smaller scale events like a decline in the price of a small cap stock, that microphone will be located at the largest scale that, in the absence of mass media coverage, can reach a large number of users. For many smaller companies, that level appears to be online discussion boards and blogs.
  5. The real key to being an effective persuader is probably summarized by the Boy Scout motto: “Be prepared.” It may be that no amount of advocacy for a particular position will really matter until people are willing to listen to it, but once they are, large scale attitude and behavioral change can be readily achieved.

Food for thought.

Davis Foulger

Additional discussion of this and other Food for thought articles can be posted in the HCTD Moodle learning environment. In this case the author is particularly interested in discussing cooperative research approaches that might be taken in testing these ideas.

Do you have a response to Davis’s pundits? Have something you would like to ponder in this column?  Please submit your responses or Food for thought toto John Howard at