Primacy-Recency
A line of message strategy research deals with the order that the most important elements in the message should occupy. The order of information not only has an impact on retention but also on attitude change. The researcher has demonstrated a primacy effect if the element placed first in the message has the greatest effect. It is a recency effect if the element placed last in the message has the maximum effect. The research model that is frequently used presents messages having the same topic, but differing arrangements, to comparable audiences, and then see which audience has the most response. All other factors being equal, any differences in responses must be due to different types of order that the messages utilize.
Early Primacy-Recency Studies
Interest in the problem of order of presentation in the type of communication situation where both sides of a controversial issue are presented goes back to 1925 of a study by Lund . On the basis of his research he enunciated a Law of Primacy in Persuasion stating that the side of an issue presented first will have greater effectiveness than the side presented subsequently. He gave his classes of college students a mimeographed communication in support of one side of a controversial issue (e.g. " a protective tariff is a wise policy for the United States") and then presented a second communication advocating a diametrically opposed stand on the same issue. He discovered that the communication coming first (whether pro or con) influenced the students more than the one coming second. Other studies by Jersild (1928) and Knower (1936) supported a primacy effect. These early studies though used two complete messages and varied the order of presentation.
Subsequent experiments however yielded different results. Cromwell (1950) performed a study in which affirmative and negative speeches on socialized medicine and labor arbitration were presented to groups of students whose opinions were measured before and after the talks. A significant "recency effect" was produced in favor of the side of the issue presented last.
To explain the contradictory results, Hovland and Mandell (1957) tried to replicate the Lund experiment using some of the original topics that were employed. Their results indicated that only one of the three groups tested showed primacy effect while the other two groups showed slight recency effects. All these studies used written communication and Hovland and his associates concluded "when two sides of an issue are presented successively by different communicators, the side presented first does not necessarily have the advantage." Also in a series of experiments Hovland, Campbell and Brock (1957) concluded that if after hearing only one side of a controversial issue, the opinion is expressed publicly in favor of that, the effectiveness of the second side is reduced, thus entailing a primacy effect. Thus, commitment to an opinion (Lund described this as 'the ideal of consistency') after hearing just the first part of an argument leads to a primacy effect.
Law of Primacy or Recency?
Despite large amount of research work on Primacy-recency problem, instead of a general "law" of primacy or recency, we have today and assortment of miscellaneous variables. Some of these tend to produce primacy ("primacy-bound variables") others produce recency ("recency-bound variables"). Still others produce either order effect, depending on their utilization or temporal placement in a two-sided communication ("free variables") (Rosnow 1967). Nonsalient, controversial topics, interesting subject matter and highly familiar issues tend towards primacy. Salient topics, uninteresting subject matter, and moderately unfamiliar issues tend to yield recency. If arguments for one side are perceived more strongly than arguments for the other, then the side with the stronger argument has the advantage-"strength" being a free variable. Another free variable is "reinforcement" (Rosnow 1967). The various attempts to fit these diverse relationships between the variables and the effects into a theoretical framework have not met with much success. But two of these frameworks within which to predict order effects in persuasion are presented below.
Anderson model for Opinion change
This model (Anderson 1959) rests on the assumption that the more opinion change asked for, the more received. Hence, if two successive communications produce proportionally the same amount of opinion change, the second communication should always have the advantage. This is because the first communication would move opinions a given amount, thereby increasing the attitudinal distance between the recipient and the second communication. If the second communication were proportionally as effective as the first, then, because it demanded more opinion change, it would have the effect of producing greater change. The primacy effects and discrepancies in certain results was explained by a two-component hypothesis. According to this hypothesis, the opinion had a superficial component which obeyed the model, and a basal component, which was largely resistant to change once it had come into being.
Miller and Campbell Model
Miller and Campbell (1959) derive their model from the work carried out by Hermin Ebbinghaus in the 19th century. Their paper takes the Ebbinghaus forgetting curve and shows how they can be used to predict primacy and recency effects in persuasion. Their studies show the possibility of a thoroughly general recency principle, with the strength of recency being maximal for a long delay between first and second communication coupled with immediate measurement after the second communication, the strength of recency being minimal when the two presentations are contiguous and the testing delayed. Also the postulation of a general recency principle also improves the case for a general primacy principle. A significant primacy effect was found under the conditions predicted to be least favourable to recency. Miller and Campbells' work (apart from some experiments that produced primacy effect) has also been confirmed by Insko (Insko 1954).
With this we come to the end of our review of the primacy-recency problem with a knowledge that the Law of Primacy-recency in Persuasion is dependent on a number of variables and primacy and recency are achievable under a certain set of conditions.
A word of caution
The primacy and recency effects on attitudes are not to be explained in terms of a primacy or recency effects upon learning or memory. Similarly it cannot be explained in terms of such factors as set, attention or reinforcement (Hovland 1957).Lund's factor of "primacy in persuasion" which we have talked about and built on through out should be distinguished from "primacy" in serial learning. The primacy and recency effects obtained on a serial position curve (which is typically U-shaped because of first and last items being better recalled) as a result of a free-recall experiment are different than the "Primacy-recency in persuasion" that we are discussing here.