Katz and Lazarsfeld: Two-Step Flow

From :
http://www.cultsock.ndirect.co.uk/MUHome/cshtml/media/kl.html

The study of the 1940 election campaign

In 1940, Lazarsfeld, Berelson and Gaudet conducted the first full-scale investigation of the effects of political mass communication. Their research focused on the 1940 Presidential election campaign and their findings were published in 1944 in The People's Choice after more research had been conducted.

The importance of social influence

Their research was originally based on something like the simplistic hypodermic needle model of media influence, whereby it was assumed that a message would be transmitted from the mass media to a 'mass audience', who would absorb the message. However, their investigations suggested that media effects were minimal, that the conception of a 'mass audience' was inadequate and misguided and that social influences had a major effect on the process of opinion formation and sharply limited the media's effect.

Limited effects

The study by Lazarsfeld et al concluded that only some 5% of people changed their voting behaviour as a result of media messages. Their exposure to election broadcasts turned out to be a relatively poor predictor of their voting behaviour, particularly when compared with other factors such as their interpersonal communication with friends, union members, business colleagues and the political tradition they had grown up in. This view of media effects was confirmed n a variety of other investigations and came to be known as the 'limited effects paradigm' of media influence.

Two-Step Flow: general conclusions
Consequently Lazarsfeld and his colleagues developed the notion of a 'two-step' flow of media messages, a process in which opinion leaders played a vitally important r?e.

This was later developed by Katz and Lazarsfeld and presented in their book Personal Influence (1955). A number of significant conclusions follow from their research:

our responses to media messages will be mediated through our social relationships, the effects of media messages being sharply limited by interpersonal relationships and group membership (this is confirmed also by Hovland who identifies our adherence to group norms as a major factor; see also the more general sections on Social Influence)
it is misleading to think of receivers as members of a 'mass audience' since that implies that they are all equal in their reception of media messages, whereas in fact some play a more active r?e than others
receiving a message does not imply responding to it; nor does non-reception imply non-response (since we may still receive the message via interpersonal communication)
there are some people amongst the media audience who act as opinion leaders - typically such people use the mass media more than the average, mix more than the average across social classes and see themselves and are seen by others as having an influence on others
Reasons suggested for the greater effectiveness of personal influence over media influence include the following:

The content and development of a conversation are less predictable than mass media messages. Consequently, the receiver cannot be as selective in advance as (s)he is able to be when choosing which media messages to attend to.
In a face-to-face conversation, the critical distance between the partners is less than in mass communication.
By direct questioning of the partner in the conversation, the assumptions underlying the conversation can be rapidly and accurately established, which is not so with mass communication.
In face-to-face interaction the communicator can rapidly adjust to the receiver's personality. (S)he has direct feedback as to the success of the communication, can correct misunderstandings and counter challenges.

Criticisms

The model is often presented graphically as shown on the right. In fact, that is somewhat misleading as it suggests that mass media messages flow first to opinion leaders and from them to the rest. Obviously, that's not the case, since you and I can both receive messages directly. The point is that the messages we receive are then modified through the pattern of our social contacts.

Katz and Lazarsfeld are perhaps also somewhat misleading when they suggest that individuals with certain characteristics are opinion leaders. It may be the case that many opinion leaders will have the characteristics they mention, but we also know that some opinion leaders in some subject areas will not have those general characteristics. However, I should mention that Katz and Lazarsfeld certainly did not take the view that opinion leaders were necessarily those formally recognized as such (e. g. celebrities, politicians etc.) Thus, their studies showed that top-down influence was relatively slight. Influence tended to be horizontal across a particular socio-economic class, except that in the 'higher' social classes there was a tendency for people to find opinion leaders in the next class up. No opinion leader was an opinion leader in all aspects of life. For example, the car mechanic in your local pub may not use the media much at all because he's always working late. Nevertheless, he knows a lot about cars and so what the rest of those in the pub 'know' from the media about different makes of car will be influenced by his views. Similarly, your Politics lecturer may not use the media anything like as much as you do, but her reading and viewing is targeted on political issues. Together with her broad knowledge of political theory and history, that is likely to make her an opinion leader as far as your Politics class is concerned. Allowing for those differences from one class to another and from one subject area to another, we probably can recognize in opinion leaders the characteristics which Katz and Lazarsfeld suggested, in particular that opinion leaders will be more active users of the mass media than others.

Katz and Lazarsfeld may also be misleading in suggesting that people are either active opinion leaders or passive followers of opinion leaders. Apart from the evidence that people can be opinion leaders on some matters and not on others, there is also the objection that some people may be neither leaders nor followers, but quite simply detached from much media output.

Much depends also on the accessibility of countervailing opinions. In the 1940s the general public would have had access to far fewer sources of information than they have today and may, broadly speaking, have had less time to access those sources. Under such circumstances it is likely that an opinion leader in the community may be especially influential. This was recognized by the Nazi party in its gradual rise to power during the 1920s and 1930s. Nazi agitation and propaganda became increasingly successful at forcing themselves onto the front pages of newspapers, thus becoming an everyday topic of conversation. They were particularly keen to capitalize on that attention, directing it in the right direction through influencing the leading members of the various small associations which were spread throughout German communities.

Where local leaders, enjoying respectability and influence, were won over, further converts often rapidly followed. In the relatively homogeneous villages in Schleswig-Holstein, where feelings about the 'Weimar system' were running high on account of the agrarian crisis, the push from one or two farmers' leaders could result in a local landslide to the NSDAP [the Nazi Party].

Kershaw (1999 : 321)

Katz and Lazarsfeld's Influence
Despite those and other criticisms, the fact remains that Katz and Lazarsfeld's research is widely accepted and still highly influential. Advertisers and spin doctors recognise that 'the best form of advertising is word-of-mouth advertising'.