Newspaper Subscribing: A Dynamic Analysis

Article excerpt

Variables related to stopping different from those related to restarting.

The post-war years have witnessed a decline in newspaper subscription per household in this country. The ratio of newspaper circulation to households dropped almost 50% from 1945 to 1985.1 Stempel has argued that a major reason for this decline is that the number of newspapers has declined, particularly in cities that had competing newspapers, and that a majority of the surviving newspapers have gained circulation.2 Nevertheless, this trend has generated serious concerns from both newspaper industry and social leaders. There are concerns about the loss of profits and the decay of public literacy and knowledge. Accordingly, mass communication scholars have made great efforts in describing and explaining why people don't read and subscribe to newspapers.

This article builds on this previous research, with two new contributions. First, it relies on a four-wave panel dataset rather than a one-shot survey. Second, it uses a dynamic modeling procedure rather than cross-sectional analysis. As the results show, newspaper subscribing behavior is more complex than static perspectives have suggested. There is more than one process going on in subscribing, and a dynamic model, tested with longitudinal data, improves our understanding of the complexity of these processes.

Static Comparison Problems

Static comparison is the fundamental approach used in the previous studies of newspaper subscribing.3 Typically, a one-wave survey is conducted and then the data are analyzed cross-sectionally. The main research interests are in the differences between socio-economic groups with regard to their readership or subscribing status. Such a comparison may be sufficient for descriptive purposes, but any attempt to go beyond that will invite problems.

First of all, static comparison requires an equilibrium assumption for the process under study, which means the process is constant over time rather than changeable. Unfortunately, this assumption is seldom met in the process of newspaper subscribing. Second, static comparison assumes that there is a single process caused by certain factors-e.g., people drop subscriptions because of social-economic disadvantages. As a matter of fact, there is usually more than one process underlying newspaper subscribing-some people drop out while others start or return. They may be two or more different processes, caused by different factors. Third, static comparison assumes that the causal factors are constant over time. It is true that some of the independent variables used in previous studies, such as sex and race, never change over time. On the other hand, there are approach is that it may obscure a substantial distinction between voluntary droppers and involuntary droppers by putting them together into the dropper group. Here the former refers to those who are involuntary starts, become non-subscribers even when no life-cycle transition happened to them, and may no longer come back; whereas the latter are the ones who are voluntary starts, then have to stop the subscription because of certain changes in personal life, but may return once the transition is over. If these two opposite trends do exist in real life, then a different analytical strategy is desired, which is the main task carried out in the following sections of this article.

The Allison Model

As stated before, this study is a secondary analysis of the ASNE subscription data. The basic strategy for analysis is a combination of two considerations: (a) differentiating the change of subscribing status into two processes-dropping out and coming back; and (b) giving more attention to the dynamic variables than to the static variables.

Operationally, the first consideration implies that two separate equations are needed, one for dropping and another for returning. When the two equations are built, a question naturally arises: Are the two processes determined by the same factors, in the same direction, and with the same strength? …