|
"Extending Validation Efforts to the Concept of News Use"
By William P. Eveland Jr., Myiah Hutchens Hively, and Fei Shen
Given the centrality of the concept of news use in political communication, it seems appropriate that efforts to better understand the implication of various conceptualizations and measurement approaches be undertaken. The work of Althaus and Tewksbury (2007) does an important service to the field by evaluating the validity of measures of media exposure. We applaud their efforts and mostly agree with their conclusions – in particular that (a) news use and political knowledge are distinct (though related) constructs and so both should be measured independently; (b) the variety of exposure measures in the ANES should be expanded to more media forms and sources; and (c) the measurement of minutes per day may be abandoned. We also believe that measures of motivation and information processing in the ANES should be expanded, although we might quibble with exactly what measures should be incorporated. In this commentary, however, we would like to focus on the argument that much more validation work needs to be done on measures of news use. Althaus and Tewksbury focus on measures of exposure (as does much of the other work validating ANES measures), but the common application of both exposure and attention measures in the political communication literature means there should be an expansion beyond validating exposure measures.
In a commentary article in Political Communication a few years ago, we offered some observations about the need to pay greater attention to how various combinations of indicators of news exposure and attention were used (Eveland, Hayes, Shah, & Kwak, 2005). We pointed out that inconsistent findings regarding the relationship between news use and political knowledge across several studies in that particular issue of Political Communcation, and the literature as a whole, might very well be due to different measurement and analytic strategies.
These concerns led us more recently (Eveland, Hively, & Shen, 2007) to address this issue empirically by examining the internal consistency, dimensional structure, and validity of measures of frequency of exposure and attention to various news media sources as predictors of political knowledge. We also considered data analysis issues that follow from the measurement issues. In that paper, we focused on political knowledge as an important validity criterion for news use measures, although we acknowledged that other criteria, such as political participation, may be equally important.
In order to tap the extent of measurement inconsistencies, we surveyed eight major journals in communication and political science that we expected to publish articles linking news media use and political knowledge. We searched the Social Science Citation Index database using the search phrase “‘media’ or ‘knowledge’” so that any study examining either knowledge or media use between 1997 and 2007 would be identified for more detailed consideration. All articles identified by the database search were examined, and a total of 32 studies were found that met our criteria.
This content analysis revealed that although the plurality of studies employed only exposure measures, a quarter of the studies used only attention measures and a third of the studies included measures of both exposure and attention. Thus, the majority of studies employed something beyond mere exposure measures.
Among the studies with exposure items available, 31% combined the indicators into some form of “exposure” scale. Among the studies with attention indicators available, 26% combined the items into some sort of “attention” scale. Among the 11 studies with both attention and exposure items available, five constructed combined exposure/attention scales and six did not. Thus, our study concluded that there is no single dominant news media measurement approach in the literature linking news use and political knowledge.
Our study also examined how news variables were used in hypothesis testing – typically in OLS regression models. Obviously, when there are multiple indicators of a given construct (or constructs) one can employ data reduction techniques to reduce the number of predictors in a model, or one can use the individual indicators as predictors. We found that approximately half of the studies (56%) employed multiple single-item indicators for news exposure or attention in their regression models.
What are the implications of these variations in approaches? As we had surmised in our 2005 Political Communication commentary article, analysis of survey data we gathered for our project found that different measurement approaches – using different indicators, and combining them in different ways – led in many cases to different substantive conclusions about the fundamental relationships under consideration. Thus, measurement does matter. And, it appears that the best measurement using common exposure and attention indicators – themselves, we believe, insufficient – was to combine matching measures of exposure and attention to a given media form or source.
These findings – and additional results reported in Eveland, Hively, and Shen (2007) – suggest that report produced by Althaus and Tewksbury (2007) to validiate measures of news exposure must be supplemented with conceptualization and empirical work on the broader construct of news use. This is not a criticism of Althaus and Tewksbury; the lack of matching exposure and attention indicators in the ANES meant they simply could not do this. So, researchers will likely have to gather their own original pilot data or encourage the ANES to modify and expand its measures to include matching exposure and attention indicators so that further validation work may be done.
The process of validating concepts and developing good measures often goes unnoticed – or simply undone. But, it is absolutely critical this research be conducted so that the field does continue to build its foundations on shaky ground. The work of Althaus and Tewksbury (2007) fortunately places this research on the top of the agenda for political communication scholars. We encourage more scholarship in this vein, expanding the concept beyond exposure to the more general concept of news media use. Doing so will, we hope, lead to more productive theorizing and more valuable and cumulative empirical findings.
William P. Eveland is an Associate Professor at The Ohio State University School of Communication. Myiah Hutchens Hively and Fei Shen are graduate students at The Ohio State University School of Communication.
REFERENCES
Althaus, S. L., & Tewksbury, D. H. (2007). Toward a new generation of media use measures for the ANES. Report to the Board of Overseers, American National Election Study.
Chaffee, S. (2001). Studying the new communication of politics. Political Communication, 18, 237-244.
Eveland, W. P., Jr., Hayes, A. F., Shah, D. V., & Kwak, N. (2005). Observations on estimation of communication effects on political knowledge and a test of intracommunication mediation. Political Communication, 22, 505-509.
Eveland, W. P., Jr., Hively, M. H., & Shen, F. (2007, November). Exposure, attention, or “use” of news? Validating measurement of a central concept in political communication and public opinion research. Paper presented at the annual meeting of the Midwest Association for Public Opinion Research, Chicago, IL.
McLeod, J. M. (2001). Steven Chaffee and the future of political communication research. Political Communication, 18, 215-224.
|
