The Impact of Published Conference Proceedings on Marketing and Consumer Research

James C. Anderson, University of Texas
George T. Haley, University of Texas
ABSTRACT - Citation analysis was used to address the question of what has been the impact of published proceedings papers upon the areas of consumer research and marketing. The Journal of Consumer Research, Journal of Marketing, and Journal of Marketing Research for the years 1975 through 1982 were the sources of citations. While the absolute number of total proceedings citations has increased over time, their percentage of reference citations has remained fairly constant. Considering citations for the primary marketing-related proceedings, ACR proceedings have received a significantly larger number of citations, and have had a significantly positive trend over time.
[ to cite ]:
James C. Anderson and George T. Haley (1984) ,"The Impact of Published Conference Proceedings on Marketing and Consumer Research", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 133-136.

Advances in Consumer Research Volume 11, 1984      Pages 133-136

THE IMPACT OF PUBLISHED CONFERENCE PROCEEDINGS ON MARKETING AND CONSUMER RESEARCH

James C. Anderson, University of Texas

George T. Haley, University of Texas

ABSTRACT -

Citation analysis was used to address the question of what has been the impact of published proceedings papers upon the areas of consumer research and marketing. The Journal of Consumer Research, Journal of Marketing, and Journal of Marketing Research for the years 1975 through 1982 were the sources of citations. While the absolute number of total proceedings citations has increased over time, their percentage of reference citations has remained fairly constant. Considering citations for the primary marketing-related proceedings, ACR proceedings have received a significantly larger number of citations, and have had a significantly positive trend over time.

Since 1971, the Association for Consumer Research has published the proceedings of its annual conference. In fact, most marketing-related annual conferences put out published proceedings. The question addressed in this paper is what has been the impact or value of published proceedings papers upon the areas of consumer research and marketing?

The criterion measure of impact which was employed was the citation of published proceeding papers in research articles, notes, and comments in the primary marketing journals The Journal of Consumer Research, Journal of Marketing, and Journal of Marketing Research. While there are other potential measures of value such as extent of readership, citation analysis (Broad 1978) has been used to address related research questions. As Drew and Karpf (1981, p. 307) have stated, "A growing literature within the sociology of science deals with the analysis of publication and citation rates. Publication rates are a straightforward index of productivity, whereas citation rates indicate the impact of an author's publications upon professional colleagues." Robinson and Adler (1978) have employed citation analysis for an assessment of the "eminence" or relative impact of marketing scholars and marketing institutions; Cox, Hammelman and Wilcox (1976) have used citation analysis to obtain a perceptual map relating business journals to one another; and the Social Science Citation Index (Institute for Scientific Information, 1982) lists published articles of researchers followed by the articles in a given period which have cited them.

Within the broad question given before, several specific questions were asked. These were:

1. What relationships, if any, are there between time in years and journals, and the number of proceedings citations?

2. Do these relationships change when self-citations are excluded?

3. What relationships, if any, are there between time in years and journals, and number of citations for each of the primary marketing-related proceedings?

4. Do these relationships change when self-citations are excluded?

METHOD

Sampling Frame

References for research articles, notes and comments appearing in The Journal of Consumer Research (JCR), Journal of Marketing (JM), and Journal of Marketing Research (JMR) were the sources of citations. The years 1975 through 1982 were used as this time frame represented years for which there was complete data for the set of journals.

Procedures

In setting up the coding categories, proceedings considered to be the primary ones in the area of marketing were broken out for separate counts with the remainder of the proceedings lumped into an "other" proceedings count. The primary marketing-related proceedings" were considered to be the annual conference published proceedings of the Association of Consumer Research (ACR), the American Marketing Association (AMA), Division 23 (Consumer Psychology) of the American Psychological Association (APA), and the American Institute for Decision Science (AIDS). As would be expected, the "other" category count was composed of a diverse group of proceedings such as proceedings from annual conferences of other disciplines (e.g., Sociology, Mathematics, Computer Science), regional conferences (e.g., Southwestern Marketing Association), smaller special interest conferences (e.g., Conserver Society Conference), and foreign conferences.

While for the purposes of this research only a count of "other citations" was needed to form the citation outcome variable and to obtain a total citation count, separate citation counts were also made for books, journal articles, unpublished manuscripts/working papers (including dissertations/theses, and research reports by private firms) and "other" other citations such as court cases. These separate counts provided For better auditing and accuracy of the data. For each count discussed, a related count was made where self-citations were excluded.

These citation counts were performed for each journal by year. After the counts were completed, a data check was conducted by duplicating the citation process described.

Analyses

Trend line correlations were computed between time in years and total proceedings citations, total proceedings citations excluding self-citations, and citations for each of the primary marketing-related proceedings. Correlations were also computed between each proceedings citation count over tine and its respective count minus self-citations.

A three-way contingency table was formed with proceedings citations versus other citations as a dichotomous outcome variable, and journal and year as predictor variables. Log-linear/logit models (Green, Carmone and Wachspress 1977; DeSarbo and Hildebrand 1980) were fit to this table using maximum likelihood logistic regression. In addition to the likelihood ratio chi-square statistics, the relative decrease in weighted unexplained variation (Goodman 1971), analogous to the usual coefficient of multiple determination (R2), was calculated for a main-effects model as a comparison to the model including the journal x year interaction (the saturated model). These analyses were repeated for the outcome variable with self-citations excluded.

Finally, chi-square tests (Siegel 1956) were computed for two-way contingency tables of interest (e.g., citations for primary marketing-related proceedings by journals).

RESULTS

Total Proceedings

The three-way contingency table for number of reference citations by year and journal is shown in Table 1. It can be seen from the Table that the trend over the past eight years has been growth in both the number of proceedings citations and total citations. [This trend of a greater number of citations per year cannot be attributed to just the number of articles, notes and comments or journal pages per year as these have remained relatively constant. For example, for the last four years the total number of articles, notes and comments has ranged between 142 and 146 with total journal pages ranging between 1281 and 1410.] A similar, though less pronounced, trend can be observed from Table 2 where self-citations have been excluded. Total proceedings citations and total proceedings citations minus self-citations are graphed by year in Figure A. Three findings from this Figure of particular interest are that 1) the trend line correlation between time in years and total citations of proceedings (TPROC) is significantly positive (.86), 2) proceedings citations maintain a fairly constant share of total citations with approximately 6%, and 3) excluding selfcitations has virtually no effect on the trend line relationship (as rTPROC, MTPROC = .99)- 1: should also be noted that the percentage of total proceedings citations which are self-citations is smaller than one might expect. on average approximately 18%.

TABLE 1

CITATIONS BY YEAR AND JOURNAL

TABLE 2

CITATIONS EXCLUDING SELF-CITATIONS BY YEAR AND JOURNAL

From the log-linear/logit analysis on the total reference citations [As the results of the log-linear/logit analysis for total reference citations minus self-citations (Table 2) were virtually the same, only the one set of results is discussed.] (Table 1), significant main "effects" were found for journal (XL = 94.6, p < .001) and year (XL = 26.8, p < .001). Examining the lambda coefficients for journal indicated that a given citation was significantly more likely in JCR, slightly though nonsignificantly more likely in JMR, and significantly less likely in JM to be a proceedings citation. Examining the lambda coefficients for year indicated basically the pattern seen in Figure A, with a given citation significantly more likely to be a proceedings citation in 1981, and significantly less likely to be one in 1977. This main effects model accounted for 77.4% of the total weighted variation to be explained. Nonetheless, this model has a XL value of 35.5 (p = .001), indicating significant variation due to the journal by year interaction. This interaction appears to be due to reversals between JCR and JMR over the years in terms of which journal has the larger proportion of proceedings citations.

FIGURE A

CITATIONS OF PROCEEDING BY YEAR

Primary Marketing-Related Proceedings

The number of citations for each primary marketing-related proceedings are grouped by year in Figure B. [Only the Figure for number of citations is shown and discussed as the lowest correlation between number of citations and number of citations minus self-citations for any marketing-related proceedings over time in years is .91 (p = .002).] Two findings that are of particular interest are that 1) over time ACR proceedings have received the most citations, and 2) the trend line correlation between time in years and number of proceedings citations is significantly positive only for ACR proceedings (r = .91, p = .002). The number of citations for AMA proceedings have, in general, increased although the sharp drop in 1982 lowered the trend line correlation (r = .66, p - .077), especially with just eight observations.

The number of citations for primary marketing-related proceedings by journal is presented in Table 3. It can be seen from the Table that ACR has the largest number of proceedings citations in JCR and JMR, whereas AMA has the largest number in JM. Overall, ACR has a significantly greater number of proceedings citations than does AMA (X(1) = 47.8, p < .001). Considering journals overall, a significantly larger number of marketing-related proceedings citations appear in JCR than in JMR (X2 = 9.8, p < .005), whereas a significantly fewer number of marketing-related proceedings citations appear in JM than in JCR (X2 = 179.7, p < .001) or JMR (X21 = 109.5, p < .001). The number of citations for primary marketing-related proceedings excluding self-citations is presented in Table 4 for comparison purposes. As can be seen from the Table, excluding self-citations does not alter the pattern of findings.

FIGURE B

CITATIONS OF PRIMARY MARKETING RELATED PROCEEDINGS BY YEAR

TABLE 3

PRIMARY MARKETING-RELATED PROCEEDINGS BY JOURNAL

TABLE 4

PRIMARY MARKETING-RELATED PROCEEDINGS CITATIONS EXCLUDING SELF-CITATIONS BY JOURNALS

DISCUSSION AND CONCLUSIONS

Citation analysis was used to address the question of what has been the relative impact of published proceedings papers upon the areas of consumer research and marketing. While there has been some controversy over the use of citation analysis and particularly the implications drawn from it (see Broad 1978), the present application--looking at how often proceedings are cited rather than individuals --is considered less controversial. That is because primary criticisms of citation counts or indexes, such as only first authors are listed and the negative effect of misspelling an author's name (or even leaving out one of the author's initials), are not relevant. Furthermore, in this research when self-citation counts were made, all authorship positions were considered.

Considering total proceedings citations, it was found that while the absolute number of proceedings citations has been significantly increasing over time, their proportion of total reference citations has remained fairly constant at about 6%. Proceedings citations were significantly more likely to appear in JCR, somewhat more likely to appear in JMR, and they were significantly less likely to appear in JM. Proceedings citations evinced an off-year in 1977 and a particularly strong year in 1981. Self-citations have no discernable effect on the relationships of proceedings citations with journals and time in years; a finding which holds for the primary marketing-related proceedings discussed next.

The ACR proceedings have had a greater impact upon the fields of marketing and particularly consumer research than any other primary marketing-related proceedings. By far, the ACR proceedings have received the largest number of proceedings citations. in addition, the number of ACR proceedings citations have shown a significantly positive trend over time. Considering particular journals, ACR proceedings have received the most proceedings citations in JCR, and JMR, while the AMA proceedings have received the most citations in JM. A related finding is that marketing-related proceedings have had the most impact upon JCR, followed by JMR, whereas these proceedings have had significantly less impact upon JM. Given that ACR proceedings are the most cited proceeding, and given the close relationship of ACR with JCR, this result is not surprising.

One possible explanation for the preeminence of the ACR proceedings relative to the AMA proceedings may be the longer paper length allowed in the ACR proceedings. By comparison, the limit for the 1983 ACR Conference is 20 pages, whereas the limit for the 1983 AMA Educator's Conference is 16 pages. (This pattern appears to hold for previous years, based upon the relative length of published papers in the two proceedings.) The additional length allows for greater depth of theoretical development and/or data analysis presentation. This can result in more substantive, and thus more citable, contributions.

REFERENCES

Broad, William J. (1978), "Librarian Turned Entrepreneur Makes Millions Off Mere Footnotes," Science, 202, 853-57.

Cox, Eli P., Hamelman, Paul W., and Wilcox, James B. (1976), "Relational Characteristics of the Business Literature: An Interpretive Procedure," Journal of Business, 9, 252-65.

DeSarbo, Wayne S., and Hildebrand, David K. (1980), "A Marketer's Guide to Log-Linear Models for Qualitative Data Analysis," Journal of Marketing, 44, 40-51.

Drew, David E., and Karpf, Ronald (1981), "Ranking Academic Departments: Empirical Findings and a Theoretical Perspective," Research in Higher Education, 14, 305-20.

Goodman, Leo A. (1971), "The Analysis of Multidimensional Contingency Tables: Stepwise Procedures and Direct Estimation Methods for Building Models for Multiple Classifications," Technometrics, 13, 33-61.

Green, Paul E., Carmone, Frank J., and Wachpress, David P. (1977), "On the Analysis of Qualitative Data in Marketing Research," Journal of Marketing Research, 14, 52-9.

Institute for Scientific Information, Inc. (ISI) (1982), Social Sciences Citation Index, 1982 Philadelphia: ISI, Inc.

Robinson, Larry M., and Adler, Roy D. (1978), "Citations Provide Objective Ratings of Schools, Scholars," Marketing News, 12, 1 and 8.

Siegel, Sidney (1956) Nonparametric Statistics For The Behavioral Sciences, New York: McGraw-Hill.

----------------------------------------