Attacking the Knowledge Gap Phenomenon

Donald P. Robin, Mississippi State University
Louis M. Capella, Mississippi State University
S. Roland Jones, Mississippi State University (student), Mississippi State University
Brenda S. Harmon,
ABSTRACT - Formulating an information strategy in which information is designed and distributed to bring about a greater increase in knowledge for all groups is an important task. The purpose of this study was to examine selected information strategy factors, determine the effect on the level of knowledge for the upper-lower socioeconomic segment, and judge the communicator's ability to close the knowledge gap.
[ to cite ]:
Donald P. Robin, Louis M. Capella, S. Roland Jones, and Brenda S. Harmon (1982) ,"Attacking the Knowledge Gap Phenomenon", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 239-243.

Advances in Consumer Research Volume 9, 1982      Pages 239-243


Donald P. Robin, Mississippi State University

Louis M. Capella, Mississippi State University

S. Roland Jones, Mississippi State University

Brenda S. Harmon (student), Mississippi State University


Formulating an information strategy in which information is designed and distributed to bring about a greater increase in knowledge for all groups is an important task. The purpose of this study was to examine selected information strategy factors, determine the effect on the level of knowledge for the upper-lower socioeconomic segment, and judge the communicator's ability to close the knowledge gap.


Formulating an effective strategy in which information is designed, packaged, and distributed to bring about greater increases in knowledge for all groups is an important task for many organizations. Theoretically, the end result of these organizations is the dissemination of public information so that people can make more intelligent decisions about everything from purchases to politics, and tax structures to legal aid. However, research and experience have shown that simply making more information available is not enough, even when that information would appear to be beneficial to reciPients.


A study conducted by Thorelli, Becker, and Engledow (1975) concerning subscribers of Consumer Reports in the United States provided a revealing profile of information seekers. Their data depicted the subscriber as the person with the money, communications skills, and motivation to use intensive, objective product information. The subscriber also exhibited a general preference for the print media. Since subscribing requires money and effort, and the print media requires communication skills, a firm could infer that selecting a different channel might make information more generally accessible. Evidence.gathered in a study by McEwen (1978) of toll-free telephone number users suggested that the solution is not so simple. In this study the profile of the information seeker was consistent with Thorelli, Becker, and Engledow in that the toll-free number user was found to be active, involved, and socioeconmically elite. It was found that the toll-free number user solicits information from a variety of sources, not primarily the print media. The implication is that eliminating the need for subscribership, thereby increasing accessibility, suggests a somewhat lesser print-oriented seeker/user but not a less active or less knowledgeable information use. In addition, increasing the accessibility of information does not broaden the user base.

These findings demonstrate that the size and nature of the gap between information seekers and non-seekers cannot be reduced by a strategy which focuses only on information availability or accessibility. Unless audience members are shown that information is relevant and meaningful, information will not be sought even if it is highly accessible. Successfully disseminating information requires a thorough understanding of what information is needed and how it can best be organized and delivered for acceptance and understanding by audiences.

With respect to both the absolute level of knowledge and changes in the level of knowledge which individuals exhibit regarding a particular subject, organizations can benefit from considering evidence gathered on the knowledge gap phenomenon. As originally introduced by Tichenor, Donohue, and Olien (1970), the knowledge gap hypothesis holds that:

"As the infusion of mass media information into a social system increases, segments of the population with higher socioeconomic status tend to acquire this information at a faster rate than the lower status segments, so that the gap knowledge between these segments tends to increase rather than decrease." (1)

The significant implication of this hypothesis for organizations trying to disseminate information is that some groups are more difficult to effectively reach than others. Thus, reaching the lower socioeconomic groups is a task that maybe more difficult for both commercial products and social ideas. The causes of knowledge gap phenomenon have been categorized into two groups.

(1) audience-related factors such as ability, motivation and media behavior which cause widening gaps; and

(2) message-related "ceiling effects" which cause narrowing gaps (Ettema and Kline, 1975).

An important hypothesis for use in trying to reach the lower socioeconomic groups and close the knowledge gap was formulated in an article by Cole and Bruner (1971) in the American Psychologist and called the "difference interpretation". The thrust of this theory is that people in various socioeconomic groups and/or from various cultures utilize their abilities differently (Ettema and Kline, 1977; Cole and Bruner, 1971). In this difference interpretation Cole and Bruner maintain, for example, that members of a higher socioeconomic class may exhibit their expertise in situations difference from members of a lower socioeconomic class. An implication of this explanation is that the knowledge gap phenomenon can be reduced by focusing on situation differences between higher and lower socioeconomic persons. The theory identifies situations in which (1) the individual is motivated to exercise their abilities; and (2) those in which it is functional for the individual to do so. Hence, it is predicted that segments of the population motivated to acquire the information and/or for which that information is functional tend to acquire the information at a faster rate. If the situation presented is attractive to the upper socioeconomic classes, then the knowledge gap tends to increase rather than decrease. If something intervenes to alter the motivation and/or function patterns of the lower socioeconomic groups it is reasonable to expect that the knowledge gap will continue to widen. Can the lower socioeconomic groups be effectively reached? If an information strategy could be developed to reach these groups, what components of the strategy would be more important? These were the questions pursued in this study.


The objective of this study was to examine selected approaches for disseminating information and determining the effect on the level of knowledge for a target lower socioeconomic group. A situation was selected in which the higher socioeconomic groups were expected to exhibit greater gains in the level of knowledge. Then, by analyzing the impact of the different information strategies on the target group, the researchers hoped to determine which if any of the strategies produced knowledge gains for the target socioeconomic group. If Cole and Bruner's difference interpretation hypothesis is essentially correct, different products or services and different media might provide different motivations for the lower socioeconomic classes. Of course, it must be functional for the class members to accumulate the information. For this reason, the study focused on a segment designated as the "upper-lower" socioeconomic class. Although the "lower-lower" socioeconomic class does save, it has the lowest savings rate and would not represent as good a market opportunity as the "upper-lower" socioeconomic class (Cunningham and Cunningham, 1981). For reasons that should become apparent, the "upper-lower" segment should have the motivation to accumulate information.

The "products" involved in this study were commercial bank services, two different types of savings accounts and a home improvement loan program. These services were selected because of their perceived social as well as commercial importance. It is important both on a moral and economic base, that lower socioeconomic segments have the financial information to make sound judgements.

The other components of the information strategy was media selection, and for this study the media involved was radio and newspaper. Two local radio stations (both AM and FM) were used for the study. Bank advertisements were scheduled during "drive times" each day. The one local newspaper was used for print advertising. All of the advertisements were one-eighth page in size and were located Run of Paper (ROP). Choosing the most appropriate media to carry the information to the target group, it is a potentially important factor. As noted previously, McEwen (1978) found that simply making information more accessible to all groups is not enough. Because different groups have different media preferences, organizations must understand and incorporate the media habits of the target group if they are to be reached. Thus, the use of specific media or media blends may better enable the organization to reach the upper-lower socioeconomic group.

Of particular importance for the upper-lower segment is that radio is an inexpensive oral medium and makes no assumption about the literacy of the listener. While the McEwen study required active participation on the part of the interested party, radio is a passive instrument. Obviously, the newspaper does require that the audience be able to read. Thus, it does not seem unreasonable for these media types to produce significant differences for the upper-lower socioeconomic segment.


Can communicators with control of the available media and other important environmental events produce a significant change in the level of knowledge in the upper-lower socioeconomic group? To determine the potential effectiveness of the information strategy, the impact on other socioeconomic groups, several objectives are achieved. First, the basic impact on the knowledge gap can be appraised. Second, the degree and nature of the influence on the lower socioeconomic groups can be estimated. Third, the importance and impact of socioeconomic segmentation can be determined.

Incorporating the three objectives with the evidence gathered from the knowledge gap literature, the following null hypothesis were developed for testing in this study:

H1: No difference in the level of knowledge occurs over a two week exposure period in the upper-lower socioeconomic segment when potentially important services are explained and different media are used.

H2: No difference in the level of knowledge occurs over a two week exposure period in the lower-lower, lower-middle, upper-middle, and upper socioeconomic segments when potentially important services are explained and different media are used


All four banks serving a small community were used in an experiment with advertising campaigns designed to inform the public of the three banking services. The campaigns were conducted over a six-week period using a Latin Square rotation of three banking services, combinations of available media, and three two-week time periods. Surveys were taken for each interval to measure changes in the level of knowledge about the three banking services. This process provided information which was used to determine the impact of different banking services and media selection. Recall post-testing techniques attempt to determine what individuals retain from advertisements. According to Runyon (1979) recall measures are particularly useful when the objective of an advertising campaign is to impart specific information. While the recall technique is appropriate in this study the ability to recall information under interview pressure may underestimate actual recall.

The community used in the testing the hypotheses was a town of approximately 12,000 in a county with population slightly over 31,000. The major source of employment is a state supported university with a student enrollment of 12,000 (not included in town and county population figures) although several small industries are also part of the community. The banking market for the area is clearly defined by the county lines. All fours banks which service this market cooperated in the study by controlling their advertising in a prescribed manner over a six-week period. Three types of banking services used and title identically by all four banks selected as potentially important to the receiver.

1) Silver Savings - a relatively new type of savings account which earns compounded daily interest at 5 1/4 % per annum;

2) Golden Savings - an older, more established type of savings account which earns interest quarterly at 5 1/2 % per annum and has withdrawals restricted to the first ten days of the calendar quarter; and

3) Home Improvement Loans.

Silver Savings accounts were relatively new thus had no advertising history. Golden Savings accounts had been established for the past ten years and had primarily been advertised through newspaper and some local radios but not to a great extent. Home improvement loans have been part of the banking services offered for some time but had received very little advertising attention in the past. The advertisements for each of the three services were "straight-forward" and focused on providing information. All of the copy simply stated the information the bank wished to convey with a minimum of body copy or pictures.

Residents surveyed in the study were selected using a stratified random sampling technique. Areas of the city were sectioned off into probable socioeconomic segments based on occupation and income of residents and property values in the area. The are city directory was then marked indicating these areas and a systematic random sampling process was used to select those residents to be surveyed from each socioeconomic group. Data were collected by means of telephone interviews conducted during the early evening hours at the end of each time period. Telephone interviews were used because the questions were short and straight-forward, and the sample size was reasonably large. In each household an adult member of the family was questioned.

The sampling procedure used systematic replacements for those people who would not respond or who were not at home. An overall 82.2 percent of those contacted (all groups over all four surveys) participated in the study. The not-at-home rate was somewhat higher at 23.4 percent of all calls placed by interviewers, but neither rate was felt to be damaging to the study. Replacement was handled by randomly selecting more names from each stratum than was necessary and using them in the order selected. A total sample of two hundred was used for each of the four interview periods and that total was divided equally for each of the five strata leaving a sample size of forty per stratum per interview period. Since there were four interview periods a total of eight hundred people were interviewed.

A 3 X 3 Latin Square design was used to test the hypothesis because it allowed the researchers an opportunity to study variations from an overall mean created by more than one variable. Every treatment was used once each category of each blocking variable as follows.


To determine the level of knowledge concerning each of the three banking services, four measurements were made. (Sample questions for the Golden Savings accounts appear in Appendix I.) Prior to the advertising programs, a pretest was conducted to gather information concerning prior knowledge about the three services. At the end of each two-week interval, another survey was conducted to again determine level of knowledge. In order to eliminate resampling previous respondents and thus avoid the learning which might occur from the questioning process after each person responded their name was "checked" to insure they would not be questioned again.

Analysis of variance (ANOVA) was used to test the significance of any variations from the overall mean. Analysis of variance was also completed on the mean scores for each time period, service type, and socioeconomic group. ANOVA is an appropriate technique because knowledge is the one metric dependent variable that represents the focus of this study, and because the predictor variable, banking services and media are nonmetric. F-ratios for each source of variation were calculated and each F-ratio was evaluated for significance to determine the confidence for each variation. The test produced a measure of variations in the average level of knowledge about those aspects of the banking services tested in the survey. Finally, multiple comparison tests were performed on the group means using the Tukey procedure.


In order to assess the changes in the level of knowledge, it is essential to evaluate the general level of knowledge which serves as the base for these changes. The mean for the entire Latin Square, or the mean level of knowledge for all time periods, all services, and all media combinations was 47.0 on a scale of 0-300 with O meaning no knowledge and 300 meaning complete knowledge. Additional insight on the general level of knowledge can be gained by observing the plot in Figure 1. It is important to note that the lower-lower socioeconomic segment has a general level of knowledge about the banking services of 15.0 significantly below that of the other groups. The progression of knowledge level from lowest to highest socioeconomic group was as expected except for the upper-middle socioeconomic segment. No further explanation is offered for the differences exhibited by that segment except that it included a large number of university faculty members. The overall evidence, however, does indicate that the upper socioeconomic groups had a relatively stronger base for acquiring additional knowledge, thus substantiating the knowledge gap hypothesis.

Knowledge Change for All Socioeconomic Groups

As indicated in Table 1, total knowledge changed from a knowledge score of 47.0 in time period zero to 65.7 in time period three for all socioeconomic groups. A rapid increase in knowledge was apparent in the first time period with a gain of 20.5 points. This was followed by a decline of 10.3 for time period two and succeeded by a subsequent increase in the final period to 65.7. The analysis of variance total F was 3.78 (p<.01). Thus, the total knowledge change for all five socioeconomic groups and all three services was significant over the total testing period. Using Tukey's range test (.05) time period zero was significantly different than time periods one and three.



The three services demonstrated various levels of knowledge growth during the testing period. The two savings accounts showed significant changes while the home improvement knowledge gain was not significant. Golden Savings knowledge increased gradually throughout all test periods beginning at 11.8 and ending at 22.4. The only significant difference was between time period zero-and time period three. Silver Savings knowledge increased dramatically in period one (to 18.6 from 10.1), however, leveled off to 14.4 at the final test phase. The only significant difference using Tukey (.05) was between period zero and period one. Home improvement knowledge increased only slightly over the entire test period from 25.2 to 28.9. No test periods were significantly different. It is interesting to note that level of knowledge increase and final knowledge levels varied by service with the greatest increases coming about in the savings accounts and the least change in home improvement loans. Initial knowledge level was highest for home improvement loans (25.2) while both savings had similar knowledge levels of around eleven points. It might be suggested that the initial levels of knowledge dictate a smaller increase in knowledge growth with some type of ceiling or "adequate knowledge" level operating to set a maximum on new knowledge gained.



Analyzing Reactions of Other Groups

Returning to the information in Table 1, total knowledge gains were undistinguished for all but two of the socioeconomic classes. The explanation for the lack of significant increase is undoubtedly different for the lower-lower socioeconomic group and the two upper groups. Cole and Bruner (1971) probably identified the reason for the lack of impact on the lower-lower segment when they suggested that the object must be functional or useful for the individual to exercise his information accumulation abilities. The lowest socioeconomic segment simply may not have enough funds to make the banking services highly meaningful.

The two upper socioeconomic groups were probably experiencing a "ceiling effect" where respondents were satisfied that they knew enough about the services to satisfy their needs. This explanation is probably especially appropriate for the highest socioeconomic segment because of the high levels of knowledge it exhibited even in the pretest. As suggested previously the upper-middle socioeconomic segments idiosyncratic behavior is something of a puzzle and can be interpreted as a lack of interest when compared to the two groups that bracket it. An analysis of the demographic composition of the group revealed little to help explain its behavior except that it contained a large proportion of university faculty families.

The lower-middle socioeconomic group exhibited a growth in knowledge that was significant at the .05 level. Much of the explanation for the knowledge growth in the class can probably be extrapolated from the analysis of the target upper-lower group that follows. However, the level of knowledge for the lower-middle segment was considerably higher than that of the upper-lower segment throughout the testing period. Because of this substantially lower beginning level of knowledge, acquisition of information was much more important for the upper-lower segment.

Reaching the Upper-Lower Socioeconomic Group

This group was the one of primary interest to the researchers. Increasing the upper-lower socioeconomic group's knowledge of banking services was the intent of the communication. To some extent the effort succeeded in accomplishing its objective. Significant changes in knowledge levels were found for all three banking services using all three media combinations.

Table 2 reflects the significant knowledge increases with knowledge being greater at the end of the testing period for all three services and two of the media. The exception is a significant decline in knowledge via radio. This data suggests the first hint of the possible imprecision of the knowledge gap hypothesis. Although knowledge levels achieved by the upper-lower socioeconomic group do not reach the levels of the upper groups, the change in knowledge is significant. Thus, a "gap" exists but not an increasing gap as proposed by the hypothesis, and further, it is a gap that can be partly closed. From this data, it appears that over a period of time lower socioeconomic groups that have an interest in a topic area can close the knowledge gap if they are presented information in a clear concise manner. The longer term change is illustrated through the significant multiple range tests. Three of the service knowledge scores, total, silver savings, and gold savings were significant between the pretest measure and the time period three measure. Home improvement knowledge significantly increased in time period one. Newspaper were effective in changing knowledge significantly from periods one to three and two to three. Combined media efforts were effective from one to two while ratio showed a significant decline in knowledge in periods two and three.




The results of this study partially confirm the existence of a knowledge gap phenomenon effect on the reception of knowledge. When comparing lower and upper socioeconomic groups over one time period and when comparing the lowest socioeconomic group with the other four groups over time the hypothesis has fairly high validity. However, this leaves the knowledge gap hypothesis incomplete. The phenomenon appears to be more complex than originally proposed. Certainly the type of information or knowledge involved has some effect on the increase in knowledge. More complex and more open-ended knowledge would probably be more susceptible to the knowledge gap phenomenon. When viewed over time, the potency of the phenomenon declines. Future studies must involve several measurement periods to avoid the possibility of a methodological artifact problem. Further, information seeing may lent itself more toward the knowledge gap phenomenon than information dissemination and "passive" receiving. Finally, the ceiling effects of information or knowledge gathering may be important barriers for all socioeconomic groups not just lower groups. The knowledge gap hypothesis appears to have application in explaining differences in knowledge gains among audiences, but the situation and type of in formation can greatly influence the validity of this hypothesis, and in at least certain instances, the gap can be partly closed.




Cole, ML and Bruner, J. S. (1971), "Cultural Differences and Inferences About the Psychological Process," American Psychologist, 26. 867-876.

Cunningham, W. H. and Cunningham, I. C. M. (1981), Marketing: A Managerial Approach, (Cincinnati, Ohio: Southwestern Publishing.

Ettema, James S. and Kline, F. Gerald (April, 1977), "Deficits, Differences, and Ceilings," Communication, 4, 179-202.

McEwen, William J. (March, 1978), "Bridging the Information Gap" Journal of Consumer Research, 4, 247-251.

Runyon, Kenneth E. (1979), Advertising and the Practice of Marketing, (Columbus, Ohio: Charles E. Merrill Publishing).

Schiffman, Leon G. and Kanuk, Leslie L. (1978), Consumer Behavior, (Englewood Cliffs, NJ: Prentice-Hall, kc.).

Thorelli, Hans B., Becker, Helmur and Engledow, Jack (1975), The Information Seekers, (Cambridge, MA: Ballinger Publishing Co.).

Tichenor, P. J., Donohue, G. A. and Olien, C. N. (Summer 1970), "Mass Media Flow and Differential Growth in R w ledge" Public Opinion Quarterly, 34, 158-170.