A Longitudinal Analysis of Consumer Behavior in the Elderly Population

Marilyn C. Fox, AT&T
A. Marvin Roscoe, AT&T
Alan M. Feigenbaum, AT&T
ABSTRACT - It is the objective of this paper to explore consumer behavior in the age 55 and over U.S. population, particularly as it relates to the telecommunications market. As reported by the U.S. Bureau of Census, the elderly population has grown substantially between 1970 and 1980, and has gained in popularity as an area for new research. In contrast to previous studies of the elderly consumer, which were methodologically oriented to snapshot comparisons with the younger consumer, this research examines elderly consumer behavior longitudinally as a total entity and within age subgroups.
[ to cite ]:
Marilyn C. Fox, A. Marvin Roscoe, and Alan M. Feigenbaum (1984) ,"A Longitudinal Analysis of Consumer Behavior in the Elderly Population", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 563-568.

Advances in Consumer Research Volume 11, 1984      Pages 563-568


Marilyn C. Fox, AT&T

A. Marvin Roscoe, AT&T

Alan M. Feigenbaum, AT&T


It is the objective of this paper to explore consumer behavior in the age 55 and over U.S. population, particularly as it relates to the telecommunications market. As reported by the U.S. Bureau of Census, the elderly population has grown substantially between 1970 and 1980, and has gained in popularity as an area for new research. In contrast to previous studies of the elderly consumer, which were methodologically oriented to snapshot comparisons with the younger consumer, this research examines elderly consumer behavior longitudinally as a total entity and within age subgroups.

The research discussed in this paper is based on a large sample drawn from a respondent panel representing all U.S. Standard Metropolitan Statistical Areas. The variables examined include an extensive battery of demographics, a series of attitude and value traits, and several measures of consumer behavior. To enable a longitudinal comparison, data was drawn from two time periods, 1976 and 1981. Based on social gerontological research, an age categorization of 55-64, 65-74, and 75+ was employed to trace the progression from mature adult to elderly.

The economic, cultural and social circumstances of the elderly are undergoing rapid change. Amid these trends, the elderly are becoming a larger proportion of our population, with continued growth projected. The results of this study of elderly consumer behavior have several implications for this increasingly important and prominent segment of our consumer marketplace.


Among readers of this paper, how many will be employed at the age of 75? Will any run in marathons when 85? Will some be sexually active at 95? Finally, how many will be alive at the start of the 22nd century? The answers to these questions are open to conjecture but it is not unlikely that each would yield some affirmative response.

Unquestionably, the elderly are experiencing improved health, and despite crises in Social Security and Medicare funding, many are now in a better financial position to enjoy it. Furthermore, the attitudes of the elderly population are changing and its presence is increasing, both numerically and politically. These trends are hard to ignore, especially in the marketing community, where a changing perception of the elderly is developing in light of their increasing buying power, and decreasing reluctance to use it in the consumer marketplace.

AT&T, a leading provider of telecommunications equipment, and voice and data transmission services, is one company with a particularly significant stake in understanding and serving the elderly market. Among improvements in technology which have affected all of our lives, new and better modes of communications are paramount in importance to the elderly (Carmone, Krauser, and Baker 1983). Thus AT&T, which has successfully applied segmentation (see Assael and Roscoe 1978) principles in its marketing programs over the last 15 years (see Veltri and Schiffman 1983) is now extending these efforts to a separate focus on the older segments of the population. The ongoing nature of this segmentation research has enabled us to take a unique approach by performing a longitudinal study of the elderly segment's consumer behavior. The remainder of this paper is devoted to describing this study.


The MARC Panel

The basis for this study, as well as a large number of other AT&T marketing studies, is the Market Analysis of Revenue and Customers (MARC) data source. The MARC Panel, as it is known, consists of detailed demographic, psychographic, and usage data on thousands of AT&T customers. The panel is constituted through a random sampling of all U.S. Bell System customers, representing each SMSA. For a number of years, on a fairly frequent and periodic basis, a sample has been selected and questionnaires mailed out to all customers in the sample. Data records of the customers who return a questionnaire are constructed from merging the questionnaire responses with pertinent information from the billing records. The current sampling process originally yields 75,000 customers and the net response (after weeding out "unreachables," unusables," and non-respondents) is more than 50%. Various means have been employed to maximize the response rate while insuring the highest quality response, sometimes resulting in a net response nearing 75% (Roscoe, Lang, and Sheth 1978).

The questionnaire contains a number of 5-point Likert scale psychographic opinion and attitudinal items, as well as questions on the customer's household which comprise the demographic information. The billing information profiles the customer's usage of long distance service and consists of number of long distance calls, minutes of long distance connection, and long distance charges. Combining this "hard" usage information with the "soft" questionnaire data sets the MARC Panel apart from most other research approaches, which rely totally on consumer-perception of past behavior or anticipation of future behavior.

Panel Demographics

The demographic information captured from the questionnaire can be categorized as family/housing, mobility, and socio-economic. Family/housing includes own or rent, number of rooms, family size, and head-of-household (HOH) gender. The mobility category, which plays a prominent role in telecommunications, includes time in current residence and number of moves in the last five years. The socioeconomic information contains the standard items of income, education, and employment. This choice of demographic variables is based on analyzing respondents' questionnaire behavior over a number of years, and selecting categories and items which have proven accuracy in terms of customer response and high prediction capability for consumer behavior.

Panel Psychographics

Although the particular items have changed to keep pace with new technology in the telecommunications marketplace, the MARC Panel has always incorporated a large number of psychographic items. The psychographics include general opinions reflecting a respondents lifestyle, as well as attitudes toward specific consumer products and services. A representative example is "On my way home for the day, I plan the telephone calls I'll make that evening." The respondent indicates degree of agreement or disagreement on a 5-point Likert scale. For the purpose of this study, common psychographic items from the 1976 and 1981 MARC Panels were selected. By means of factor analysis, these items were grouped under four factor headings (keeping-in-touch, gadget-oriented, thriftiness, and fashion-innovator), which will be discussed individually below in the section on results

Study Structure

As a final preliminary to presenting the actual results, it is important to elaborate on how the study was structured. The key issue was deciding on who should be classified as elderly. The delineation scheme used is taken from the paper by Sherman and Schiffman (1983) which bases the choice on a review of social gerontological research. There are three sub-groups of elderly. The 55-64 age bracket is referred to as mature adult, 65-74 as young elderly, and 75-plus as old elderly. Tabulations on each variable were done for each of these groups and the sample as a whole. A further breakdown divided the categories by retired and employed, an important distinction which is discussed further in the section on results. Finally, for each of these subcategories, the results tabulated were for both the 1976 and 1981 Panels. This approach provided a good longitudinal comparison over a period long enough to expect changes, but not so long that comparisons would be considered meaningless due to the pace of technological advancement. An entryless chart representing the classification scheme just described, is diagramed below.




Sample Demographics

The results described in this section are derived from a comparison of all elderly (age 55+) subjects from the 1976 and 1981 MARC Panel respondents, as profiled in the following table.



The elderly respondent split between elderly employed and retired is exhibited because definite differences between the groups were evident in the results. It is interesting to note that while the proportion of elderly to the total population of respondents increased only modestly (less than 5%) from 1976 to 1981, the breakdown of employed versus retired elderly respondents shifted dramatically and is now essentially consistent with the known UT.S. population split between retired and employed in the 55 and over age bracket. The increased response by retired elderly may be indicative of a hypothesized trend for that group to become more involved and increasingly seek to be heard.

The longitudinal comparison on demographic data revealed noticeable differences in essentially every category, both in the elderly population, and in the population as a whole. In 198 1, a larger percentage of subjects from both groups owned property and, on the average, lived in larger residences than in 1976. Also, a decrease in mobility was evidenced during that period for the elderly and the whole population. Education and income were other areas in which both groups showed improvement but it was more marked for the elderly as shown in the tables below.





It should be noted that the longitudinal educational improvement shown by the elderly is partially attributable to a demographic process known as the cohort impact. This term refers to the shifts in age categories over time so that some of the younger, more educated population has moved into the elderly category. Inflation, another time-related effect,-likewise accounts for some of the income improvement in the elderly segment

The two groups showed more differences in degree of change for household makeup as the tables below illustrate.





The trend in head of household gender is especially interesting and a further breakdown reveals that it is most significant in the 65 and over population. The bulk of the longitudinal shift is due to men who became 65 between 1976 and 1981 X and remained employed rather than retiring. Because men represent a substantial majority of the older working population, this data supports other figures which indicate a trend toward retirement-age people remaining employed. It is reasonable to expect that within several years, the employed-retired distribution shift will intensify, and will be accompanied by major changes in marketing approaches to the elderly household.

The analysis above revealed differences in demographics over the five year time period. The magnitude of the differences was smallest when longitudinally comparing the youngest elderly segment, and progressively larger for the older segments. However, for all age segments, including the whole sample population, the direction of change for any given demographic variable was the same. The overall similarity in longitudinal demographic changes for the elderly, as compared to the total population,suggests that the longitudinal psychographic differences to be discussed below are due to actual trends, rather than demographically dissimilar study PoPulations.

Psychographic Factor Analysis

As mentioned earlier, a factor analysis of the psychographic items resulted in four factor groups. The first, fashion innovator, contained items on interest in new styles of telephones and other new consumer products, as well as inclination and likelihood of moving in the near term.

Gadget orientation, the second factor, focused on attitudes about the various "premium" styles of telephones with features such as pushbuttons, dial-lights, and special styling.

A question on bargain shopping behavior and a self-assessment item on the use of money comprised the third factor, thriftiness.

The final factor, keeping-in-touch, focused on actual telephone conversational patterns and other closely related items. The keeping-in-touch factor is of special interest to the authors of this study because it has the most relevance to the marketing of long-distance services, the mission of their specific AT&T organization.

The Fashion-Innovator Factor

In the "fashion innovator" factor category, the general item "I enjoy trying new products when they first come out" showed significantly more agreement by each segment of the elderly in 1981 than in 1976. The elderly response to this item is shown in the Appendix, Figure A. The specific companion item "If a new style of telephone were introduced, I'd be more likely to get it than my friends" met an across-the-board low level of elderly enthusiasm in 1976, while in 1981 the 75+ group agreement doubled, with major increases for other elderly groups. In contrast, essentially no longitudinal change was registered for the sample population as a whole. The mobility items for the fashion innovator showed no change for any groups.

The Gadget-Oriented Factor

A very interesting result occurred for the "gadget-oriented" factor. There were no changes in the decorative and fancy phone items. However, push-button phones, which have no functional value, were favored more in 1981. "A pushbutton phone is for people with more money than they know what to do with" was agreed to by less than a third of the elderly, about the same low level as in 1976. However, t-r.e item "I prefer push-button phones even though more expensive" showed a significant increase in agreement for each elderly group as pictured in the Appendix, Figure B. Thus, the elderly agreed that push-button phones are expensive but did not see them as frivolous. Furthermore, "I find push-button telephones unattractive" was agreed to more often (see Appendix, Figure C) by the 1981 elderly, demonstrating that it is not aesthetics, but must be functionality, that is the motivator for the elderly's positive attitude. Adding decisive credence to the functionality argument was the moderate across-the-board decrease in agreement on the item "A dial-light on a telephone is an example of an unnecessary luxury." Overall, this analysis of the "gadget-oriented" factor suggests that, to the elderly, the steak sells rather than the sizzle.

The Thriftiness Factor

The "thriftiness" factor strongly supported a widespread societal belief that the elderly are thrifty or at least perceive themselves that way and did in 1976 as well. The perception aspect of that statement may be key because a full three-quarters of the total respondent population perceived themselves that way, despite the increasingly high levels of individual bankruptcy that exist in the U.S. today. Interestingly though, a lower percentage of all elderly groups and the total sample population agreed with the item, "I shop for low prices," than perceived themselves to be thrifty. This was the case in both 1976 and 1981, with essentially no change in the shopping item. Clearly, the two items are not synonymous, which might lead one to infer that a somewhat expensive but high-quality functional product is attractive to most consumers. including elderly.

The Keeping-In-Touch Factor

The fourth and final factor, "keeping-in-touch," is epitomized by the now familiar suggestion "Reach out, reach out and touch someone," the theme of a well-known AT&T long-distance advertising campaign. Thus, it is not surprising that one of the items under that factor is "I probably make more long-distance calls than most people I know." More agreement was evidenced for this item in 1981 across the board, with the elderly at a somewhat lower agreement level than that registered for the total sample population. The longitudinal shift for the elderly is featured graphically in Figure D (see Appendix). The functionality issue raised earlier was further demonstrated by the two other "keeping-in-touch" items for which marked shifts occurred. Graphs for "I prefer to have several phones in my home as a convenience" and "My home is small so I don't need more than one phone" appear in the Appendix, Figures E and F. The responses for both pointedly indicate that the elderly, as does the full population of respondents, fully acknowledge the need for enough (usually meaning multiple) telephones in their homes. In 1976, economies were sometimes achieved by minimizing telephones. In 19817 when "keeping-in-touch" was more important, the functional advantage of having a truly adequate number of telephones outweighed the cost. The positive results for the previously mentioned "keeping-in-touch" item, perceived increase in long-distance usage, are substantiated by the actual usage data which is discussed below.

Usage Results

There are two points of considerable interest with respect to the actual usage results. First, for each of the three measures employed, significant changes occurred in the elderly categories as well as in the total sample population. Second, even with these usage increases, the overall current usage level is low enough so that the potential for improvement is almost staggering. The results are recorded in the table below.



Paralleling the psychographic results, usage was highest in the youngest elderly group and lowest in the oldest elderly group. The employed had greater usage in each elderly group than did the retired, which is also consistent with consumer orientation patterns which emerged in the psychographic analysis. Finally, usage has clearly increased longitudinally, just as more positive buying behavior was evidenced in the 1981 results than in 1976.


Product Implications

This section examines some of the marketing implications of this study, beginning with the most immediate concern, the long-distance business.

The growth in long-distance volume that has been experienced recently follows specific marketing efforts targeting the upscale consumer. The results of this study on the "keeping-in-touch" factor suggests that specific targeting of the elderly is warranted. Given the low level of elderly long-distance usage and the improved financial circumstances for many in that segment, a campaign which clearly communicated "keeping-in-touch" needs met by long-distance is recommended to reach the elderly.

Beyond long-distance, the study also revealed valuable information for telecommunications equipment manufacturers and the consumer market as well. The responses for the "fashion-innovator" and "gadget-oriented" factors indicate that the elderly will express interest in telephone equipment with value and functionality. The results also reveal a strong generic interest among the elderly in new products. It is hard to say what actual market potential underlies these trends, but few current marketing campaigns address the elderly market and this must change if we are to discover the answer

Segmentation Implications

Perhaps the most valuable of the study results lies in the area of implications for market segmentation. The elderly are not one, but actually several different markets, with age and employment as the "subsegmentation" factors. As might be expected, the 55-64 employed group was the most inclined to consume within the elderly population. In fact, the 55-64 employed group has a close demographic resemblance to the total sample population (all ages and employment conditions), with significant deviations limited to the mobility and income categories. Furthermore, the psychographic response profiles match very closely as well. It would seem, therefore, that the 55-64 employed category should be distinct.

In each of the elderly sub-categories, the split between employed and retired was noteworthy. Aggregating 55 and older employed and comparing to retired reveals significant differences in essentially every category of demographic, psychographic, and usage results. However, the trends for both groups were similar. Therefore, it appears that the use of age and employment status in elderly market segmentation is generally advisable,but some marketing approaches for the elderly populations as a whole are feasible.

Based on the above discussion, the following segmentation scheme should be included among serious alternatives: 55-64,employed; 55-64, retired; 65 ;employed; 65+, retired.

To wrap up the Conclusions section, we will comment on events from 1976-1981 which have ramifications on the interpretation of this study.


In 1977, the "registration" ruling was issued by the Federal Communications Commission (FCC). The ultimate outcome of that was the emergence of an across-the-board, truly competitive telecommunications equipment market. The impact on this study was greatly increased awareness of the phone as a consumer good rather than a device supplied by a local utility firm. Specialty telephone stores proliferated and telephones were subsequently offered in a wide variety of department store retail outlets. Since the average consumer had little inkling of the coming events, in 1976, it seems safe to say that longitudinal differences in responses on the 1981 psychographics were accentuated by the considerably more telecommunications-sophisticated population. Further complicating the picture was the Retirement Act of 1978 which essentially eliminated mandatory retirement at age 65. Undoubtedly, this had some effect on the employed-retired split for the 65 and over group, although its relative impact through 1981 was probably a good deal less pronounced than that of the FCC ruling. While new regulatory decisions and legislation affecting the consumer market have become almost daily occurrences, the "registration" ruling and Retirement Act of 1978 were major events. They are extremely relevant to this study and their impact should be considered in an overall analysis.

Further Research

Despite limitations and caveats mentioned throughout the course of this paper, this study brings out the increasing importance of the elderly as a market segment and some of their distinctive aspects. Further, it has brought to our attention several areas we need to learn more about through further studies. This additional research would be of two major types. First, we would like to replicate this study in several modified forms, using various refinements that would enable firmer conclusions to be reached. Second, a need exists for other special studies outside of this framework which would focus on particular related issues. The remainder of this section is devoted to a separate examination of these areas.

Study Replication

The term "elderly" conjures up a certain image in the mind but based on this study's results, a definition of this population is not clearcut. Using several scenarios for age segmentation, statistical tests should be applied to the demographic data to determine the best differential age cutoff values. Further research should also draw on select groups of demographics as controlling variables for longitudinal comparison, although the demographic analysis of the 1976 and 1981 respondents in this study supported comparison of them without control.

Future studies should also look more closely at gender differences, which among other things, would involve looking beyond just head-of-household (e.g., marital status). The Sherman and Schiffman paper mentioned earlier contains a discussion of two hypotheses with important implications for gender differences in elderly behavior. Briefly stated, the "leveling hypothesis" asserts that equalizing circumstances occur as the genders advance in age, while the "jeopardy hypothesis" says that becoming old exacerbates already existing societal inequities for women. If either of these hypotheses are true, or if they both apply situationally, then gender differences among the elderly have a different pattern than for the younger population. In that event, particularly considering the changing role of women in society, it is especially important to feature gender in a Prominent role in future studies.

Related Research

Having performed this study, we are now armed with the conviction that the elderly segment is an improving market. However, even with the improvements in study methodology mentioned above, we would only be clarifying the structure of the market while discovering little more of its needs and dynamics. As an example of knowledge we currently lack, we would want to understand how brand preference is manifested in elderly consumer behavior. This is of particular importance in technology-based markets, where the elderly are not as informed as younger people who more frequently use state-of-the-art equipment in their work. As a consequence, the elderly use more limited criteria in making purchase decisions of this type. An even more basic issue is to determine how the elderly are reached through marketing and advertising media. Finally, there are many unanswered questions on pricing effects, which have been raised elsewhere by Schewe (1983).


Beyond acquiring expertise and improving our techniques in developing marketing approaches to the elderly, we must offer functional products and services of high value to them With respect to AT&T's lines of business, some possibilities include emergency, security, and energy management hookups Shopping, banking, convenience services, and consumer hotlines might also be of interest. The growing trend in information access has resulted in proliferating use of a telephone-personal computer configuration for database retrieval. Therefore, services which provide capsulized voice information on health matters, financial alternatives senior citizen activities, etc. might serve as an elderly person's alternative to computerized databases.

The elderly have been a neglected and sometimes invisible market, as noted by Bartos (1980). The overriding finding of our work is that an untapped potential exists. An approach to better understanding this market has been presented and product/service concepts related to the authors' industry have been suggested. It remains to follow up on this and other work so that the business community can better serve the elderly market for their mutual benefit.















Assael, Henry and A. Marvin Roscoe, Jr. (1976), "Approaches to Market Segmentation Analysis," Journal of Marketing, 40, 67-75.

Bartos, Rena (1980), "Over 49: The Invisible Consumer Market," Harvard Business Review, 58, 14-18.

Carmone, Frank, Cheryl Krauser, and George Baker (1983), "The Mature Consumer and Technology,' presented at the Association for Consumer Research Conference Special Session, Chicago, Illinois.

Roscoe, A. Marvin, Jr., Dorothy Lang, and Jagdish Is.Sheth (1975), "Follow-up Method, Questionnaire Length, and Market Differences in Mail Surveys," Journal of Marketing, 39, 20-27.

Schewe, Charles (1983) "Buying and Consumer Behavior of the Elderly: Findings from Behavioral Research," presented at the Association for Consumer Research Conference. Special Session, Chicago, Illinois.

Sherman, Elaine and Leon Schiffman (1983), "Applying Age/ Gender Theory from Social Gerontology to Understand the Consumer Well-Being of the Elderly,' presented at the Association for Consumer Research Conference, Special Session, Chicago, Illinois.

Veltri, John and Leon Schiffman 1983). 'Fifteen Years of Consumer Lifestyle and Value Research at AT&T,' presented at the Personal Values and Consumer Behavior Workshop, Oxford, Mississippi- University of Mississippi.