There is a lack of research on the everyday lives of older people in developing countries. This exploratory study used structured observation and content analysis to examine the presence of older people in public fora and considered the methods’ potential for understanding older people’s social integration and inclusion. Structured observation occurred of public social spaces in six cities each located in a different developing country and in one city in the United Kingdom, together with content analysis of the presence of people in newspaper pictures and on television in the selected countries. Results indicated that across all fieldwork sites and data sources, there was a low presence of older people, with women considerably less present than men in developing countries. There was variation across fieldwork sites in older people’s presence by place and time of day and in their accompanied status. The presence of older people in images drawn from newspapers was associated with the news/non-news nature of the source. The utility of the study’s methodological approach is considered, as is the degree to which the presence of older people in public fora might relate to social integration and inclusion in different cultural contexts.
By 2050, there will be approximately 2 billion people aged 60 years or more and the number of older people will outnumber the young. The most rapid changes will take place in developing countries, and projections indicate that by then four of five older people will live in developing countries [
When proposing a need for more social research on older people in developing countries, the methodological approach to be adopted requires particularly careful consideration. Applying conventional research methods might be problematic. Using survey methods and instruments created for use in developed countries raises issues of validity, if concepts and language are simply imported without thought for differing cultural contexts. Difficulties in accessing older people and the use of standardised but culturally biased procedures and formats for delivering questions and eliciting responses threaten reliability and representativeness (cf. [
Observational methods and content analysis might circumvent some of the problems that arise in using conventional research methods in the context of developing countries. It can be argued that the actual presence or absence of older people in public fora, such as social and cultural spaces, is a good marker of their level of social integration (cf. [
Studies that document the presence or absence of older people in public fora might therefore provide valuable data that will contribute to our understanding of the place of older people in their respective societies. However, there are a very limited number of such studies, particularly in developing countries. There is some information, primarily from developed countries, about places that are important to older people. For example, Londoners identified a religious building as an important place in their lives [
With regard to the presence and form of presentation of older people in broadcast and print media there is a considerable body of work available, but there is a knowledge gap around images of older people in non-Western cultures with cross-cultural research extremely limited [
Considering developing or non-Western countries, a study on the presentation of older people on Taiwanese television [
When studying older people’s presence in public fora, it is important to consider the differing life situations of older women and men in developing and developed countries. First, there are health differences between women and men, with greater healthy life expectancy among women in both developing and developed countries [
Furthermore, there are gender differences with regard to media representation. As indicated above, research has shown that older women are particularly underrepresented across a range of media. Feminist critique of media representations suggests that underrepresentation may play a significant role in how women interpret and experience ageing [
Our primary research question is: how present are older people in public fora in developing countries? A secondary research question is: is the presence of older people in public fora related to their gender? Our site selection, of six cities in six developing countries and one city in a developed country, and the use of two data collection methods were motivated by several considerations. First, as this was an exploratory study it was important to examine the validity, reliability, and feasibility of the methods adopted across a wide range of settings. The selection of a range of developing countries fulfilled this consideration, as such countries would provide challenges and contexts as fieldwork sites that would contrast with those found in the developed country; there would also likely be differences between the selected developing countries themselves. Second, there is a tremendous gap between our understanding of the demographic changes affecting older people in the developing world and our understanding of the public lives of older people in such countries. Our study, with its primary focus on developing countries, would thus both increase the amount of data available on the presence of older women and men in social spaces and in the media in such countries and offer some insight into how their presence in public fora in developing countries contrasts with their presence in similar fora in a developed country, with opportunity for cautious comparison due to the use of standardised methods across fieldwork sites. Finally, the use of three different data sources—non-participant structured observation in social spaces and content analysis of images of older people both in newspapers and on television—would allow for triangulation of data as a way of evaluating the robustness of findings.
This study involved gathering data in six developing countries—Syria, Jordan, Egypt, Thailand, Kenya, and Tanzania—and in one Western country—the UK—over the course of 2007. In each of these countries, structured observations were carried out in public social places where people were likely to be present and content analysis was performed of images in newspapers and television programmes in which they were likely to be depicted.
Country selection was guided by a need for variation in geographical location, political structures, and ethnic and social mixes, as well as demography. The chosen developing countries offer diversity of cultures, with varying degrees of contrast and comparison with the selected site in the UK. The selected countries represent, respectively, the overwhelmingly Arabic Muslim Syria and Jordan and Hamitic Muslim Egypt; the multiplicity of ethnic groups in Black East Africa, with a predominantly Christian culture and minority Muslim culture in Kenya, and strong Christian and Muslim cultures in Tanzania; and the distinctive, largely Buddhist Thai culture of the Indian subcontinent, with its Chinese and Malay subpopulations. The UK has predominantly a white population with strong Christian heritage but has increasingly diverse ethnic and religious minorities [
The procedures used for structured observation and content analysis are described in detail below. For all data sources, an older person was defined as a person whose appearance suggested an age of 60 or more. For all procedures, weekends in non-Muslim countries and Fridays in Muslim countries were not used for data collection in order to avoid “atypical” days producing biased data.
A site selection frame was developed for the structured observations of older people. As detailed in Table
Observations of older people in public places and in media.
Data source | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
Public spaces | Newspapers | Television | ||||||||
Observations | Read | Pictures examined | Channels watched | Observation periods | Total time | Scenes examined | ||||
Periods | Time | |||||||||
|
% |
|
|
% |
|
|
% |
|
% | |
City/country | ||||||||||
Haleb/Syria | 42 | 15.9 | 3 | 114 | 7.96 | 8 | 6 | 22.2 | 165 | 27.6 |
Amman/Jordan | 42 | 15.9 | 3 | 136 | 9.49 | 5 | 5 | 18.5 | 77 | 12.9 |
Luxor/Egypt | 42 | 15.9 | 3 | 236 | 16.5 | 4 | 5 | 18.5 | 50 | 8.35 |
Bangkok/Thailand | 36 | 13.6 | 3 | 450 | 31.4 | 0 | 0 | 0 | 0 | 0 |
Nairobi/Kenya | 30 | 11.4 | 2 | 170 | 11.9 | 0 | 0 | 0 | 0 | 0 |
Arusha/Tanzania | 30 | 11.4 | 2 | 93 | 6.49 | 6 | 5 | 18.5 | 111 | 18.5 |
Sheffield/UK | 42 | 15.9 | 2 | 234 | 16.3 | 4 | 6 | 22.2 | 196 | 32.7 |
Total |
|
|
|
|
|
|
|
|
|
|
A time sampling frame for recording of data from observations was developed based on previous work on the observation of older people in residential care environments [
Due to local circumstances, some adjustments were made to the site selection frame. Where the city center was non-residential, as in Amman and Bangkok, the old center and the area around the railway station were chosen, respectively, as they are focal points of city life. In Nairobi, the city center was not residential but it was not safe to observe outside this area. In Luxor and Arusha, a main bus stop was selected as the bus station was out of town; in Bangkok the train station was used, as the bus station was in a different area; in Bangkok, Nairobi, and Arusha there were no parks within the area observed so this part of the observation schedule could not be carried out. Some adjustments to the sampling frame were also required to reflect local circumstances. In Nairobi and Arusha, tourists are advised not to go out on foot after dark, so observations could not be done after 8 pm. If a building was not open during a given phase of the day, data could not be collected. Religious buildings, cafes, and post offices were affected by closure. Public spaces selected for observation were in total closed on 11 occasions between 7 and 9 am, on 10 occasions between 8 and 11 pm, on 8 occasions between 5 and 8 pm, and on 1 occasion between 9 and 12 am.
Media content analysis was undertaken of newspaper pictures and television channels, as detailed in Table
A sampling frame was developed in which for each country two local language papers and one English language paper were chosen, unless English was the local language (UK, Kenya, and Tanzania) where two English papers were chosen only. Within this frame, as the researchers had restricted prior knowledge of the countries’ media, the newspapers were arbitrarily selected from amongst those papers prominently displayed on at least two news stalls on a given day in each city. Eighteen papers in total were selected. All were dailies and all were published within the relevant country. Circulation and ownership varied.
Within each newspaper, every picture was examined and data extracted on people represented in terms of gender, age group, and number. Where older people were depicted, a brief description of their activity was recorded. The page number and section were noted, together with the headline for English papers and a brief description for other language papers, except when the picture contained only a face with no indication of context. Pictures advertising a product were recorded as adverts and the products were identified where possible. The sections in which the picture appeared were used as base categories for the analysis and finally combined into two superordinate categories: news and non-news.
Within the sampling frame developed for television, terrestrial television channels were prioritized, followed by channels from the region in which the country was situated, and if this was insufficient, a sample of the remaining channels was chosen. Television channels were recorded in all countries but Thailand and Kenya, as in these two countries the researcher was not able to access them within a context that allowed for reliable data recording. In Syria, eight channels were selected, all Arabic. In Jordan, five channels were selected, three of which were Arabic and the other two are Canadian and Turkish. In Egypt, four channels were selected, two Arabic and one American and there was one other channel of unknown origin. In Tanzania, six channels were selected: these were Tanzanian, French African, South African, Indian, British, and American. In the UK, four of the five British terrestrial channels were selected. In total, 27 channels were selected.
The time sampling frame for recording of data from television was developed via adaptation of the frame used for the observation of social spaces described above and via extensive piloting. Each selected channel in a country was studied for a five-minute period during each of the same six phases of the day used for social space observations. A five-minute period was judged, a useful time frame for coding television material as it provides sufficient time to avoid the whole period being taken up by advertisements or credits and to establish reliably the type of programme being broadcast.
During these five-minute periods, every scene was recorded and data extracted on people represented in terms of gender, age group, and number. By “scene,” what is meant is a linear representation or presentation of events or narrative, commencing and ending on an edit or camera switch. Where older people were depicted, a brief description of their activity was recorded. The programme type and where possible the name of the programme were recorded. For programmes in English, the programme type was generally evident through the language. For other channels, the type was decided based on the action and appearance of the scenes. Analytic categories were decided by identifying common programme types in UK and finally combined into two superordinate categories: news or non-news.
As with observations of social spaces, some adjustments to the sampling frame had to be made. It was not possible to observe television from 12 to 2 pm in Jordan, 8 to 11 pm in Egypt, and 7 to 9 am in Tanzania, due to pragmatic constraints on the researcher’s access to media during these phases.
As the data collected was largely non-normal in distribution, nonparametric analytic techniques were used. Parametric analyses were used only where they provided valuable additional descriptive information. Where repeated testing of the same sample occurred, Bonferroni corrections to reduce Type I errors were applied, with an appropriate reduction in alpha for each analysis. Where category cell
In total, 310 older people were observed in social spaces. This was a mean of 1.17 (
The number of older people observed per period varied considerably across cities (see Table
Observations of older people by gender in seven cities.
City | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Haleb | Amman | Luxor | Bangkok | Nairobi | Arusha | Sheffield | Total | |||||||||
|
% |
|
% |
|
% |
|
% |
|
% |
|
% |
|
% |
|
% | |
Men | 24 | 85.7 | 44 | 97.8 | 9 | 100 | 17 | 89.5 | 2 | 100 | 0 | 0 | 88 | 42.5 | 184 | 59.4 |
Women | 4 | 14.3 | 1 | 2.22 | 0 | 0 | 2 | 10.5 | 0 | 0 | 0 | 0 | 119 | 57.5 | 126 | 40.6 |
Total |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Observation period | ||||||||||||||||
Mean | 0.64 | 1.07 | 0.21 | 0.53 | 0.07 | 0 | 4.93 | 1.17 | ||||||||
(SD, range) | (1.45, 0–7) | (1.56, 0–7) | (0.47, 0–2) | (1.06, 0–4) | (0.25, 0-1) | (0.00, 0-0) | (9.32, 0–39) | (4.16, 0–39) | ||||||||
Median | 0 | 0.50 | 0 | 0 | 0 | 0 | 0 | 0 | ||||||||
(IQR) | (0–0.25) | (0–2.00) | (0-0) | (0–0.75) | (0-0) | (0-0) | (0–5.25) | (0-1.00) |
There was significant variation across cities in the number of older men observed per period (
With regard to older people observed per phase of day, the largest number of older people was observed between noon and 2 pm (
There was also considerable variation across cities in the proportion of older people observed who were accompanied as opposed to observed ones alone. While nearly half of the older people observed in Luxor, Nairobi, and Sheffield were accompanied, only 15.8 per cent of older people observed in Bangkok were accompanied, 14.3 per cent in Haleb, and 8.89 per cent in Amman. There was a significant association between the accompanied versus alone status of older men and city (
Of those observed older people who were accompanied, 84 (79.2%) were accompanied by older people only, while 20 (18.9%) were accompanied by people of other age groups only and two (1.89%) accompanied by mixed age groups of younger and older people. There was a significant association between city status (developed versus developing) and whether or not the accompanying person was old (Sheffield 94.6%, developing cities 40.0%,
A total of 1433 pictures were extracted from newspapers. With regard to the images of people whose ages could be reliably determined, 217 older people were observed in total, in comparison to 513 children or babies, 988 middle-aged people, and 2109 young people. There was thus significant variation in the absolute level of presence of the different age groups (Friedman
The proportion of pictures containing older people varied significantly across countries (see Table
Images of people by age and gender in pictures in newspapers from seven cities.
City | ||||||||
---|---|---|---|---|---|---|---|---|
Haleb | Amman | Luxor | Bangkok | Nairobi | Arusha | Sheffield | Total | |
|
|
|
|
|
|
|
|
|
Pictures examined | 114 (7.96) | 136 (9.49) | 236 (16.5) | 450 (31.4) | 170 (11.9) | 93 (6.49) | 234 (16.3) | 1433 (100) |
Picture content | ||||||||
Older men | 9 (7.89) | 3 (2.21) | 23 (9.75) | 30 (6.67) | 15 (8.82) | 4 (4.30) | 4 (1.71) | 88 (6.14) |
Older women | 0 (0.0) | 0 (0.0) | 2 (0.85) | 9 (2.00) | 11 (6.47) | 1 (1.08) | 4 (1.71) | 27 (1.88) |
Non-older men | 69 (60.5) | 98 (72.1) | 177 (75.0) | 293 (65.1) | 114 (67.1) | 61 (65.6) | 126 (53.8) | 938 (65.6) |
Non-older women | 14 (12.3) | 27 (19.9) | 45 (19.1) | 172 (38.2) | 49 (28.8) | 20 (21.5) | 58 (24.8) | 385 (26.9) |
Note: percentages do not total 100 within countries as some pictures contained images of people from more than one age/gender category while age/gender category could not be determined in other pictures.
We examined the association between picture content (news or non-news material) and the gender and age of the person depicted. Significant associations were obtained for depiction of an older woman (news 1.51%, non-news 3.43%,
Content analysis was undertaken of a total of 599 television scenes. With regard to the images of people whose ages could be reliably determined, 106 older people were observed in total, in comparison to 170 children or babies, 370 middle-aged people, and 855 young people. There was thus significant variation in the absolute presence of the different age groups (Friedman
The proportion of scenes containing older people varied significantly across countries (see Table
Images of people by age and gender in scenes on television from five cities.
City | ||||||
---|---|---|---|---|---|---|
Haleb | Amman | Luxor | Arusha | Sheffield | Total | |
|
|
|
|
|
|
|
Scenes examined | 165 (27.6) | 77 (12.9) | 50 (8.35) | 111 (18.5) | 196 (32.7) | 599 (100) |
Scene content | ||||||
Older men | 5 (3.03) | 7 (9.09) | 7 (14.0) | 3 (2.70) | 12 (6.12) | 34 (5.68) |
Older women | 2 (1.21) | 3 (3.90) | 3 (6.00) | 4 (3.60) | 5 (2.55) | 17 (2.84) |
Non-older men | 76 (46.1) | 48 (62.3) | 26 (52.0) | 68 (61.3) | 122 (62.2) | 340 (56.8) |
Non-older women | 61 (37.0) | 22 (28.6) | 22 (44.0) | 63 (56.8) | 84 (42.9) | 252 (42.1) |
Note: percentages do not total 100 within countries as some scenes contained images of people from more than one age/gender category while age/gender category could not be determined in other scenes.
There was no significant association between presence of older people in scenes and the period of day in which the scenes were broadcast, and there is no significant association between programme type and presence of older people.
Our data, drawn from the structured observations of public social spaces and from the content analysis of images in newspapers and on television channels, is mutually reinforcing: data from all sources indicated a low presence of older people in public fora. Overall, this low presence was particularly true for the developed country sites relative to the developed country site in our study, with women additionally considerably less present than men in developing countries. These findings support the other work that suggests that older people are infrequently visible in social spaces (e.g., [
One would expect a lower presence of older people in cities in developing countries due to the lower proportions of older people in these countries compared to developed countries. However, when the median number of older people per observation period is considered as the benchmark, Sheffield had a lower median of older people observed in social spaces than Amman. Thus, older people were generally absent from social spaces in both Sheffield and the developing country cities, but only for Sheffield there were a small number of periods when older people were frequently observed. Social spaces were not evenly used by the older people observed, with most older people observed in markets, transport hubs, and post offices. The fact that so few older people were seen in locations more associated with leisure, such as cafes and hotels, is interesting. Research in Western countries indicates that economic and health factors impact older people’s leisure activities [
Older people in our sample of developing country cities, in comparison to those in Sheffield, are less likely to be accompanied when in public. Leaving aside the very low absolute level of presence of older people in the developing country cities sampled, the finding that older people in these cities are relatively unaccompanied when observed is intriguing. Being unaccompanied may indicate that an older person has a low availability of family relationships, which would place that person at higher risk of poor health, difficulty in accessing social care, and mistreatment and abuse [
When considering older women, 94.4 per cent of those observed were recorded in Sheffield. This confounding of gender with developing versus developed country city status prevented a more nuanced examination of how gender influences women’s visibility in social spaces when comparing developed and developing contexts. However, this extremely low level of visibility of older women in our sampled developing country cities does reflect the predominant finding of gender inequalities in such countries (e.g., [
The finding that older men were significantly more likely to be depicted in newspapers from developing countries in comparison to those from the UK, especially when taking population composition into account, suggests less media discrimination against older men in developing countries than in the UK. Previous research on the representation of older people in the media supports our finding that older women were less likely to be present than older men [
The most valuable feature of this study is its uniqueness. There are no other directly comparable studies and extremely few of any kind concerned with social integration and exclusion of older people that present data from several countries with differing cultural, political, demographic, and socioeconomical profiles. As this is an exploratory study, it is essential that there is transparency about its methodological successes and limitations so that future research can improve upon this work. Many factors will have influenced the validity and reliability of the data collected and it is important to review these factors.
As mentioned in Section
While acknowledging that carrying out observations in a different set of places within the same cities may well have produced a different data profile, a strength of the study is its capacity to triangulate the data from the three different sources of social spaces, newspaper pictures, and television. The first data source is indeed geographically very tightly defined and therefore highly suspect with regard to any claim for representation of the population from which it is selected. However, the second two data sources are national/regional, not local, and therefore can be argued to be reasonably robust with regard to how they reflect the cultures from which they are drawn. The information that emerges from all three data sources is mutually reinforcing such that the credibility of the data profile drawn from observation of social spaces is strengthened due to its similarity to those drawn from newspapers and television.
Our researcher will have approached the data collection process with conscious and unconscious assumptions and beliefs, which will have been embedded within that researcher’s own cultural framework. This cultural anchoring will have affected the data collection process across the different cultures in which it occurred in fundamental and unquantifiable ways. For our study, all data was collected by the same individual. This approach was decided upon partly out of financial necessity and partly to ensure consistency of observations across all sites/countries. The researcher was well prepared through briefing, training, and piloting, and the use of a highly structured observation protocol will have maximized reliability during data collection. However, traditional indices of reliability available when more than one researcher is used during data collection (e.g., interrater reliability) cannot be determined for our study [
For all observations, there were difficulties in estimating age. Anyone appearing over the age of 60 was recorded as an older person, and where there was any uncertainty the person was excluded from the “older person” category. Estimates of age were further affected by the distance in observations used to limit the impact of the observer on the observed for social spaces data collection, the researcher’s relative unfamiliarity with people from certain ethnic groups, and clothing that obscured faces. Thus, age estimation might be anticipated to be more reliable in the UK than in some of the developing country fieldwork sites. A further issue is whether the age of 60 was a suitable age at which to draw the boundary of old age, given the discrepancy in life expectancy across the sampled countries noted in Section
Overall, despite the challenges encountered in this study, it has produced interesting and original findings, which suggest that observational methods and content analysis have utility as ways of exploring social integration and inclusion of older people in both developed and developing countries. While again emphasizing the exploratory nature of our study, there are still broad implications of our findings for policy around older people. Firstly, a lack of presence of older people in public fora implies a lack of representation also in other areas of society such as policy and planning. There is awareness that more engagement is needed with older people in policy and planning, and there have been efforts to achieve this [
Nevertheless, we would reinforce the notion that a lack of visibility of older people, a public absence of those in later life, has been identified as a form of exclusion [
There may be some concern that moving from a demonstration of the public presence or absence of a social group to the inference that such a group is socially excluded is too great leap. It may be unwise to apply a single model of social exclusion in vastly different cultural contexts without adaption to the local setting (see [
Ethical approval for the study was provided by the Sheffield Hallam University Faculty Research Ethics Committee, Faculty of Development and Society.
The authors declare that there is no conflict of interests regarding the publication of this paper.