TAO/TAO Survey Among Elderly - Wave 2

The issues of the psychosocial consequences of Internet usage in the elderly as well as the consequences of seniors’ participation in online communities have produced a considerable amount of research in the recent decades. At the start of TAO, the literature on these issues was vague and revealed conflicting results. Other problems of the research at that time were small sample sizes which have a high risk of selection bias and cross-sectional study designs from which one cannot draw causal inferences. Furthermore, data from the countries that participated in TAO were nearly not available.

Methodological considerations
Therefore, the TAO survey utilized a rationale which tried to facilitate a comparatively strong research design which allowed drawing causal conclusions. Methodologically, the claim of causality requires high standards. Within health research or epidemiology, causality is given – among other criteria – when:
 * there is a clear temporal relation (before/after),
 * the relation is plausible,
 * the relation is strong,
 * the relation is specific,
 * the relation is consistent with the current knowledge,
 * the study results can be replicated.

In order to secure many of these criteria, the main features of the TAO survey research design were The main outcome domains were defined as mental well-being and (real life) social inclusion.
 * a large sample size,
 * a longitudinal data collection in two waves,
 * the utilization of well-known and – if available – tested questionnaires and instruments,
 * the application of multivariate statistics.

The data collection was carried out via the Web-platform Surveymonkey(R) for elderly onliners and simultaneously with a paper and pencil questionnaire for elderly offliners. The questionnaire contained the following contents:
 * Sociodemographic characteristics (including social inclusion variables; mainly taken from the European Social Survey Questionnaire)
 * Computer- and Internet use (e.g., intensity, duration, contents; mainly taken from the GVU WWW User Survey)
 * Mental well-being (Mental Health Index – 5)
 * Psychosocial Consequences of Internet Use (18 items subscale of the Internet Consequences Scale – ICONS)

Wave 2 Results
The results of our second wave survey can be summarized as follows: The psychosocial consequences of Internet usage as measured by the Psychosocial Consequences Scale of the Internet Consequences Scale (ICONS) deteriorated significantly, though still remains on a comparable high level.
 * The Internet usage of the participating seniors in our survey remained stable during the 12 months between measurement point 1 and measurement point 2.
 * Real-life social inclusion patterns did not change significantly between the two survey waves, but the slight changes that could be observed suggest that for older adults the importance of information about ongoings in the world and the Internet as a tool for pleasure (games etc.) decreases over time, while its importance for staying in touch with other people and socal activities in Web 2.0 areas grows.
 * Quality of life as measured by the Mental Health Index-5 remained stable on a very high level.

These results should be discussed methodologically and in terms of seniors’ living situations. Methodologically, the 12 months time frame was possibly not suffiicently long to find any major changes either in the Internet usage or in terms of social inclusion. Another drawback was that we were unfortunately unsuccessful in recruiting enough beginning Internet users for the first wave to allow us to to find any effects of Internet usage on other variables.

A more general issue relates to the differences of cross-sectional and longitudinal surveys, which are well-known in social and medical research. When conducting the cross-sectional analyses we found many associations between Internet usage and our main outcome variables, quality of life and social inclusion. Against our working hypotheses from D5.1, these associations did not turn out to be causal. In our sample, Internet usage and Web 2.0 usage as such do not lead to better outcomes in terms of quality of life and social inclusion. We can preliminarily conclude that the reason for this lies in processes of self-selection. In other words, the reason for the positive cross-sectional correlation between Internet usage and favorable outcomes could also be that persons with a higher quality of life and higher levels of social inclusion are more likely to be using the Internet. Certain seniors employ a given usage pattern due to their lifestyle and due to sociodemographic characteristics.

Besides methological issues, one can conclude that the Internet has become an integral part of seniors’ lives. Being on the Internet does not change anything in terms of real life, neither for the good nor for the bad. From this point of view, one can see Internet usage as a broadening of the social sphere. Being on the Internet, however, does not lead to negative consequences either, as some commentators have assumed (Dickinson, 2006; Gilhooly et al., 2009; Hoogeboom et al., 2010; Xie, 2007).

We cannot explain with our data the slight, but significant, deterioration of scores in the Psychosocial Consequences Scale. One could hypothesize, however, that the public discussion about dangers on the Internet (privacy breaches, fraud etc.) had an impact on our participants. Overall, this two wave survey was an interesting study and proved to be useful getting insight into Internet usage patterns and motivations and to assess how Internet usage in this age group of the so-called ‘silver market’ affects social inclusion and well-being.

General conclusions
To sum up the results from Wave 1 (W1) and Wave 2 (W2), we found interesting associations between low as well as high social inclusion with Internet usage, an association between good mental well-being and Internet usage, and we also found an association between high social inclusion and social media usage. These associations, however, did not turn out to be causal relations when W1 data were matched with W2 data.

In the light of these results, we conclude that Internet and social media usage in our survey do not enhance well-being and social inclusion. Contrary to many assumptions, we assume that Internet and social media usage are rather ‘selected’ by seniors according to their needs and according to their mental capabilities. Thus, socially well included seniors, which build the majority of onliners, use the Internet to broaden their communication pathways. In our survey, only few socially isolated seniors used the Internet for staying in contact with others.

From a psychological perspective, the ‘selection’-hypothesis makes more sense than the ‘causation’-hypothesis. Using the Internet and social media is ambitious and challenging to elderly beginners. The prerequisites are rather high in terms of self-esteem and self-efficacy; computer anxiety, on the other hand, needs to be low. Being online and using social media is much easier when the mental well-being is high, social inclusion is strong, and – not to forget – one is used to use the computer. Many predominantly male participants in our survey have started using the computer in their working life and have maintained the use of a computer after retirement.