A researcher at the University of Gothenburg has refuted popular notion that old people are a rather homogenous group - as they are considered to be ill, lonely and unable to take care of themselves. 'But the truth is', he says, 'that the differences among people grow with age'.
As part of his doctoral thesis, Bo G Eriksson conducted a study that followed people from their 70th to their 90th year of life.
The expert studied participants of the extensive and unique so-called H-70 study, which is based on a group of randomly selected individuals born in 1901 and 1902 who were followed closely over their entire lifetimes.
Eriksson's study focused on the period from their 70th to their 90th year of life. And it turned out that people become more and more different as they age.
'The perception of old people having similar interests, values and lifestyles can lead to age discrimination. However, I found that, as people age, these stereotypes become more and more untrue,' says Eriksson.
Eriksson also studied differences in causes of death with increasing age, and again found indications of possible age discrimination.
Eriksson explored how social conditions can affect longevity, and found four mechanisms at work. The first two relate to creation of social facts. Examples of social facts include promises and agreements that strengthen the identities of individuals. The third mechanism relates to how a person builds and maintains self esteem by successfully responding to challenges. The fourth mechanism consists of everyday conversations, which decrease anxiety and offer support in everyday decision making, improves attention and gives the brain and the memory a healthy workout.
"Taken together, these mechanisms also contribute to increased everyday activity, which has some beneficial physical effects," says Eriksson.
Moreover, Eriksson applied two different methods to predict people's lifespan: one that researchers commonly use when calculating probability and one that is based on artificial neural networks (ANN), which is common in research on artificial intelligence. It turned out that the ANN method was more effective in complex situations where traditional methods do not work. ANN may therefore be appropriate in evaluations of results produced with traditional research methods.