Actual and imagined life satisfaction 1

Actual and imagined life satisfaction 2 Abstract Do differences in reported life satisfaction between societies reflect real

Can we compare life satisfaction between nationalities? – Evaluating actual and

differences or do they reflect cultural differences in the way people rate their

imagined situations

experiences? Study 1 showed that imagining better or worse life situations affected student respondents’ ratings of their own life satisfaction, as predicted by range-

Friedel Bolle

frequency theory. Study 2 investigated how German and Polish students rated their

Europa University Viadrina

actual life satisfaction and how satisfied they would be under three imagined scenarios. Study 3 similarly compared Danish and Hungarian students. Both studies

Simon Kemp

found significant differences in the rating of the hypothetical situations, and moderate

University of Canterbury

correlations between ratings of satisfaction in the hypothetical situations and reality, but in neither study were national differences in actual satisfaction predicted by differences in hypothetical satisfaction. Overall, the results suggest that national differences in rated life satisfaction are real rather than reflecting differences in how satisfaction is rated.

Author notes. We are grateful to Anika Köhler, Geza Sapi, and Marta Sernec for their help in obtaining the samples of Studies 2 and 3. Simon Kemp gratefully acknowledges the sponsorship of the Alexander von Humboldt Foundation for this research. Correspondence should be directed to Friedel Bolle, Europa Universität Viadrina, Frankfurt (Oder), Germany; email: [email protected], or to Simon Kemp, Psychology Department, University of Canterbury, Christchurch, New Zealand; email: [email protected].

Actual and imagined life satisfaction 3 Can we compare life satisfaction between nationalities? – Evaluating actual and imagined situations

Actual and imagined life satisfaction 4 number of issues. The one we focus on here is the comparability of the responses between people and particularly societies. As Kahneman and Kruger (2006, p. 18) put it: “If Jim says that he is ‘very satisfied’ and Tim says that he is only ‘satisfied’, is Jim

In this paper, we describe two different ways of using scenarios to shed some

really more satisfied than Tim? But maybe Tim is the type of person who rarely uses

light on an issue of considerable and growing importance: How seriously should

superlatives to describe himself … while Jim tends to extremes in his self-

people’s self-reports of their life satisfaction be taken?

descriptions.”

There is now a considerable body of research into people’s subjective well-

Comparability of responses may be a more obvious problem when comparing

being. (For reviews, see, e.g., Diener et al., 1999; Easterlin, 2006; Frey & Stutzer,

individuals than when comparing societies, since one would expect many individual

2002; Kahneman, Diener & Schwarz, 1999.) Unsurprisingly, such research has led to

differences to average out when societies are compared. But it is quite possible that

a number of applications, most obviously in suggesting to people how they might lead

there are systematic differences between societies. For example, maybe some

happier and more satisfying lives (e.g. Parducci, 1995), but also in attempting to

societies do use superlatives more often than others.

evaluate policies and ways of life pursued by societies. For example, there have been

This possibility increases if we consider the ways in which people come to use

suggestions that national indices of subjective well-being should supplement more

different response categories, either on a numerical scale or with verbal labels such as

conventional measures such as per capita gross domestic product and life expectancy

“very satisfied” or “not at all satisfied”. There has been considerable research on this

(e.g. Diener, 2000; Diener & Seligman, 2004). A number of large-scale surveys, such

subject, and it has often been shown that the way people assign stimuli to different

as the Eurobarometer and the World Values Survey, incorporate a global measure of

categories can be affected by a number of contextual variables (see Poulton, 1989, for

life satisfaction. A World Map of Happiness can now be tracked down on the internet

a review). One influential account of how people assign categories to stimuli that are

(e.g. White, 2007).

ordered on some continuum is Parducci’s (1965) range-frequency theory. Very

The global measures of life satisfaction that are incorporated in surveys are

basically, the theory suggests that the boundaries between categories are a

normally simple questions which are answered using a simple rating which has

compromise between placing them at equal intervals as determined by the perceived

between 3 and 11 response categories. For example, the Eurobarometer asks: “On the

overall range and placing equal numbers of stimuli in each category. Parducci (1995)

whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied

himself explored the implications of range-frequency theory for how people’s

with the life you lead?” It is also fairly common to ask “All things considered, how

happiness might be affected by the sequence of their individual positive and negative

satisfied with your life as a whole these days?” and for respondents to answer on a

experiences, concluding that people are likely to be happiest when they have a large

scale from 0 to 10 (e.g. Frey and Stutzer, 2000). Using such simple measures raises a

number of their own best experiences. (For another, rather different account of how

Actual and imagined life satisfaction 5 people might construct satisfaction from past experience, see Kahneman and Krueger, 2006).

Actual and imagined life satisfaction 6 The second use of scenarios in this paper follows a somewhat different consideration. Suppose we find a difference between two societies in their average

The focus of the present paper is on how people might categorise their life

rating of life satisfaction. Of course, there are a variety of credible reasons for the

satisfaction by comparison with other people rather than summing up their own past

difference, but, in the interests of simplicity, we consider just two. Either there is a

experiences. However, effects of range and frequency might be expected here too.

real difference in life satisfaction between the two societies, or the difference only

The well-known finding that improvements in either national or personal living

reflects differences in the way people in the two societies assign ratings to essentially

circumstances do not generally lead to substantial, long-term improvements in rated

the same life satisfaction. How could we distinguish these two?

life satisfaction is consistent with this account of contextual effects (e.g. Brickman,

One possibility, which is investigated in the second and third studies below, is

Coates, & Janoff-Bullman, 1978; Easterlin, 1995; but see also Veenhoven & Hagerty,

to ask the respondents not only to rate their own life satisfaction but also to rate the

2006). More generally, one would expect differences in both the range and frequency

life satisfaction they would expect to have in different situations. If people in the two

of life satisfaction between societies, and this might lead us to expect differences in

societies (or from different cultural backgrounds) give similar ratings for how

the way people within those societies use the categories. For example, if it were

satisfied they would be in the same hypothetical situation, it is an indication that they

possible to occasionally see very miserable people in Society A but not B (perhaps

assign ratings to states of life satisfaction in much the same way. If, however, people

because they are hidden away in B), then the category boundaries would be shifted

in the two societies rate their satisfaction in the hypothetical situation rather

downwards in Society A relative to B. According to the range-frequency theory (and

differently, this is an indication that they might rate the same level of life satisfaction

ignoring the effects of empathy), most individuals in A might then see themselves as

differently as well.

more satisfied with their lives in consequence. Experimentally, the effect of context is tested in Study 1 by presenting student

Study 1 In this study, respondents read and assessed the satisfaction of an imaginary

respondents with information about the love-life, study experience, friendships,

fellow student in three different scenarios, with half the respondents reading “good”

financial situation and health of hypothetical others. Some respondents were

and half “poor” scenarios. They were then asked to rate their own present satisfaction

presented with scenarios in which the hypothetical others appeared to have good lives,

and how satisfied they would be in a standard imaginary scenario. We hypothesised

and some were presented with scenarios in which the others had relatively poor lives.

that respondents reading the poor scenarios would form categories and ratings centred

The expectation was that the former students would in consequence rate their own life

on less appealing objective scenarios than those who read the good scenarios.

satisfaction less favourably than the latter.

Consequently we expected that when they came to rate their own current satisfaction

Actual and imagined life satisfaction 7

Actual and imagined life satisfaction 8

and how satisfied they would be in the standard scenario, those who had read poor

The scenario that was common to the two questionnaires read:

scenarios would rate their own lives and the standard scenario as more satisfying.

“Imagine:

Method Two questionnaires were prepared. These are called the good and poor questionnaire. The cover page was the same for both. It was headed Student Life Satisfaction Questionnaire, and simply informed respondents they would be asked “a few questions about how life might be for a student in different situations”. The following three pages each contained a scenario for “a student at this university”, and immediately following the scenario the respondent was asked to rate the satisfaction they thought the student might experience on two different scales. A verbal scale presented the four alternatives: very satisfied; fairly satisfied; not very satisfied; and not at all satisfied. Respondents were also asked to rate the student’s satisfaction on a scale from 0 to 10, “where 0 indicates ‘completely dissatisfied’ and 10 indicates ‘completely satisfied’”. In both questionnaires, the scenarios were sequenced so that the expected

The student does not have a permanent partner or boy/girlfriend but (s)he knows one or two people who seem interested in her/him and (s)he finds interesting in return. Some of the subjects (s)he is studying at university are boring, but others are fairly interesting. Generally (s)he gets average grades. (S)he has a large number of friends in Christchurch, many of whom (s)he finds to be very interesting people. (S)he would like to have more money but (s)he has enough to live on as a student. Generally (s)he has good health, but of late (s)he has sometimes had difficulty sleeping.” The better and worse scenarios featured variations on the same five themes (love life; university study; friendship; money; and health). By way of example, the

worst scenario of the three viewed was first, the expected best scenario was second,

best love life description read “The student has a permanent partner or boy/girlfriend.

and an intermediate scenario was last. The scenarios differed between the good and

(S)he is very proud to have this person as a partner, who virtually everyone in the

poor questionnaires so that the expected worst scenario of the good questionnaire

university believes to be drop-dead gorgeous”. The worst health description read:

(sequenced first in that questionnaire) was identical to the expected best scenario of

“The student has a permanent disability. This disability restricts her/his movement a

the poor questionnaire (sequenced second in that questionnaire). Before the first

good deal, causes a reasonable amount of pain, and the prognosis is that the condition

scenario, respondents were asked “to imagine a student at this university whose life

will slowly but surely get worse.” Occasionally, the same theme description was used

has a number of features, and we ask you to rate how satisfied you think the student

for two scenarios.

would be with his or her life. Please imagine someone about your age and of the same

After reading and rating the scenarios, the respondents was asked for their age

sex, but not a particular person you know.” Before next two scenarios they were told

and gender and to rate their satisfaction with “your own life as a whole these days”

to perform the same task but with different life circumstances.

using both the verbal and numerical scales. Finally, on the last page of the

Actual and imagined life satisfaction 9

Actual and imagined life satisfaction 10

questionnaire, the respondent was asked to imagine themselves in the circumstances

scale: own life satisfaction t(77) = - 3.51, p < .001; satisfaction with common scenario

of the common scenario. This was presented in full but reworded from the third to the

t(76) = - 2.61, p < .05. Numerical scale (/10): own life satisfaction t(77) = 2.60, p