were the second and the third autho

were the second and the third author of the article, and Aune Valk from the University
of Tartu. The questionnaire consisted of several parts from which only the self-esteem
measure was relevant for this study. Global self-esteem was measured by the SISE
among 1,395 native Estonians (56% females and 44% males) with the mean age of 43.5
years (SD = 17.6) ranging from 15 to 74 years. The sample was randomly selected from
the National Census and was representative of the Estonian-speaking population in
Estonia concerning residence geography, age, gender, and educational level. About 28%
of the respondents had completed a basic education, 55% secondary school level, 14%
had a university degree, and about 3% declined to declare their educational level. About
8% were the residents of the capital city, 30% were living in different towns, and 62%
in rural areas. The survey was carried out by the TNS Emor, the major service marketing
research and consulting company in Estonia.
Sample R2*. Data from the second sample was derived from a larger study which
aimed to study changing values, attitudes, attributes and behaviour patterns within
European polities. The European Social Survey (ESS) is a biennial multi-country survey
covering over 20 nations. The project is directed by a Central Co-ordinating Team led
by Roger Jowell at the Centre for Comparative Social Surveys, City University,
London. Estonia participated in the second round of the ESS which took place from
December 2004 to January 2005. The survey was representative of all persons aged 16
and over (no upper age limit) resident within private households in Estonia, regardless
of their nationality, citizenship or language. The sample was selected by strict random
probability methods at every stage and respondents were interviewed face-to-face. The
survey was carried out by the Statistical Office of Estonia. It should be emphasized that
the SISE was not included in the general ESS questionnaire (applied in all participating
countries) but only in the Estonian survey. The Estonian version of the ESS
SELF-ESTEEM ACROSS THE LIFE SPAN 9

questionnaire was administrated to the 1,469 native Estonians (59% females and 41%
males) with the mean age of 47.9 years (SD = 19.8) ranging from 16 to 98 years.
Among these participants, about 44% had completed a basic education, 37% had
finished secondary or a vocational school, and 19% had a university degree. The
complete data for the SISE was available from 1,457 participants (99% of the total
sample).
Sample R3. The third representative sample of 655 adult individuals (52% females
and 48% males) completed the Estonian version of the RSES in 2000 as a part of a
wider study of the integration processes in Estonia. This project was coordinated by Jüri
Kruusvall and Raivo Vetik from Tallinn University. The mean age of the participants
was 42.0 years (SD = 17.3) ranging from 15 to 74 years and this sample was
representative of the Estonian-speaking population in Estonia. In this sample, 36% had
completed a basic education, 47% had finished secondary or a vocational school, 16%
had a university degree, and about 1% declined to declare their educational level. The
internal reliability coefficient (Cronbach α) of the RSES was α = .81. The survey was
carried out by the Saar Poll Ltd, the second largest social and market research company
in Estonia.
Sample R4. The representative sample of Estonian adolescents (N = 2,708; mean
age = 14.9 years ranging from 11 to 18, SD = 2.04) was drawn from 27 socially and
geographically representative schools from all of the 15 Estonian counties, including the
capital city of Tallinn, smaller towns (e.g., Tartu) and rural areas in 2001. Similarly to
general population at this age, the current sample contained approximately equal
numbers of boys (47%) and girls (53%). The project was coordinated by the first and
the second author of this article. The internal reliability coefficient was α = .81.
SELF-ESTEEM ACROSS THE LIFE SPAN 10

Sample I5. This self-recruited Internet sample consisted of 23,248 Estonian-speaking individuals (63% females and 37% males; mean age was 29.3 years ranging
from 9 to 89, SD = 8.73) who visited a non-commercial, advertisement-free web site
and completed an on-line version of the RSES during 2000-2001. After filling in the
entire questionnaire, the participants received the online feedback-sheet with a
comparison of their results with the other. While only information about the
participants’ gender and age was asked to complete, they did not provide any personal
identification, therefore, complete anonymity was assured. In total, about 10% of all
Estonian Internet-users participated in this study whereas about the third of the Estonian
population from 15 to 74 years had an access to the Internet during the data collection
according to the regular e-track survey conducted by the TNS Emor (www.emor.ee, for
details). In this sample, the internal reliability coefficient of the RSES was α = .82.
In this study, all the samples were divided into 9 age groups using intervals
identical with that used by Robins and his colleagues (2002): 9-12, 13-17, 18-22, 23-29,
30-39, 40-49, 50-59, 60-69, and 70-90 years. Two representatives samples tested by the
RSES – adults (R3) and adolescents (R4) – were combined for subsequent analyses to
reflect the entire life span from 11 to 74 years with a partial overlap in two age groups:
13-17 and 18-22 years (3% vs 97% and 12% vs 88% for the R3 and R4, respectively).
The numbers of participants across age groups for the samples are reported in Table 1.
Insert TABLE 1 about here
SELF-ESTEEM ACROSS THE LIFE SPAN 11

RESULTS
Age Differences in Self-Esteem Measured by the Single-Item Self-Esteem Scale
Figure 1 shows cross-sectional trajectories of global self-esteem across the life span for
all samples. The third curve from the top (crosses) is reproduced from Robins et al.
(2002; Table 2, p. 429) and is based on the data collected from 326,641 individuals over
the Internet who expressed their agreement with a single statement “I see myself as
someone who has high self-esteem” on a 5-point scale. The lowest curve (squares)
present the mean levels of self-esteem across the age groups based on answers to the
SISE of 2,852 Estonians who were randomly selected from the whole population
(averaged data of Samples R1* and R2*). These two normative trajectories of self-esteem are not very similar. According to data reported by Robins and his colleagues
(2002), self-esteem gradually rises through adulthood and reaches it highest peak in the
60s. Unlike Robins’ data (2002), two representative Estonian samples (R1*+R2*) tested
by the SISE scale reach the maximum level in the 30s and demonstrate after that a
steady decline toward lower self-esteem level in accordance with the data reported by
McMullin and Cairney (2004). Concerning the mean levels, the multinational Internet
sample demonstrates significantly higher scores compared to the Estonian
representative samples in all age groups. However, the mean level differences between
these samples are hard to interpret because of differences in language (English vs.
Estonian, respectively). Because Pearson correlation is insensitive to the mean level
differences, we computed congruence coefficients between the mean levels of self-esteem trajectories of a multinational Internet sample and two nationally representative
Estonian samples. The results confirmed that even a small number of cases is enough to