GLOBAL SELF-ESTEEM ACROSS THE LIFE SPAN: A CROSS...

GLOBAL SELF-ESTEEM ACROSS THE LIFE SPAN:A CROSS-SECTIONAL COMPARISON BETWEENREPRESENTATIVE AND SELF-SELECTEDINTERNET SAMPLES

Helle PullmannJuri AllikAnu Realo

Department of Psychology, University of Tartu, Tartu, Estonia; and TheEstonian Centre of Behavioural and Health Sciences, Tartu, Estonia

The cross-sectional trajectory of global self-esteem across the life spanwas examined administering the Rosenberg and Single-Item Self-Esteemscales to Estonians (N¼ 29,463) who were either randomly selected fromthe National Census to represent the population or self-recruited throughthe Internet. The results (a) challenge the recent conclusion of a universalage trajectory of self-esteem, (b) demonstrate that self-recruited Internetdata collection method is biased compared to random sampling, and(c) present that different self-esteem items have dissimilar trajectories.A variance component analysis confirmed that age differences in self-esteem are relatively small compared to interindividual differences andmeasurement error.

Based on a huge data set collected from 326,641 individuals over theInternet, Robins and his colleagues proposed that there was a

Received 4 December 2005; accepted 15 June 2007.

This project was supported by grants 5677, 5287, and 7020 from the Estonian Science

Foundation and by an Estonian Ministry of Science and Education grant SF0182585s03 to

Juri Allik. Portions of this article were presented at the 11th Biennial Meeting of International

Society for Studies for Individual Differences, July 2003, Graz, Austria. The authors thank

Juri Kruusvall, Ly Kastik, Kaia Laidra, Aune Valk, and the Estonian Coordinating Team of

the European Social Survey for their assistance with collecting the data. The authors are also

grateful to Earl of Carlisle, two anonymous reviewers, and especially Dr. Jeffrey W. Elias for

their helpful comments and suggestions.

Address correspondence to Helle Pullmann, Department of Psychology, University of

Tartu, Tiigi 78, 50410 Tartu, Estonia. E-mail: [email protected]

Experimental Aging Research, 35: 20–44, 2009

Copyright # Taylor & Francis Group, LLC

ISSN: 0361-073X print/1096-4657 online

DOI: 10.1080/03610730802544708

universal trajectory of self-esteem across the life span (Robins,Trzesniewski, Tracy, Gosling, & Potter, 2002; Robins & Trzesniewski,2005). On average, self-esteem estimated by a single-item scale (‘‘I seemyself as someone who has high self-esteem’’) is relatively high inchildhood, drops during adolescence (particularly for girls), risesgradually throughout adulthood, and declines sharply in old age. Thistrajectory presumably holds across gender, socioeconomic status,ethnicity, and nationality (Robins et al., 2002). However, all thesenormative changes in the mean level of self-esteem are relatively smalland do not exceed one third of standard deviation.

These findings, along with some earlier studies (e.g., Demo, 1992;McCarthy & Hoge, 1982; Robins, Norem, & Cheek, 1999), challengethe standard view according to which there are no systematic age dif-ferences in global self-esteem (Wylie, 1979), the latter constructdefined as the individual’s generalized positive or negative attitudetoward the self as a human being (Rosenberg, 1965). The consider-able overlap with the Big Five personality dimensions may explainthe normative trajectory of self-esteem. Indeed, it is well documentedthat measures of self-esteem correlate strongly with two Big Five per-sonality dimensions, Neuroticism and Extraversion (Costa, McCrae,& Dye, 1991; Judge, Erez, Bono, & Thoresen, 2002; Kwan, Bond, &Singelis, 1997; Pullmann & Allik, 2000; Robins, Tracy, Trzesniewski,Gosling, & Potter, 2001) and that these relationships remaininvariant across at least 53 cultures (Schmitt & Allik, 2005). Basedon similar observations, Judge and colleagues (2002) proposed thatself-esteem represents a general Neuroticism factor and contains onlya small amount of unique information that goes beyond it (Judge,et al., 2002). This may indicate that self-esteem originates from per-sonality dispositions, which, according to the Five-Factor Theoryof personality (McCrae & Costa, 1999; Allik & McCrae, 2002), aredeeply rooted in biology and remain relatively stable throughoutthe whole life course. Indeed, self-esteem appears to be a heritabletrait (about 30% to 50%), although being also subject to consider-able environmental influences that in most cases are probably notshared by members of a twin pair (Kendler, Gardner, & Prescott,1998; Roy, Neale, & Kendler, 1995; for a review see Neiss, Sedikides,& Stevenson, 2002). Based on cross-sectional and longitudinal stu-dies, Costa and McCrae (2002) identified a specific pattern of changesfor each main personality trait; the changes are quite small comparedwith the general stability of personality traits. In most cultures, neur-oticism, extraversion, and openness appear to decrease with age afterlate adolescence, whereas agreeableness and conscientiousnessincrease (Costa & McCrae, 2002; McCrae et al., 2000, 2004; Roberts,

Self-Esteem Across the Life Span 21

Walton, & Viechtbauer, 2006). To the extent the inverse value ofself-esteem is an indicator of neuroticism, the reported gradual riseof self-worth through adulthood is compatible with the normativetrajectory of the decrease in neuroticism.

Albeit impressive, single-item self-esteem data collected throughthe Internet (Robins et al., 2002) probably do not give an adequatepicture about the normative trajectory of self-esteem across the lifespan. Firstly, Internet samples are likely to be selective, which placeslimits on external validity. For instance, not all people have access tothe Internet (Walsh, Kiesler, Sproull, & Hesse, 1992) and there is stilla large number of people who have only a limited access to a com-puter despite exponential growth of the Internet (Barbeite & Weiss,2004; Birnbaum, 2004b; Crawford, Couper, & Lamias, 2001).Further, those who have better access may not be representative ofthe population in general (Smith, 2001; Birnbaum, 2004b). Moreover,participants in the Internet-based research may have a differentmotivation for taking a test (Buchanan & Smith, 1999) and as wasargued by Reips (2000), the typical Internet test-taker is more moti-vated than a typical laboratory participant. Until the problem ofself-selection has been resolved, the results collected through theInternet cannot be generalized to the general population (Birnbaum,2004a). Although Robins and his colleagues (2002) provided severalstrong arguments about the validity of the Internet method (e.g.,similarity of the results across gender and nationality), the data col-lected through the Internet, as Robins, Tracy, et al. (2001) admit,retain in the status of preliminary findings until they are replicatedon randomly selected, representative samples of individuals. There-fore, future research should examine age differences in self-esteemusing representative samples of individuals, which would greatlystrengthen the generalizability of these conclusions and alleviate con-cerns about the Internet sample (Robins et al., 2002, p. 432).

Secondly, the conclusions from studies based on single-item mea-sures might not replicate because there may be idiosyncrasies aboutthe single item, or interactions of item with cohort or sample. Theseare real dangers because the large sample size makes trivial effects sig-nificant, and artifacts may be replicable across subsamples. Forexample, a national probability sample of 16,000 respondents inCanada, who answered an abridged six-item Rosenberg Self-EsteemScale (RSES; Rosenberg, 1965), showed no gradual increase of self-esteem throughout adulthood. On the contrary, self-esteem graduallydeclined with increase of age for both sexes and all income groups(McMullin & Cairney, 2004). This discrepancy may be due tosampling bias of the Internet sample, but it may be also that the

22 H. Pullmann et al.

sum-score of six items behaves differently from the single-itemmeasure of self-esteem. There is no evidence that the normativetrajectories of single items of the same scale are identical. Althoughcorrelation between the RSES and single-item measure of self-esteemis relatively high in all age groups (Robins, Hendin, & Trzesniewski,2001), this does not preclude a possibility that respondents of differ-ent age groups interpret the single item slightly differently, whichcould lead to selective elevation or drop of the mean score of this itemin different age groups.

The main aim of the present study was to examine age differencesin self-esteem across the entire life span, including representative sam-ples. Except for a few studies (e.g., McMullin & Cainer, 2004), mostof the literature describing self-esteem across the life span relies onnonrepresentative samples of particular age cohorts. The second goalof the study was to examine the validity of the Internet data collectionmethod and of the single-item measure of self-esteem. More specifi-cally, we examined normative age trajectory of global self-esteemamong three nationally representative Estonian samples and com-pared these results with the data of self-recruited cross-sectional sam-ple collected via the Internet. We focus on normative changes inglobal self-esteem (henceforth a cross-sectional trajectory across thelife span), which is assessed by mean-level differences in self-regardacross age groups over time. Mean-level change refers to changesin the average self-esteem level of a population and indicates whetherthe sample as a whole is increasing or decreasing on this trait.Although cross-sectional studies are unable to isolate age effects fromperiod and cohort effects, they can estimate the existence of the meanlevel changes in self-esteem across age.

METHODS

Measuring Instruments

Rosenberg Self-Esteem Scale (RSES)The Estonian version of the Rosenberg Self-Esteem Scale (Pullmann& Allik, 2000) was administered to the participants to measure theirglobal self-esteem. Rosenberg (1965) defined self-esteem as a feelingof self-acceptance, self-respect, and positive self-evaluation that isconceptualized as a relatively enduring characteristic rather thansomething that shifts abruptly from one situation to another. Itemsof the RSES (see Table 2) were answered on a 5-point scale rangingfrom 1 (strongly disagree) to 5 (strongly agree). All results arereported in terms of the mean value across all 10 items.


Single-Item Self-Esteem Scale (SISE)An Estonian version of the Single-Item Self-Esteem Scale (‘‘Mul onkorge enesehinnang’’) was used in this study as a literal translationof the original SISE item ‘‘I have high self-esteem’’ (Robins, Hendin,et al., 2001). Items of the SISE were answered on a 5-point scale ran-ging from 1 (strongly disagree) to 5 (strongly agree). For mnemonicpurposes, samples in which the SISE was used instead of the 10-itemRSES scale are marked with an asterisk (�).

Samples

Five different samples were used in this study. The first three sampleswere nationally representative samples of adult Estonians randomlyselected from the National Census (Rahvastikuregister), the fourthwas a geographically representative sample of adolescents, and thelast sample self-recruited through the Internet. According to the Stat-istical Office of Estonia (2000), there were about 690,000 (53%females and 47% males) native Estonian-speaking individuals inage from 15 to 74 years living in Estonia in 2000. From those, about42% had completed a basic education, 46% had finished a secondaryor vocational school, and 13% had a university degree. In this study,representative samples were stratified to match National Census dataon the variables of age, gender, educational attainment, and theapproximate proportion between urban-rural residencies.

Sample R1�

The first sample came from a wider study of social capital, culturalvalue dimensions, and identity in Estonia in 2002. The project princi-pal investigators were the second and third authors of the article, andAune Valk from the University of Tartu. The questionnaire consistedof several parts from which only the self-esteem measure was relevantfor this study. Global self-esteem was measured by the SISE among1395 native Estonians (56% females and 44% males) ranging from15 to 74 years old, with the mean age of 43.5 years (SD¼ 17.6). Thesample was randomly selected from the National Census and was rep-resentative of the Estonian-speaking population in Estonia regardingresidence geography, age, gender, and educational level. About 28%of the respondents had completed a basic education, 55% secondaryschool level, 14% had a university degree, and about 3% declinedto declare their educational level. About 8% were the residents ofthe capital city, 30% were living in different towns, and 62% in ruralareas. The survey was carried out by the TNS Emor, the major servicemarketing research and consulting company in Estonia.


Sample R2�

Data from the second sample was derived from a larger study thataimed to study changing values, attitudes, attributes, and behaviourpatterns within European polities. The European Social Survey(ESS) is a biennial multicountry survey covering over 20 nations.The project is directed by a Central Coordinating Team led by RogerJowell at the Centre for Comparative Social Surveys, City University,London. Estonia participated in the second round of the ESS, whichtook place from December 2004 to January 2005. The survey was rep-resentative of all persons aged 16 and over (no upper age limit) resi-dent within private households in Estonia, regardless of theirnationality, citizenship, or language. The sample was selected by strictrandom probability methods at every stage and respondents wereinterviewed face-to-face. The survey was carried out by the StatisticalOffice of Estonia. It should be emphasized that the SISE was notincluded in the general ESS questionnaire (applied in all participatingcountries) but only in the Estonian survey. The Estonian version ofthe ESS questionnaire was administrated to 1469 native Estonians(59% females and 41% males) ranging from 16 to 98 years old, withthe mean age of 47.9 years (SD¼ 19.8). Among these participants,about 44% had completed a basic education, 37% had finished sec-ondary or a vocational school, and 19% had a university degree.The complete data for the SISE was available from 1457 participants(99% of the total sample).

Sample R3The third representative sample of 655 adult individuals (52%females and 48% males) completed the Estonian version of the RSESin 2000 as a part of a wider study of the integration processes inEstonia. This project was coordinated by Juri Kruusvall and RaivoVetik from Tallinn University. The participants ranged from 15 to74 years old, with mean age 42.0 years (SD¼ 17.3), and this samplewas representative of the Estonian-speaking population in Estonia.In this sample, 36% had completed a basic education, 47% had fin-ished secondary or a vocational school, 16% had a university degree,and about 1% declined to declare their educational level. The internalreliability coefficient (Cronbach a) of the RSES was a¼ .81. The sur-vey was carried out by the Saar Poll Ltd., the second largest socialand market research company in Estonia.

Sample R4The representative sample of Estonian adolescents (N¼ 2708; meanage¼ 14.9 years, range 11 to 18, SD¼ 2.04) was drawn from 27socially 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 populationat this age, the current sample contained approximately equal num-bers of boys (47%) and girls (53%). The project was coordinatedby the first and second authors of this article. The internal reliabilitycoefficient was a¼ .81.

Sample I5This self-recruited Internet sample consisted of 23,248 Estonian-speaking individuals (63% females and 37% males; mean age¼ 29.329.3 years, range 9 to 89, SD¼ 8.73) who visited a noncommercial,advertisement-free Web site and completed an online version of theRSES during 2000 to 2001. After filling in the entire questionnaire,the participants received the online feedback sheet with a comparisonof their results with others’. Although the participants were asked tocomplete gender and age information, they did not provide any per-sonal identification, therefore, complete anonymity was assured. Intotal, about 10% of all Estonian Internet users participated in thisstudy, whereas about a third of the Estonian population from 15 to74 years old had access to the Internet during the data collectionaccording to the regular e-track survey conducted by the TNS Emor(see www.emor.ee, for details). In this sample, the internal reliabilitycoefficient of the RSES was a¼ .82.

Table 1. Number of participants and descriptive statistics of the global self-

esteem scales by age groups for the samples

The Single-Item Self-Esteem Scale (SISE) The Rosenberg Self-Esteem Scale (RSES)

R1� R2� R3 1 R4 I5

Age

groups N M SD N M SD NR3 NR4 M SD N M SD

9–12 — — — — — — — 471 3.76 .68 33 3.71 .83

13–17 94 3.06 1.02 56 3.11 .89 64 1833 3.74 .66 539 3.87 .71

18–22 118 3.03 1.05 140 3.12 .95 55 404 3.80 .63 4831 4.03 .66

23–29 152 3.22 1.02 148 3.08 .94 84 — 3.91 .63 8359 4.06 .65

30–39 257 3.28 .93 209 3.02 .98 112 — 3.84 .63 6283 4.06 .64

40–49 227 3.10 1.01 205 2.94 1.06 107 — 3.70 .68 2493 4.01 .66

50–59 196 3.06 1.00 251 2.83 .97 104 — 3.78 .67 642 3.99 .69

60–69 253 3.13 1.01 205 2.87 .98 85 — 3.90 .60 58 4.01 .66

70–90 98 3.19 1.17 243 2.68 .91 44 — 3.99 .55 10 3.96 .78

Total 1395 3.14 1.01 1457 2.92 .98 655 2,708 3.77 .66 23,248 4.04 .65

Note. N¼number of participants; M¼mean; SD¼ standard deviation; NR3¼number of

participants in the sample R3; NR4¼ number of participants in the sample R4.


In this study, all the samples were divided into 9 age groups usingintervals 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 70–90years. Two representatives samples tested by the RSES—adults(R3) and adolescents (R4)—were combined for subsequent analysesto reflect the entire life span from 11 to 74 years, with a partial over-lap in two age groups: 13–17 and 18–22 years (3% versus 97% and12% versus 88% for the R3 and R4, respectively). The numbers ofparticipants across age groups for the samples are reported in Table 1.

RESULTS

Age Differences in Self-Esteem Measured by the Single-ItemSelf-Esteem Scale

Figure 1 shows cross-sectional trajectories of global self-esteem acrossthe life span for all samples. The third curve from the top (crosses) isreproduced from Robins et al. (2002; Table 2, p. 429) and is based onthe data collected from 326,641 individuals over the Internet whoexpressed their agreement with a single statement ‘‘I see myself assomeone who has high self-esteem’’ on a 5-point scale. The lowestcurve (squares) present the mean levels of self-esteem across the agegroups based on answers to the SISE of 2852 Estonians who were ran-domly selected from the whole population (averaged data of samplesR1� and R2�). These two normative trajectories of self-esteem are notvery similar. According to data reported by Robins and his colleagues(2002), self-esteem gradually rises through adulthood and reaches ithighest peak in the 60s. Unlike Robins’s data (2002), two representa-tive Estonian samples (R1� þR2�) tested by the SISE scale reach themaximum level in the 30s and demonstrate after that a steady declinetoward lower self-esteem level in accordance with the data reportedby McMullin and Cairney (2004). Concerning the mean levels, themultinational Internet sample demonstrates significantly higherscores compared to the Estonian representative samples in all agegroups. However, the mean level differences between these samplesare hard to interpret because of differences in language (English versusEstonian, respectively). Because Pearson correlation is insensitive tothe mean level differences, we computed congruence coefficientsbetween the mean levels of self-esteem trajectories of a multinationalInternet sample and two nationally representative Estonian samples.The results confirmed that even a small number of cases is enoughto claim that the form of trajectories is not identical since the coef-ficient was negative (r(8)¼�.36, p¼ .382).


It is noteworthy that although two Estonian samples, R1� andR2�, were both randomly drawn from the National Census andexactly the same question was used, the mean levels of the SISE dif-fered slightly across the life span. The maximum difference reached0.51 scale among the oldest age group (age 70 to 79 years, d¼ .51,p< .001) and the analysis of variance (ANOVA) revealed that theoverall difference between these two samples was statistically signifi-cant, F(1, 339)¼ 13.44, p< .001. After adjustment for educationallevel and gender, the analysis of covariance (ANCOVA) revealed asignificant effect for the samples across age groups, F(7, 2821)¼2.32, p< .05. The mean values of the SISE scale across age groupsfor these samples are reported in Table 1.

Figure 1. The cross-sectional life span trajectories of global self-esteem

measured by the RSES and SISE. Vertical bars denote 0.95 confidence inter-vals. Data for the multinational Internet sample are adapted with permission of

the first author from ‘‘Global self-esteem across the life span,’’ by R. W.

Robins, K. H. Trzesniewski, J. L. Tracy, S. D. Gosling, & J. Potter (2002),

Psychology and Aging, 17, 429. Copyright 2002 by the American Psychologi-

cal Association.


Age Differences in Self-Esteem Measured by the RosenbergSelf-Esteem Scale

In Figure 1 (top) are presented two cross-sectional age trajectoriesmeasured by the Estonian version of the RSES for the two types ofsamples: nationally representative sample and the Estonian self-recruited Internet test-takers. The upper curve (circles) demonstratesmean-level changes of self-esteem across the age range based on datareceived from 23,248 individuals who volunteered to fill in thequestionnaire at an Internet site (I5) and the curve marked with tri-angles is an aggregate of two nationally representative (R3þR4)samples.

Table 2. A variance component analysis of the RSES Items and the total

scores of 26,611 individuals

Variance component analysis (%)

Correlation rItems of the RSES Error Sex Age Sample Interactions

(1) I feel I’m a person of

worth, at least on an equal

plane with others

84.77 1.02 3.53 10.15 0.53 0.07��

(2) I feel I have a number of

good qualities

90.27 0.24 0.70 7.15 1.64 �.01

(3) All in all, I am inclined to

feel that I am a failure (R)

92.73 0.08 2.90 4.08 0.21 0.05��

(4) I am able to do things as

well as most other people

88.12 0.02 0.79 9.86 1.21 0.07��

(5) I feel I do not have much

to be proud of (R)

95.39 1.61 1.42 0.97 0.61 0.04��

(6) I take a positive attitude

toward myself

98.74 0.00 0.00 0.50 0.76 �.03��

(7) On the whole, I am

satisfied with myself

98.87 0.00 0.50 0.00 0.63 �.09��

(8) I wish I could have more

respect for myself (R)

93.71 0.08 1.24 4.86 0.11 0.04��

(9) I certainly feel useless at

times (R)

99.22 0.00 0.38 0.35 0.05 0.02��

(10) At times I think that I am

no good at all (R)

86.30 0.05 4.52 8.17 0.96 0.09��

Total Score 88.85 0.00 3.00 7.74 0.41 0.05��

Note. The participants’ gender, age group, and the type of samples (nationally representative

R3þR4versus self-recruited Internet I5) are treated as three random variables. (R)¼Reversed

scored item; interactions¼ Sex�Age, Sex�Sample, Age�Sample, and Sex�Age� Sample;

r¼Pearson correlation coefficient between age and the RSES items and the total score;��p< .001.


Generally, self-esteem level of adolescents in the self-recruitedInternet sample (I5) roses gradually until it reaches a plateau inyoung adulthood and retaines this level (4%� 1.5% points) till theend of the age range (correlation with age was about zero). Thus,opposite to the multinational Internet sample (Robins et al., 2002),self-esteem is at its lowest level during childhood among the Estonianself-recruited Internet sample.

The mean values of global self-esteem scale are systematicallyhigher in the Estonian self-recruited Internet sample compared tonationally representative samples. Namely, the mean score of theRSES of the representative sample is systematically and statisticallysignificantly (p< .05) below the mean self-esteem level in the Estonianself-recruited Internet sample across the age groups from 13 to 59years (Cohen’s d¼ .18–.47). In total, the RSES mean in the Internetsample is statistically significantly (p< .0001) higher compared to thenationally representative samples, d¼ .42 (see Table 1).

Therefore, the results of this study indicate that Estonians with ahigher level of self-esteem are more likely to answer the self-esteemquestionnaire posted on the Internet than those who are randomlyor semi randomly selected from the general population. What couldbe a possible explanation for this discrepancy? One possibility isthat the Estonian self-recruited Internet sample mostly consisted ofindividuals who were better educated and positioned at the higherlevel of the social stratification than average, as it is known that suchindividuals have generally higher global self-esteem (see Twenge &Campbell, 2002). As the respondents of the Internet sample did notprovide information about their educational level, this hypothesiscannot be tested in this study. However, the results of ANCOVArevealed that in the adult representative sample (R3) more highlyeducated individuals tended to have a higher self-esteem after adjust-ment for gender, F(2, 646)¼ 5.14, p< .01. More specifically, parti-cipants with a university degree had statistically significantly(p< .0001) higher self-esteem measured by the RSES (M¼ 4.02,SD¼ .55) compared to adults who had completed only basic or sec-ondary level of education (M¼ 3.77, SD¼ .67 and M¼ 3.86,SD¼ .64, respectively). Interestingly, there was no significant differ-ence in mean scores of the RSES between individuals in the Internetsample and participants with a university degree in the representativeadult sample (d¼ .03, ns). To conclude, the results of the currentstudy confirm that Estonians with a higher educational level havehigher self-esteem than persons with less educational attainment.As the mean level of the RSES in the self-recruited Internet sampleis comparable to those who had a university degree in the nationally


representative sample, it is very likely that the Internet sample alsoconsisted mostly of highly educated participants.

Item Life Span Trajectories

The mean scores of the RSES are considerably higher than the meanvalue of the SISE for all of the samples. This indicates that indivi-duals are inclined to agree with the SISE statement ‘‘I have highself-esteem’’ more reluctantly than with most of the other 10 itemsincluded in the Estonian version of the RSES. Indeed, a moredetailed inspection shows that the individual RSES items have notonly different levels of endorsement but they also have quite differentlife span trajectories: some of the items are relatively stable with mini-mal variations in the mean levels from childhood to older age, themeans of other items increase across age groups, and, finally, thethird group of items reach the maximum value in early adulthoodand start to decline afterwards. The cross-sectional developmentaltrajectories across different age groups for the all 10 RSES itemsfor a representative (R3þR4) and self-recruited Internet sample(I5) are presented in Figure 2.

Next we investigated what amount of variance in respondents’ self-esteem scores could be attributed to differences in age, sex, and sam-pling procedures as well as to the interactions of these effects. Toanswer this question, a variance component analysis with minimumvariance quadratic unbiased estimation (MVQUE) was performedfor the self-esteem measure scores. MVQUE is a variation of therestricted maximum likelihood estimation technique and it was cho-sen because there is no weighting of the random effects, so an iterat-ive solution for estimating their variance components is not required.Like correlation, the basic goal of variance component estimation isto assess the population covariation between random factors and adependent variable. The advantage of MVQUE is that it is not lim-ited to linear relationships between random factors and dependentvariables.

In our first analysis, the RSES scores of 26,611 individuals aretreated as a dependent variable and the participants’ gender, agegroup, and the type of samples (nationally representative R3þR4versus self-recruited Internet I5) as three random variables. Theresults show that less than 11% of the total variance in the RSESscores is attributable to the random factors (i.e., age group, gender,and sample type), only 0.41% to the interaction between variousrandom factors, 7.74% to differences in the type of sample, and,finally, only 3% to age differences (Table 2). Interestingly, there


are two items (item 1, ‘‘I feel I’m a person of worth, at least on anequal plane with others’’ and item. 10, ‘‘At times I think that I amno good at all’’) that demonstrate age dependence exceeding the levelof 3%. The endorsement of these two items increases with age(Table 2, last column). Moreover, these two items also demonstratethe largest effect of the sample, 10% and 8%, respectively, indicatingthat the items that are most sensitive to age are simultaneously mostvulnerable to the sampling bias. Five items, including both positivelyand negatively worded statements, show very modest (less than 1%)total variance caused by age.

Finally, we analyzed similarity of the life span trajectories by a cor-relational analysis. The cross-correlations between the item trajec-tories of the representative (R3þR4) and self-recruited Internetsample (I5) are shown in Table 3. The Pearson correlation evaluatesonly similarity in shape being insensitive to the mean level and scat-ter. On the basis of this, it is possible to claim that the first five RSES

Figure 2. The mean values of the RSES items across different age groups

for the representative (left panel) and self-recruited Internet (right panel)samples. Item numbers correspond to those in the original RSES (Rosenberg,

1965) and in the Estonian version of the RSES (Pullmann & Allik, 2000); see

Table 2.


items in the Internet sample behave differently from the respectiveitems in the representative sample (the main diagonal in Table 3).Only trajectories of three items, items 6 (‘‘I take a positive attitudetoward myself’’), 9 (‘‘I certainly feel useless at times’’), and 10 (‘‘Attimes I think that I am no good at all’’), are sufficiently similar inthe both samples. If one observes the trajectory of the item 6, self-esteem would appear to increase gradually from age 12 to 90 throughthe whole life span. This is a replicable finding: the correlation acrossthe nine age groups for random versus Internet samples is r¼ .96

Table 3. Correlations between the life span trajectories of the 10 RSES items

in the Internet (I5) and representative (R3 1 R4) samples

Self-recruited internet

sample (I5)

Representative samples (R3 1 R4)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

(1) I feel I’m a person of

worth, at least on

an equal plane with

others

�.07 �.09 �.12 .47 .00 .23 0.82 .04 .58 .56

(2) I feel I have a

number of good

qualities

�.10 .27 �.24 �.15 �.00 �.24 �.48 .11 .09 2.79

(3) All in all, I am

inclined to feel that

I am a failure (R)

.46 �.55 .07 �.01 .58 2.79 .50 .81 .28 �.54

(4) I am able to do

things as well as

most other people

�.07 .25 �.49 �.11 .18 �.33 .06 .29 .28 2.92

(5) I feel I do not have

much to be proud

of (R)

.06 �.10 �.39 .10 .29 �.43 .56 .52 .54 �.63

(6) I take a positive

attitude toward

myself

2.82 .83 �.65 .33 �.64 .96 .09 2.81 .39 .39

(7) On the whole, I am

satisfied with myself

�.43 .50 2.77 .23 .00 .18 .64 �.01 .72 �.35

(8) I wish I could have

more respect for

myself (R)

2.80 .76 2.77 .36 �.48 .80 .39 �.61 .59 .23

(9) I certainly feel

useless at times (R)

�.60 .63 2.79 .33 �.21 .31 .27 �.20 .71 �.42

(10) At times I think

that I am no good

at all (R)

�.45 .52 �.09 .37 �.45 .84 �.05 2.69 .21 .79

Note. Significant correlations p< .05 are shown in bold. The correlation between identical

items (the main diagonal) is underlined.


(p< .001). But if one chooses the item 9, the conclusion may be thatself-esteem grows in the first half of life, afterwards at age 30 starts todecline, and again to increase after age 50 (except for a drop after age70). Similarly, the latter pattern is also replicable across sample types,r¼ .71 (p< .05). However, these two patterns of age trends in generalself-esteem across the life span are unrelated, r¼ .31 (p¼ .41) andr¼ .39 (p¼ .29), respectively. This is a demonstration that singleitems may provide replicable findings in large samples, but theseare not necessarily generalizable to other, ostensibly similar items.

DISCUSSION

Taken together, this research makes three essential contributions to thedebate whether there are systematic normative age differences in globalself-esteem. Firstly, we presented some results that challenge generaliza-tion of the recently reported self-esteem trajectory (Robins et al., 2002;Robins & Trzesniewski, 2005; Trzesniewski, Robins, Roberts, & Caspi,2004) across different cultures. Second, we provided evidence that theself-recruited Internet data collection method is essentially biasedcompared to a random sampling from the general population. Finally,we found that different items measuring global self-esteem have dissimi-lar trajectories across the life span and therefore, the single-itemmeasures, like the SISE (Robins, Hendin, et al., 2001), may not begeneralizable across samples and comparable in similar way concerningquestionnaires consisting of several self-esteem items.

Does Global Self-Esteem Have a Universal Trajectoryacross the Life Span?

Surprisingly, none of the four normative self-esteem trajectories inthis study resembled any other trajectory portrayed in Figure 1. Cur-iously, the most dissimilar were trajectories of two large Internet sam-ples, the multinational English-speaking Internet sample (Robinset al., 2002) and the Estonian Internet sample (I5; this study). Thereare many possible reasons for this discrepancy. The first possibility,of course, is the difference between instruments used for measuringof global self-esteem. The multinational study was based on a singleanswer to the SISE item, whereas Estonian participants responded tothe RSES containing 10 items. As we demonstrated, the normativetrajectories of individual items through age were remarkably differentindicating that the wording of self-esteem items is sensitive to age.This suggests that different age groups have a tendency to responddifferently to the RSES items. For example, older respondents agreed


less frequently with the statement about feeling useless from time totime (item 9) compared to younger participants. At the same time,some other items (e.g., item 7: ‘‘On the whole, I am satisfied withmyself’’) did not exhibit remarkable differences between differentage groups. There have been several previous attempts to differen-tiate various facets inside global self-esteem. For instance, it wasproposed that global self-esteem consists of two subfactors, positiveand negative self-esteem, whereas positively and negatively wordeditems in the RSES have a tendency to group into two separate factors(e.g., Kohn & Schooler, 1969; Owens, 1994). However, the divisionof the self-esteem trajectories into distinct types does not correspondto the distinction between positive and negative self-esteem. Anyhow,the results of this study support the opinion proposed by severalresearchers (e.g., Carmines & Zeller, 1979; Corwyn, 2000; Dunbar,Ford, Hunt, & Der, 2000; Greenberger, Chen, Dmitrieva, & Farruggia,2003; Horan, DiStefano, & Motl, 2003; Marsh, 1996; Tomas &Oliver, 1999) that the wording of items has a systematic and endur-ing effect on the measurement of global self-esteem.

In addition to the differences in measurement instruments (i.e.,SISE versus RSES), the discrepancy between the results of the multi-national and Estonian Internet samples could be due to culturaland=or language differences, the pinpointing and explanation ofwhich is beyond the scope of this article. Furthermore, the life spantrajectories of two Estonian samples, R1� and R2�, randomly drawnfrom the National Census were rather dissimilar, although the ident-ical single-item measure was used. This may mean that there is nouniversal cross-sectional self-esteem trajectory through the age rangeand the observed differences between different age groups remainwell within the limits of the overall measurement accuracy. Thus,the results of this study do not support the idea of a single andinvariable trajectory of global self-esteem across the human life span.If these data support any trajectory, it is a flat one with some randomperturbations from the mean level.

How Typical are Self-Recruited Internet Samples?

Gosling and his colleagues (2004) argued that accumulating evidenceindicates on the consistency between the Internet and traditionalpencil-and-paper methods of gathering psychological data. Previousstudies have shown, for instance, that the administration mode,traditional or computer-based, had little effect on the psychometricproperties of the RSES (Vispoel, Boo, & Bleiler, 2001). Thisstudy supported this notion; the internal reliability of the self-esteem


measure was virtually identical in the Internet and traditionalpencil-and-paper samples. Also, the congruence between factor struc-tures of the RSES in these two samples was very high, consistent withnear-perfect equivalence.

Nevertheless, the mean values of global self-esteem measures weresystematically different in the self-recruited Internet sample andother samples representing more adequately the entire Estonian popu-lation. Most revealing was the contrast between the Estonian self-recruited Internet (I5) and other nationally representative (R3þR4)samples. This finding demonstrated that those participants who, ontheir own initiative, discovered the questionnaire on the Internetand volunteered to answer scored generally higher on the RSES scalethan those who were randomly selected from the National Census.

Why do Internet samples score systematically higher on self-esteemquestionnaires compared to nationally representative samples? In2000, when the Estonian Internet study was carried out, about a thirdof the Estonian population had an access to the Internet (comparedwith 52% in the United States and 23% on average in EuropeanUnion weighed by population; Bauer, Berne, & Maitland, 2002).However, just 4 years later, already 52% of 16- to 74-year-oldEstonians were using the Internet (TNS Emor, Spring 2004). A reg-ular survey conducted by the TNS Emor revealed that, comparedwith the entire population, Estonian Internet users in 2000 had abetter education (15 to 16 study years on average) and contained adisproportional number of managers, top officials, and executives(about a half of all Internet users).

This fact could explain the difference between the general popu-lation and Internet users because it is known that the individuals withhigher income and better education have higher levels of self-esteemthan persons with lower income and less educational attainment(McMullin & Cairney, 2004; Twenge & Campbell, 2002). Generally,both education and income have been found to be highly correlatedwith Internet use. Pettit (1999) examined whether respondents of anonline version of a psychological questionnaire were an unusual setof volunteers and found that 90% of the respondents indicated thattheir standard of living was higher compared with that of others.Also, in terms of the highest level of education obtained, theserespondents tended to be fairly well educated—almost 17% of thevolunteers were postgraduates, more than 37% had obtained bache-lors degrees, more than 18% had obtained community collegeeducation, and more than 25% had been educated at highschool. Similarly, according to a report published by the U.S.National Telecommunications and Information Administration in


1999, although 61.6% of those with a college degree used the Inter-net, only 6.6% of those with elementary school education or less usedthe Internet (National Telecommunications and Information Admin-istration, 1999). However, more research is needed on this topic aswell as on the relationship between self-esteem and educational levelthroughout the life course.

The analyses of this study revealed that the self-esteem level ofparticipants in the representative samples was similar to that of indi-viduals, who on their own initiative had volunteered to answer theself-esteem questionnaire over the Internet. Therefore, excluding per-haps the youngest and the oldest age groups, we can conclude thatthe Estonian self-recruited Internet sample mostly consisted of indivi-duals who were better educated and positioned at higher level ofsocial stratification, which may account for their generally higherlevel of global self-esteem. The analysis provided an indication thatthis might very likely be the case: there were no significant differencesbetween the two groups when the mean scores of the Internet samplewere compared with the mean scores of those in the representativesample who had a university degree. Nevertheless, as educationallevel of self-recruited Internet respondents in the Internet samplewas not provided, it was impossible to substantiate this hypothesisunconditionally in the current study.

Normative Stability of Self-Esteem

Research on general self-esteem development throughout human lifehas produced inconsistent results. For example, some studies havefound a mean-level rise of self-esteem during adolescence whereasother studies registered no change or even a drop in self-esteem (cf.O’Malley & Bachman, 1983). The reason for these discrepanciesmay be in nonequivalent measuring instruments or variation in thetested samples (Robins et al., 2002; for cohort effects see Twenge &Campbell, 2001). This study, however, points in the direction ofanother interpretation, according to which the main cause of incon-sistent results is the lack of a strong and systematic relationshipbetween age and global self-esteem. More specifically, possibleage effects are too small to be observed, even in large samples.Although individuals may have different developmental self-esteemtrajectories—besides those who are consistent over time, there areothers whose perception of self-worth is decreasing or increasing overtime (cf. Zimmerman, Copeland, Shope, & Dielman, 1997)—theopposite tendencies seem to cancel each other out when the meanscores are studied on a cross-sectional level. It appears that the


magnitude of the age-related changes is considerably smaller thanvariations in the mean level of self-esteem produced by sampling oreven small changes in the wording of items. For example, the greatestdifference between the means of two age groups studied by Robinsand his colleagues (2002) was 0.41 scale points, which is about 9%of the whole range of the 5-point SISE measure. Due to the fact thatthe standard deviation of the whole sample in the mentioned studywas 1.31 scale units, the maximal age-related change they were ableto observe was approximately one third of the standard deviation.The results of this study suggested that belonging to a certain agegroup accounted only for 3% of the variance in the mean levels ofthe RSES total scores while the influence of sampling (self-recruitedversus random sampling) had a much greater (7.74%) effect on themean values of the self-esteem.

Most theories of self-esteem suggest that feeling of self-worth ori-ginates either from life experiences or interpersonal relationships. Forexample, it was assumed that self-esteem is related to achievementand mastery experience. In turn, failures in personal life contributenegatively to the self-esteem (Rosenberg, 1979). According to thesociometer theory, however, people have a fundamental motive tomaintain connectedness with other people and for that purpose theyroutinely monitor how much they have been accepted or rejectedby others (Baumeister & Leary, 1995; Leary, Haupt, Strausser, &Chokel, 1998). People’s feelings about themselves reflect, in a sub-stantial extent, how they believe that they are perceived and evaluatedby others (cf. Shrauger & Schoeneman, 1979). All these approachesassume that the explanation of individual variation of self-esteemcan be found in environmental factors—socialization, life experience,or culture. Contrary to this perspective, another line of findings indi-cates that a substantial part of self-esteem can be understood as anenduring tendency to feel and think about oneself. From one thirdto one half of the variance of this tendency is controlled by genes(Kendler, et al., 1998; Roy, et al., 1995). Like most personality traits,shared environmental influences on self-esteem are small or negligible(Neiss, et al., 2002). The fact that self-esteem does not arise solelyfrom environmental circumstances is also supported by observationsthat rank-order stability is similar to other personality traits(Trzesniewki et al., 2004). Most therapeutic interventions or schoolprograms aiming to boost self-esteem have demonstrated a resistanceto change and have produced only temporary or limited benefits(Baumeister, Campbell, Krueger, & Vohs, 2003). Although cross-cultural psychologists have argued that the way how selfhood isconstructed differs fundamentally across regions of the world, a


recent large-scale comparative study has been more favorable for thetheoretical position holding that self-esteem is a universal phenom-enon that most likely stems from common human motivations: theinternal reliability and factor structure of the RSES was generalizableacross 53 nations and its scores correlated in an identical way with thekey personality traits of neuroticism and extraversion (Schmitt &Allik, 2005).

Global self-esteem appears to be a relatively enduring individualdifference that is strongly associated with core personality dimen-sions. Many studies have shown that scores on the RSES correlatesignificantly with two of the Big Five dimensions, Neuroticism andExtraversion (Costa et al., 1991; Judge et al., 2002; Kwan et al.,1997; Pullmann & Allik, 2000; Robins, Tracy, et al., 2001). The cor-relation between self-esteem and emotional stability is comparable tothe average relationship between alternative measures of neuroticism.For instance, according to a recent meta-analysis of the publishedarticles in 10 leading journals, the average correlation betweenself-esteem and Neuroticism was �.64 (Judge et al., 2002). In a largecomparative study involving 53 nations, the mean uncorrected corre-lations of the RSES scores with the Neuroticism scale of the Big FiveInventory (Benet-Martinez & John, 1998) was �.41 across 53 studiednations (Schmitt & Allik, 2005). For a comparison, correlationsbetween six facets of the NEO-PI-R that measure Neuroticism rangefrom �.31 to �.64, with the mean value of �.48 (Costa & McCrae,1992, Appendix F). A particularly strong negative correlation withthe Neuroticism score suggests that the main function of positiveself-evaluation is to buffer and protect its holder from frustrationand anxiety (Pyszczynski, Greenberg, Solomon, Arndt, & Schimel,2004; Rosenberg, Schooler, Schonbach, & Rosenberg, 1995). Theseobservations seem to indicate that the uniqueness of self-esteem mea-sures is overwhelmed by its commonality with other measures ofemotional stability (Judge et al., 2002). Thus, measures of globalself-esteem seem to represent reasonably well a general neuroticismfactor as it is conceptualized in the Big Five with perhaps somecontamination from extraversion.

Placed into a broader context of personality traits, the relative stab-ility of global self-esteem across the life span is less surprising. TheFive Factor Theory (McCrae & Costa, 1996, 1999) was proposed inorder to explain an extraordinary stability of personality traits in time(McCrae & Costa, 2003), a powerful effect of genes and vanishinglysmall effect of the shared environment on personality traits (Plomin& Caspi, 1999), and their transcendence across boundaries of culture(Allik & McCrae, 2002). Based on cross-sectional and longitudinal


studies, Costa and McCrae (2002) proposed that on the backgroundof a general temporal stability of personality traits, there are still spe-cific age curves for each personality trait. Among all five personalitydimensions, neuroticism is the most stable across the life span: in somecountries, such as the United States (Costa & McCrae, 2002; McCraeet al., 2004), the mean level of self-reported neuroticism decreases withage and in some other countries, such as Estonia and Italy, it remainsbasically the same across the whole life span (Allik, Laidra, Realo, &Pullmann, 2004; Costa, McCrae, Martin et al., 2000) or have a curvi-linear trend (McCrae et al., 2004). If the main function of positive self-evaluation is to buffer and protect its holder from frustration andanxiety, then it is expected that self-esteem, like other componentsof emotional stability, is also relatively stable across the life span. Thisprovides a support to a standard view, according to which normativeage differences in global self-esteem are small and most likely beyondour ability to measure them.

Finally, the results of this study warn about limitations of the self-recruited Internet studies. Firstly, such convenience-type studiesmight be better suited for testing theoretical ideas concerningrelationship between various concepts rather than providing pointestimates of parameters that are intended to apply to a population.Secondly, biases related to Internet samples seem to affect meanlevels more than patterns of covariation (i.e., mean levels are moreaffected than, say, factor structures).

REFERENCES

Allik, J., Laidra, K., Realo, A., & Pullmann, H. (2004). Personality development

from 12 to 18 years of age: Changes in mean levels of traits and in their structure.

European Journal of Personality, 18, 445–462.

Allik, J. & McCrae, R. R. (2002). A Five-Factor Theory perspective. In R. R.

McCrae & J. Allik (Eds.), The Five-Factor Model across cultures (pp. 303–321).

New York: Kluwer Academic=Plenum Publishers.

Barbeite, F. G. & Weiss, E. M. (2004). Computer self-efficacy and anxiety scales for

an Internet sample: Testing measurement equivalence of existing measures and

development of new scales. Computers in Human Behavior, 20, 1–15.

Bauer, J. M., Berne, M., & Maitland, C. F. (2002). Internet access in the European

Union and in the United States. Telematics and Informatics, 19, 117–137.

Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high

self-esteem cause better performance, interpersonal success, happiness, or heal-

thier lifestyles? Psychological Science in the Public Interest, 4, 1–44.

Baumeister, R. F. & Leary, M. R. (1995). The need to belong: Desire for interperso-

nal attachments as a fundamental human motivation. Psychological Bulletin, 117,

497–529.


Benet-Martinez, V. & John, O. P. (1998). Los cinco grandes across cultures and

ethnic groups: Multitrait multimethod analysis of the Big Five in Spanish and

English. Journal of Personality and Social Psychology, 75, 729–750.

Birnbaum, M. H. (2004a). Methodological and ethical issues in conducting social

psychology research via the Internet. In C. Sansone, C. C. Morf, & A. T. Panter

(Eds.), Handbook of Methods in Social Psychology (pp. 359–382). Thousand Oaks,

CA: Sage.

Birnbaum, M. H. (2004b). Human research and data collection via the Internet.

Annual Review of Psychology, 55, 803–832.

Buchanan, T. & Smith, J. L. (1999). Using the Internet for psychological research:

Personality testing on the World Wide Web. British Journal of Psychology, 90,

125–144.

Carmines, E. G. & Zeller, R. A. (1979). Reliability and validity assessment. Beverly

Hills, CA: Sage.

Corwyn, R. F. (2000). The factor structure of global self-esteem among adolescents

and adults. Journal of Research in Personality, 34, 357–379.

Costa, P. T., Jr. & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO

PI-R) and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa,

FL: Psychological Assessment Resources.

Costa, P. T., Jr. & McCrae, R. R. (2002). Looking backward: Changes in the mean

levels of personality traits from 80 to 12. In D. Cervone & W. Mischel(Eds.),

Advances in personality science (pp. 219–237). New York: Guilford.

Costa, P. T., Jr., McCrae, R. R., & Dye, D. A. (1991). Facet scales for agreeableness

and conscientiousness: A revision of the NEO Personality Inventory. Personality

and Individual Differences, 12, 887–898.

Costa, P. T., Jr., McCrae, R. R., Martin, T. A., Oryol, V. E., Senin, I. G., Rukavishnikov,

A. A., et al. (2000). Personality development from adolescence through adulthood:

Further cross-cultural comparisons of age differences. In V. J. Molfese & D. Molfese

(Eds.), Temperament and personality development across the life span (pp. 235–252).

Hillsdale, NJ: Erlbaum.

Crawford, S. D., Couper, M. P., & Lamias, M. J. (2001). Web surveys: Perceptions of

burden. Social Science Computer Review, 19, 146–162.

Demo, D. H. (1992). The self-concept over time: Research issues and directions.

Annual Reviews in Sociology, 18, 303–326.

Dunbar, M., Ford, G., Hunt, K., & Der, G. (2000). Question of wording effects in

the assessment of global self-esteem. European Journal of Psychological Assess-

ment, 16, 13–19.

Gosling, S. D., Vazire, S., Srivastava, S., & John, O. P. (2004). Should we trust

Web-based studies? A comparative analysis of six preconceptions about Internet

questionnaires. American Psychologist, 59, 93–104.

Greenberger, E., Chen, C., Dmitrieva, J., & Farruggia, S. (2003). Item-wording and

the dimensionality of the Rosenberg Self-Esteem Scale: Do they matter? Person-

ality and Individual Differences, 35, 1241–1254.

Horan, P. M., DiStefano, C., & Motl, R. W. (2003). Wording effect in self-esteem

scales: Methodological artifact or response style? Structural Equation Modeling,

10, 435–455.


Judge, T. A., Erez, A., Bono, J. E., & Thoresen, C. J. (2002). Are measures of self-

esteem, neuroticism, locus of control, and generalized self-efficacy indicators of a

common core construct? Journal of Personality and Social Psychology, 83, 693–710.

Kendler, K. S., Gardner, C. O., & Prescott, C. A. (1998). A population-based twin

study of self-esteem and gender. Psychological Medicine, 28, 1403–1409.

Kohn, M. L. & Schooler, C. (1969). Class, occupation, and orientation. American

Sociological Review, 34, 659–678.

Kwan, V. S., Bond, M. H., & Singelis, T. M. (1997). Pancultural explanations for life

satisfaction: Adding relationship harmony to self-esteem. Journal of Personality

and Social Psychology, 73, 1038–1051.

Leary, M. R., Haupt, A. L., Straussedr, K. S., & Chokel, J. T. (1998). Calibrating

sociometer: The relationship between interpersonal appraisals and state self-

esteem. Journal of Personality and Social Psychology, 74, 1290–1299.

Marsh, H. (1996). Positive and negative global self-esteem: Substantively meaningful

distinction or artifacts. Journal of Personality and Social Psychology, 70, 810–819.

McCarthy, J. D. & Hoge, D. R. (1982). Analysis of age effects in longitudinal studies

of adolescent self-esteem. Developmental Psychology, 18, 372–379.

McCrae, R. R. & Costa, P. T., Jr. (1996). Toward a new generation of personality

theories: Theoretical contexts for the Five-Factor Model. In J. S. Wiggins

(Ed.), The Five-Factor Model of personality: Theoretical perspectives (pp. 51–87).

New York: Guilford.

McCrae, R. R. & Costa, P. T., Jr. (1999). A Five-Factor Theory of personality. In

L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research

(2nd ed., pp. 139–153). New York: Guilford.

McCrae, R. R. & Costa, P. T., Jr. (2003) Personality in adulthood: A Five-Factor

Theory perspective (2nd ed.). New York: Guilford.

McCrae, R. R., Costa, P. T., Jr., Hrebıckov�aa, M., Urb�aanek, T., Martin, T. A., Oryol,

V. E., et al. (2004). Age differences in personality traits across cultures: Self-

report and observer perspective. European Journal of Personality, 18, 143–157.

McCrae, R. R., Costa, P. T., Jr., Ostendorf, F., Angleitner, A., Hrebıckov�aa, M., Avia,

et al. (2000). Nature over nurture: Temperament, personality, and lifespan devel-

opment. Journal of Personality and Social Psychology, 78, 173–186.

McMullin, J. A. & Cairney, J. (2004). Self-esteem and the intersection of age, class,

and gender. Journal of Aging Studies, 18, 75–90.

National Telecommunications and Information Administration. (1999). Falling

through the net: Defining the digital divide. Washington, DC: United States Depart-

ment of Commerce.

Neiss, M. B., Sedikides, C., & Stevenson, J. (2002). Self-esteem: A behavioural

genetic perspective. European Journal of Personality, 16, 351–367.

O’Malley, P. M. & Bachman, J. G. (1983). Self-esteem: Change and stability between

ages 13 and 23. Developmental Psychology, 18, 327–379.

Owens, T. J. (1994). Two dimensions of self-esteem—Reciprocal effects of positive

self-worth and self-deprecation on adolescent problems. American Sociological

Review, 59, 391–407.

Pettit, F. A. (1999). Exploring the use of the World Wide Web as a psychology data

collection tool. Computers in Human Behavior, 15, 67–71.


Plomin, R. & Caspi, A. (1999). Behavioral genetics and personality. In L. A. Pervin

& O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed.,

pp. 251–276). New York: Guilford.

Pullmann, H. & Allik, J. (2000). The Rosenberg Self-Esteem Scale: Its dimension-

ality, stability, and personality correlates in Estonian. Personality and Individual

Differences, 28, 701–715.

Pyszczynski, T., Greenberg, J., Solomon, S., Arndt, J., & Schimel, J. (2004). Why do

people need self-esteem? A theoretical and empirical review. Psychological Bull-

etin, 130, 435–468.

Reips, U.-D. (2000). The Web experiment method: Advantages, disadvantages, and

solutions. In M. H. Birnbaum (Ed.), Psychological experiments on the Internet

(pp. 89–117). San Diego: Academic Press.

Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global

self-esteem: Construct validation of a single item measure and the Rosenberg

Self-Esteem Scale. Personality and Social Psychology Bulletin, 27, 151–161.

Robins, R. W., Norem, J. K., & Cheek, J. M. (1999). Naturalizing the self. In L. A.

Pervin and O. P. John (Eds.). Handbook of personality: Theory and research (2nd

ed., pp. 443–477). New York: Guilford.

Robins, R. W., Tracy, J. L., Trzesniewski, K. H., Gosling, S. D., & Potter, J. (2001).

Personality correlates of self-esteem. Journal of Research in Personality, 35, 463–482.

Robins, R. W. & Trzesniewski, K. H. (2005). Self-esteem development across the

lifespan. Current Directions in Psychological Science, 14, 158–162.

Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J. (2002).

Global self-esteem across the life span. Psychology and Aging, 17, 423–434.

Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level

change in personality traits across the life course: A meta-analysis of longitudinal

studies. Psychological Bulletin, 132, 1–25.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton

University Press.

Rosenberg, M. (1979). Conceiving the self. New York: Basic Books.

Rosenberg, M., Schooler, C., Schoenbach, C., & Rosenberg, F. (1995). Global

self-esteem and specific self-esteem: Different concepts, different outcomes.

American Sociological Review, 60, 141–156.

Roy, M. A., Neale, M. C., & Kendler, K. S. (1995). The genetic epidemiology of

self-esteem. British Journal of Psychiatry, 166, 813–820.

Schmitt, D. P. & Allik, J. (2005). Simultaneous administration of the Rosenberg

Self-Esteem Scale in 53 nations: Exploring the universal and culture-specific

features of global self-esteem. Journal of Personality and Social Psychology, 89,

623–642.

Shrauger, J. S. & Schoneman, T. J. (1979). Symbolic interactionist view of the

self-concept: Through the looking glass darkly. Psychological Bulletin, 86,

549–573.

Smith, T. W. (2001, October). Are representative Internet surveys possible? Proceedings

of Statistics Canada Methodology Symposium ‘‘Achieving Data Quality in a

Statistical Agency: A Methodological Perspective.’’ Ocotber 17–19, 2001,

Quebec, Canada.


Statistical Office of Estonia. (2000). Statistical Yearbook of Estonia. Tallinn.

Tomas, J. M. & Oliver, A. (1999). Rosenberg’s Self-Esteem Scale: Two factors or

method effects. Structural Quation Modeling, 6, 84–98.

Trzesniewski, K. H., Robins, R. W., Roberts, B. W., & Caspi, A. (2004). Personality

and self-esteem development across the lifespan. In P. T. Costa & I. C. Siegler

(Eds.), Recent advances in psychology and aging (pp. 163–185). Amsterdam, The

Netherlands: Elsevier Science.

Twenge, J. M. & Campbell, W. K. (2001). Age and birth cohort differences in self-

esteem: A cross-temporal meta-analysis. Personality and Social Psychology Review,

5, 321–344.

Twenge, J. M. & Campbell, W. K. (2002). Self-esteem and socioeconomic status: A

meta-analytic review. Personality and Social Psychology Review, 6, 59–71.

Vispoel, W. P., Boo, J., & Bleiler, T. (2001). Computerized and paper-and-pencil

versions of the Rosenberg Self-Esteem Scale: A comparison of psychometric

features and respondent preferences. Educational and Psychological Measurement,

61, 461–474.

Walsh, J. P., Kiesler, S., Sproull, L. S., & Hesse, B. W. (1992). Self-selected and

randomly selected respondents in a computer network survey. Public Opinion

Quarterly, 56, 241–244.

Wylie, R. C. (1979). The self-concept. Lincoln, NE: University of Nebraska Press.

Zimmerman, M. A., Copeland, L. A., Shope, J. T., & Dielman, T. E. (1997). A longi-

tudinal study of self-esteem: Implications for adolescent development. Journal of

Youth and Adolescence, 26, 117–141.


GLOBAL SELF-ESTEEM ACROSS THE LIFE SPAN: A CROSS...

Documents

Transcript of GLOBAL SELF-ESTEEM ACROSS THE LIFE SPAN: A CROSS...