Page 4 - 2011-ciani

56 A. C. Ciani and C. Capiluppi

Figure 1. Sample classification. Mainlanders: subjects resident and born in the Trapani area, on mainland Sicily, with no known relatives coming from the Egadi
archipelago. Islanders: subjects resident in the Egadi archipelago with at least one grandparent native to the islands; most of them were born in the islands (Native),
but some of them were born incidentally on the mainland (Not native). Ancient Origin Islanders: a subset of Islanders subjects with all four grandparents native to the
islands; according to the historical accounts most of them descend directly from the founders of the island population. Mixed Origin Islanders: a second subset of
Islanders with 1–3 grandparents native to the islands. Immigrants: subjects resident in the Egadi archipelago without grandparents native to the islands; they include
all 1st, 2nd and 3rd generation immigrants to the archipelago, with origins in various parts of Italy; they are further classified into four subsets. Adapted Immigrants:
immigrants resident for at least 20 years, who immigrated after 10 years of age; these individuals have shared the island environment with islanders for a long time
but do not have recent ancestry in common. Recent Immigrants: immigrants resident for less than 20 years who immigrated after 10 years of age; these individuals
share the least environment and genetic ancestry with islanders. Child Immigrants: subjects who immigrated before 10 years of age; these individuals have shared
the same environment with islanders since infancy; this group is dealt with separately because early experience could influence adult personality traits. Native
Immigrants: 2nd and 3rd generation immigrants (i.e. descendants of immigrants) born on the Egadi islands; they have shared with islanders the same environment
and early experience since birth but have no recent shared genetic ancestry. In order to investigate which mechanism could be at the origin of personality differences,
we further classified the Islanders into Emigrant and Sedentary, according to whether they had emigrated from the island during their lifetime. Emigrant: subjects
who emigrated permanently (not just for the purposes of studying), to mainland Italy or another country and started a family away from the island. Some of these
individuals were sampled on the nearby mainland (Trapani), others were sampled in the islands during a family visit or as they returned to the island, once retired, at a
late age. Sedentary: islanders who never left the island during their lifetimes, if not for the purposes of studying.

regression analysis (we found significant effects for sex and differences were not significant (Table 2). A further
age on conscientiousness and emotional stability, and for all comparison between mainlanders and the subset of native
the three variables on the other traits). To eliminate the effect islanders with ancient origin ancestry (n ¼ 380) shows a
of sex, age and education level on group comparisons, all the highly significant difference for extraversion ( p < .001),
statistical tests reported in the paper ( p-values summarized in openness to experience ( p < .001) (À9.8 and À6.6 corrected
Tables 2 and 3) were obtained by ANCOVA with sex, age and T-scores, respectively), and a significant difference for
education level as covariates1. Comparisons between groups conscientiousness ( p ¼ .042). While for emotional stability
with less than 20 subjects were not considered. For each the difference was on the critical level ( p ¼ .051) (Table 2).
significant comparison, we reported the T-score difference
estimated by the ANCOVA model to evaluate the effect size. Second question. Can these differences be ascribed to a
possible genetic origin, or are they due to individual
RESULTS flexible adaptation to the environment?

For each sub-sample we reported sex, age, education level, In Table 2 we show that immigrants (n ¼ 79) did not differ
age at the time of immigration and years lived on the islands from mainlanders (n ¼ 106) on any traits except emotional
by immigrants in Table 1. We reported, for each subgroup stability (þ2.9 T-scores, p ¼ .035). Immigrants significantly
considered in the analysis, the Big Five personality traits differed from original islanders (n ¼ 408) on extraversion
expressed in T-scores in Figure 2, to estimate the measured ( p < .001) and openness to experience ( p ¼ .001) (þ7.3
personality mean T-scores in each population subset. and þ4.1 T-scores, respectively). The subjects who
immigrated to the islands showed a personality profile that
First question. Are there differences in personality was undistinguishable from the mainland population, except
traits between islanders and mainlanders? for higher emotional stability; in contrast, immigrants were
significantly more extraverted and more open to experience
Islanders (n ¼ 528) differ from mainlanders (n ¼ 106) on than islanders (Figure 4).
extraversion and openness to experience after controlling for
the effect of the covariates (Figure 3): islanders were To better discriminate between genetic and environmen-
significantly less extraverted and less open to experience tal hypotheses, in Table 3 we show eight new comparisons of
(À8.3 and À6.0 corrected T-scores,2 respectively). The other relevant sample subsets. The first part of Table 3 tests the
effects predicted by the genetic hypothesis; the second part of
1The Kolmogorov–Smirnov and the Shapiro–Wilk normality tests showed the table tests differences predicted by the individual flexible
that the normality assumption held up for personality traits scores response to the environment hypothesis.
distribution.
2Effect size estimate: T-scores difference estimated by the ANCOVA model The first and second ‘genetic’ comparisons aim to detect
with sex, age and education as covariates. the effect of possible genotype differences between subsets
of immigrants and islanders who have shared the same
environment; the third comparison aims to detect a gradient

Copyright # 2010 John Wiley & Sons, Ltd. Eur. J. Pers. 25: 53–64 (2011)
DOI: 10.1002/per