Page 90 - Ana Kozina and Nora Wiium, eds. ▪︎ Positive Youth Development in Contexts. Ljubljana: Educational Research Institute, 2021. Digital Library, Dissertationes (Scientific Monographs), 42.
P. 90
positive youth development in contexts
other adults and neighbours. Almost all non-target loadings were poor. For
Empowerment, almost all target loadings were almost fair (all exceeded .38;
ps < .001), except for items that relate to self-perception and family deci-
sions. Almost all non-target loadings were poor, except for two items that
refer to school rules. The most troublesome of the DA latent constructs are
Boundaries and expectations and Constructive use of time. Only two tar-
get loadings of items of Boundaries and expectations were almost fair (ps <
.001). The other 6 Boundaries and expectations items’ target loadings were
poor. Non-target loadings varied between .01 and .82.
As for Constructive use of time, all of the target loadings were poor
(below .21), indicating that both Constructive use of time and Boundaries
and expectations are problematic constructs and should be probed into.
For Commitment to learning, target loadings for 4 out of 7 items were high-
er than .36 (ps < .001). Almost all non-target loadings were poor. Only 6
out of 13 target loadings of Positive values were fair (ps < .001), indicating
Positive values may be a problematic construct. Almost all non-target load-
ings were poor. Moreover, Social competencies appear to be questionable as
well since only two target loadings were almost fair (higher than .42; ps <
.001). Almost all non-target factor loadings were poor. As for Positive iden-
tity, almost all target loadings were higher than .52 (ps < .001) while only
one item’s target loading was poor (p < .05). All non-target loadings were
poor.
The following correlated errors were included in the model. They were
based on modification indices when they could be justified by the con-
tent: item 15 with 21 for Boundaries and expectations (both relate to school
rules), item 35 with 36 for Positive values (both refer to being responsible)
and item 56 with 57 for Positive identity (both are connected with strate-
gies to deal with difficult situations). We examined a second ESEM model,
which contained factors that refer to different contexts (i.e. personal (self),
social, family, school, community). Even after including 12 modification in-
dices (as justified by the content), the model did not show an adequate fit:
χ2(1579) = 6093.76, p <. 001, CFI = .905, RMSEA = .038, 90% CI [.037, .039],
SRMR = .031. Target loadings for the Personal factor varied from -.10 to .68
(ps < .001) while non-target loadings were poor. Target loadings for Social
assets were poor (ranging from .15 and .46; ps < .001) and non-target load-
ings were also poor. As for Family assets, target loadings were from .24
to .61 (ps < .001) while non-target loadings were poor. Target loadings for
School assets ranged from .07 to .67 (ps < .001) whereas non-target loadings
90
other adults and neighbours. Almost all non-target loadings were poor. For
Empowerment, almost all target loadings were almost fair (all exceeded .38;
ps < .001), except for items that relate to self-perception and family deci-
sions. Almost all non-target loadings were poor, except for two items that
refer to school rules. The most troublesome of the DA latent constructs are
Boundaries and expectations and Constructive use of time. Only two tar-
get loadings of items of Boundaries and expectations were almost fair (ps <
.001). The other 6 Boundaries and expectations items’ target loadings were
poor. Non-target loadings varied between .01 and .82.
As for Constructive use of time, all of the target loadings were poor
(below .21), indicating that both Constructive use of time and Boundaries
and expectations are problematic constructs and should be probed into.
For Commitment to learning, target loadings for 4 out of 7 items were high-
er than .36 (ps < .001). Almost all non-target loadings were poor. Only 6
out of 13 target loadings of Positive values were fair (ps < .001), indicating
Positive values may be a problematic construct. Almost all non-target load-
ings were poor. Moreover, Social competencies appear to be questionable as
well since only two target loadings were almost fair (higher than .42; ps <
.001). Almost all non-target factor loadings were poor. As for Positive iden-
tity, almost all target loadings were higher than .52 (ps < .001) while only
one item’s target loading was poor (p < .05). All non-target loadings were
poor.
The following correlated errors were included in the model. They were
based on modification indices when they could be justified by the con-
tent: item 15 with 21 for Boundaries and expectations (both relate to school
rules), item 35 with 36 for Positive values (both refer to being responsible)
and item 56 with 57 for Positive identity (both are connected with strate-
gies to deal with difficult situations). We examined a second ESEM model,
which contained factors that refer to different contexts (i.e. personal (self),
social, family, school, community). Even after including 12 modification in-
dices (as justified by the content), the model did not show an adequate fit:
χ2(1579) = 6093.76, p <. 001, CFI = .905, RMSEA = .038, 90% CI [.037, .039],
SRMR = .031. Target loadings for the Personal factor varied from -.10 to .68
(ps < .001) while non-target loadings were poor. Target loadings for Social
assets were poor (ranging from .15 and .46; ps < .001) and non-target load-
ings were also poor. As for Family assets, target loadings were from .24
to .61 (ps < .001) while non-target loadings were poor. Target loadings for
School assets ranged from .07 to .67 (ps < .001) whereas non-target loadings
90
