You are on page 1of 12

Infant Behavior & Development 35 (2012) 71–82

Contents lists available at SciVerse ScienceDirect

Infant Behavior and Development

Temperamental profiles and language development: A replication and


an extension
Valentina Garello ∗ , Paola Viterbori, M. Carmen Usai
Department of Education Sciences, University of Genoa, Corso Podestà 2, 16128 Genova, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Individual differences in child temperament are associated with individual differences in
Received 28 April 2010 language development. The present study examined the relationship between tempera-
Received in revised form ment and language ability in 109 twenty-four- to 30-month-old children. Parents and
10 November 2010
day-care teachers completed two questionnaires: the Primo Vocabolario del Bambino (Caselli
Accepted 20 September 2011
& Casadio, 1995) and the Questionari Italiani del Temperamento (Axia, 2002). Researchers
administered the First Language Test (Axia, 1993) to assess productive and receptive lan-
Keywords:
Temperament
guage in each child.
Language development Replicating previous research (Usai, Garello, & Viterbori, 2009), day-care teachers iden-
Toddlers tified three temperamental profiles: most of the children fit into the first profile, typical
Individual differences of the Italian population; another profile was made up of easily distractible and not very
persistent children, with a poor capacity to modulate motor activity; and the third profile of
children were inhibited in new situations. A relationship was found between temperament
assessed by day-care teachers and different levels of linguistic competence. In particular,
the groups of “inattentive” and “inhibited” children showed poorer lexical and morpho-
logical abilities and a more immature vocabulary, characterised by the presence of more
primitive components of the lexical repertory compared to the group of “typical” chil-
dren. Unlike the results from day-care teachers, temperament questionnaires completed
by parents revealed a 4-cluster-solution. Also, for parents, the “typical” profile is charac-
terised by the largest vocabulary (productive and receptive) and the most mature semantic
production.
© 2011 Elsevier Inc. All rights reserved.

The present study is the replication and extension of a preceding research aimed to identify specific temperamental
profiles in a sample of 28-month-old children and to analyse the differences in language ability among the profiles found
(Usai, Garello, & Viterbori, 2009).
Temperament can be defined as constitutionally based individual differences in reactivity and self-regulation (Rothbart
& Bates, 1998). Several different temperament taxonomies are described (see e.g., Buss & Plomin, 1984; Chess &
Thomas, 1984; Gartstein & Rothbart, 2003; Goldsmith, 1996; Rothbart, 1981) with relatively independent dimensions.
An interesting line of inquiry was aimed to find particular temperament types, in the sense of groups of individuals
with very specific configurations of temperamental characteristics. In the literature, evidence has been found that cer-
tain combinations of temperamental dimensions could be much more likely in that they characterise more individual
children.
A good example for preschooler population is the Caspi and Silva (1995) study where five temperamental types in a
cohort of 3 years old Neozeland children were derived: 40% of children were characterised as age and situation-appropriate

∗ Corresponding author. Tel.: +39 010 20953706/726; fax: +39 010 20953728.
E-mail address: valentina.garello@unige.it (V. Garello).

0163-6383/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.infbeh.2011.09.003
72 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

and they were grouped in a “well-adjusted” category; 10% of them were rated as irritable, impulsive, emotionally labile,
and impersistent on tasks and they were classified in a “undercontrolled” group; 8% of them were shy, fearful, and socially
ill at ease and constituted a “inhibited” group; finally, a “reserved” type (15% of the sample) was characterised by chil-
dren apprehensive in the novel situation, and a “confident” type (28% of the sample) comprised children with an outgoing
approach.
More recently Van den Akker, Deković, Prinzie, and Asscher (2010), studying longitudinally a sample of children aged 30
months, found three temperamental profiles: a “typical” type that comprised the majority of the sample and was similar
to the well-adjusted profile identified by Caspi and Silva (1995); an “expressive” type characterized by the highest levels
of anger proneness and activity level, and intermediate levels of social fear; a “fearful” type characterized by the highest
levels of social fear, and intermediate levels of anger proneness and activity level. Other authors found similar temperament
types: Komsi et al. (2006) and Stifter, Putnam, and Jahromi (2008) identified respectively a “resilient” or “low reactive” type
similar to the “typical” type, an “undercontrolled” or “exuberant” type analogous to the “expressive” type, and, finally, an
“overcontrolled” or “inhibited” type correspondent to the “fearful” type. Otherwise, Aksan et al. (1999) were able to find two
temperament types: one high in control and low in both approach and negative affectivity, and the other low in control and
high in approach and negative affectivity.
In our previous study, exploring the relationship between temperament and language ability in a sample of 28-month-old
children, we identified three temperamental profiles: a profile typical of the Italian population that grouped most of the
children and was characterised by good social orientation, low inhibition to novelty, medium-high positive emotionality
and attention; a second profile made up of easily distractible and not very persistent children, with a poor capacity to
modulate motor activity; and a third profile with children who were inhibited in new situations. A comparison of the
three groups based on their language ability revealed important differences in language development. In particular, the
group of “inattentive” children was characterised by a more restricted expressive vocabulary, the lowest decontextualised
production and a more immature vocabulary composition compared to the other groups. Furthermore, “inattentive” children
showed higher percentages of risk indices for language acquisition, such as a restricted vocabulary size and an absence of
morphological ability (use of inflected forms of nouns, adjectives and verbs). The results confirmed a relationship between
temperamental and linguistic profiles and suggested that children with poor attention control may show greater delays in
language acquisition.
These results are in line with a growing body of literature that revealed links between specific temperamental dimensions
and both expressive and receptive vocabulary (e.g., Dixon & Shore, 1997; Dixon & Smith, 2000; Kubicek, Emde, & Schmitz,
2001; Morales et al., 2000; Mundy & Gomes, 1998; Slomkowski, Nelson, Dunn, & Plomin, 1992). For example, Bloom (1993)
found that children who learn to speak earlier show greater emotional stability compared to late talkers and suggested that
emotional stability may promote the attentive states necessary to learn the meaning of words. Furthermore, Slomkowski et al.
(1992) found that temperamental features, such as affect-extraversion (i.e., high interest in people, high cooperativeness,
low fearfulness, and high happiness), at 2 years of age were positively related to several language measures at 7 years
and, similarly, Dixon and Shore (1997) demonstrated that smiling-laughter and duration of orienting at 13 months were
associated with language style at 21 months.
In our previous study (Usai et al., 2009), we had children evaluated by day-care teachers. Teachers completed two
scales assessing the temperament and language development of their pre-schoolers. A major limit of this study was that
the measures were recorded by the same observer, which might have strengthened the associative link between cer-
tain temperamental dimensions and specific language skills. Teachers may consider children who display temperament
attributes such as high attentional control, adaptability and low activity more competent because these characteristics
match with the teacher’s idea of a model student (Keogh, 1989). Research has also suggested that, at least with older children,
teacher-perceived temperament affects the student–teacher relationship and influences both attitudes toward the student
and expectations of pupils’ abilities (Keogh, 1989, 1994; Martin, 1989; Moritz Rudasill & Rimm-Kaufman, 2009; Mullola
et al., 2010; Rothbart & Jones, 1998). For these reasons, a child temperament evaluation from parents may add important
information.
A second limitation of the previous study was the Italian version of the McArthur–Bates Communicative Development
Inventories language assessment. This scale was developed for child language evaluation by parents and little is known
about its use in the day care setting. Some recent data, however, suggest that teacher and parent reports of vocabulary size
and composition are not the same in children between 24 and 36 months. Generally, parents’ reports exceed teacher reports
in the number of words identified (Vagh, Pan, & Mancilla-Martinez, 2009).
A final limitation of the previous study is that it was conducted with a sample of children each evaluated at 28 months
of age, thereby reducing the generalizability of the results to other age levels.
Taking into account the previous limitations, the present study explores whether differences in the temperaments of
24–30-month-olds, as perceived by both teachers and mothers, are associated with children’s language skills evaluated by
both informants and a standardised test.
In particular, we expect to replicate the results of the previous study by identifying three temperamental profiles with
different language abilities. We expect the profiles characterised by low attention levels to be associated with poorer language
performance, both on parent and teacher reports and on standardised tests. Secondly, we seek to identify temperamental
profiles from mothers’ questionnaires and we expect to find the same association between temperamental profiles and
language skills.
V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82 73

1. Method

1.1. Participants

The mothers and day-care teachers of 109 children, aged 24–30 months (52 males and 57 females; M = 27.8, SD = 2.1),
participated in this study. The children were enrolled in 14 different day care centres. Children with health problems or
exposure to a language other than Italian were excluded from the study and not included in the final sample.
Parents and day-care teachers assessed the children, using two questionnaires evaluating language ability and temper-
ament. Language development was evaluated also using a standardised test. Eleven children were not available at the time
of this assessment because of prolonged illness or family transfer.
Parents were asked to indicate the mother’s highest level of academic achievement. Of the 99 mothers who provided this
information, 23% had middle school education, 52% had high school education and 25% had a university degree. Ten mothers
did not provide information about their academic level.

1.2. Procedure and design

Language and temperament were assessed using two questionnaires: Il Primo Vocabolario del Bambino “Words and Sen-
tences” (PVB, Caselli & Casadio, 1995), which is the Italian version of the MacArthur–Bates Communicative Development
Inventory (CDI, Fenson et al., 1993), and the Questionari italiani del temperamento (QUIT, Axia, 2002), a scale evaluating the
temperament characteristics of children. Parents and day-care teachers completed both questionnaires over a period of 10
days. Thirty-three day-care teachers took part in the study.
Although the PVB and QUIT questionnaires are designed to be self-explanatory, the day-care teachers attended a specific
training course on the correct use of these tools. Parents were given instructions about the questionnaires when inform
consent was given to them.
After the participants had completed the questionnaires, a group of independent testers, blind to children’s temperament
and PVB scores, administered the First Language Test (Axia, 1993) to each child to assess their productive and receptive
language.

1.3. Language measure

The “Words and Sentences” PVB version includes a 670-word vocabulary checklist, organised into 23 semantic categories.
This part of the PVB form is designed to assess vocabulary size and composition.
Vocabulary composition refers to the proportion of each semantic categories on the total. This information is important
because, while the overall vocabulary during the course of development grows constantly, the average proportion of indi-
vidual semantic categories does not remain constant. At an early stage, the lexicon is largely made up of nouns, then the
proportion of predicates (verbs and adjectives) grows steadily and finally an increase of the proportion of function words can
be observed. Concurrently onomatopoeic terms, proper names and interactive terms (such as words with social or pragmatic
function) decrease significantly (Bates et al., 1994; Caselli et al., 1995).
The second part of the PVB form is designed to assess morphological development: the use of inflected forms of nouns,
adjectives and verbs. In the third part, the PVB asks parents and day-care teachers if the child is combining words.
Furthermore, three risk indicators of language delay were extracted from the PVB: vocabulary size 1.5 SDs below the
mean of the normative sample, absence of morphological ability and absence of combinatory language. These indicators
were selected based on a previous longitudinal study that found a predictive relationship between these three indicators
assessed at 24–30 months and the overall linguistic development measured a year later (Viterbori, Zanobini, Scopesi, & Usai,
2006).
The PVB has been extensively used in studies of early language skills of Italian toddlers (Caselli, Casadio, & Bates, 1999;
D’Odorico & Fasolo, 2007; D’Odorico, Carubbi, Salerni, & Calvo, 2001).
The First Language Test (Test del Primo Linguaggio, TPL, Axia, 1993) evaluates language abilities between 12 and 36 months.
In the present study, we administered four scales assessing receptive vocabulary and syntax, and expressive vocabulary and
syntax. In both scales, each receptive vocabulary and syntax item is presented in a four picture multiple-choice format. The
child must indicate the figure corresponding to the word or the sentence produced by the examiner. To measure production,
the child must name pictures of objects and describe vignettes illustrating simple actions (e.g., a child who is eating). The
score for each scale is obtained by summing the number of correct items. The TPL has been used in the study of typical
and atypical language development in Italian children (Cipriani, Chilosi, Pfanner, Villani, & Bottari, 2002; Volterra, Capirci,
Caselli, & Vicari, 2004).

1.4. Temperament measure

The Questionari Italiani del Temperamento (QUIT, Axia, 2002) assessed the child’s temperament as perceived by both moth-
ers and day-care teachers. The 12–36-month version of the QUIT consists of 56 six-point Likert-type questions of children’s
day-to-day behaviours. For each item, response ranges from 1 (“never”) to 6 (“always”). The QUIT detail six temperamental
74 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

Table 1
Means and standard deviations for vocabulary (PVB – words and sentences version).

Age in months Teachers Parents Normative data

M SD M SD M SD

24–25 107 82 225 194 331 167


26–27 122 81 314 165 334 174
28–29 218 180 273 138 435 155
30 299 167 496 134 446 168

Note: PVB, Primo Vocabolario del Bambino; T, teachers; P, parents.

dimensions: social orientation (pleasure in social situations/interactions with others; e.g. “He/she immediately looks at the
person who calls him/her”), inhibition to novelty (emotional reaction to novelty experience; e.g. “When he/she enters a
new environment, it took a long time before he/she is at ease”), motor activity (gross motor activity, speed of reaction to
environmental experience; e.g. “When he/she plays, he/she is always on the move”), positive emotionality (intensity of
positive emotional reactivity/expression; e.g. “When he/she plays, he/she laughs a lot”), negative emotionality (intensity
of negative emotional reactivity/expression; e.g. “If someone annoys him/her, he/she cries for a long time”) and attention
(ability to focus attention and to shift attention from one focus to another; e.g. “He/she looks carefully at every new object”).
Higher scores on the QUIT dimensions correspond to higher levels of the labelled characteristic. Aspects of temperamental
difficulty are qualified as follows: high motor activity, low attention, high inhibition to novelty, low social orientation, low
positive emotionality and high negative emotionality. Aspects of temperamental easiness are considered as follows: low
motor activity, high attention, low inhibition to novelty, high social orientation, high positive emotionality and low negative
emotionality (Axia, 2002). For the present study, in the day-care teachers’ evaluation, the Cronbach alphas ranged from .61
to .85; in the case of parents, the Cronbach alphas ranged from .60 to .72, except for the attention scale which Cronbach
alpha was below .60 (˛ = .48).

1.5. Statistical analyses

First, we explored the concurrent associations between temperamental characteristics and linguistic indicators. Calcu-
lating numerous correlations increases the risk of a type I error, i.e., to erroneously conclude the presence of a significant
correlation. For this reason, we adjusted the level of statistical significance of correlation coefficients by using a Bonferroni
correction.
Secondly, a non-hierarchical K-means cluster analysis (Rapkin & Luke, 1993) was performed to verify the existence of
particular aggregations of temperamental characteristics on both teachers’ and mothers’ questionnaires. We expected to
identify the same three temperamental profiles as in our previous study (Usai et al., 2009). The K-means procedure is
intended to minimize within-cluster variance (or to maximize within-category homogeneity) on criterion variables while
maximizing differences between clusters.
The variables were standardized to ensure that the differences in the means and standard deviations did not affect the
distances in forming the clusters. Kruskal–Wallis test was used to verify that the clusters were significantly different on the
criterion variables.
The cross-validation of the clusters was assessed using a split-sample replication in which the cluster solution was
generated in a two randomly selected subgroups of approximately half of the sample. The children were assigned to the new
clusters obtained in the two halves and the two solutions were then compared for agreement with the full sample solution
by using Cohen’s . The two resulting s were averaged (Barbaranelli, 2002).
To evaluate the generalizability of the cluster solution, the full sample of children was split into two age groups (24–27
months and 28–30 months) and the clustering procedure was applied to each group to verify whether the full sample cluster
solution was also adequate for both younger and older children. The children were assigned to the new clusters obtained in
the two halves and the two solutions were then compared for agreement with the full sample solution by using Cohen’s .
Having identified a cluster solution, comparisons were carried out between the scores obtained on the PVB and the TPL
from each temperamental profile. Kruskal–Wallis test was used to compare linguistic competence among all the profiles.
Profiles were then compared two-by-two using Mann–Whitney test.

2. Results

2.1. Descriptive analyses

Means and standard deviations for temperament and language (PVB and TPL) measures were compared to normative
data (Tables 1–3). The score distribution for both temperament characteristics and language ability were not different
from the normative sample except for the PVB scores of day-care teachers. In their evaluation, children were significantly
under the vocabulary score mean. Specifically, the mean vocabulary production assessed by teachers at 24–30 months
was comparable to the 10th percentile of the normative sample (Caselli & Casadio, 1995). Because the normative data
V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82 75

Table 2
Means and standard deviations for TPL vocabulary and syntax scales.

Age in TPL
months

Vocabulary Syntax
scale scale

Production Comprehension Production Comprehension

Sample (M ± DS) Normative data Sample Normative Sample (M ± DS) Normative Sample Normative
(M ± DS) (M ± DS) data data (M ± DS) data
(M ± DS) (M ± DS) (M ± DS)

24 8.44 ± 7.32 10.70 ± 5.91 12.64 ± 5.63 13.90 ± 4.94 8.67 ± 14.76 5.50 ± 6.69 5.11 ± 7.13 5.63 ± 5.72
25–27 10.04 ± 6.35 12.40 ± 5.48 16.52 ± 3.15 15.80 ± 3.67 13.98 ± 14.79 10.10 ± 9.51 10.92 ± 6.79 9.76 ± 5.95
28–30 12.49 ± 5.71 14.60 ± 3.75 16.64 ± 4.57 17.10 ± 3.75 19.33 ± 17.58 20.20 ± 14.20 12.88 ± 6.46 12.80 ± 6.18

Note: TPL, Test del Primo Linguaggio (First Language Test).

was obtained from parents’ evaluations of children, the normative data cannot be compared to the data generated by
teachers (Caselli, Pasqualetti, & Stefanini, 2007). In fact, we assume that the child–teacher relationship is different from the
child-parent relationship, and this difference may affect teachers’ evaluations.

2.2. Comparison between day-care teachers and parents on child temperament evaluation

Preliminary analyses on parent and teacher ratings of child temperament dimensions revealed some significant dif-
ferences. Compared to parents, teachers reported lower scores for the dimensions of social orientation [4.08 ± .71 versus
4.31 ± .63; t(100) = −2.435; p < .05], motor activity [3.01 ± .96 versus 3.55 ± .68; t(97) = −5.639; p < .001] and positive emo-
tionality [4.5 ± .83 versus 4.73 ± .58; t(97) = −2.210; p < .05]. Conversely, they reported a higher score for inhibition to novelty
[2.97 ± .83 versus 2.75 ± .68; t(97) = 2.210; p < .05].

2.3. Correlations between temperament dimensions and language production

As expected, dimensions of temperament were significantly correlated with measures of language (assessed by the same
informant, see Table 4).
According to the day-care teacher data, two dimensions of temperament, positive emotionality and attention, are corre-
lated with the semantic categories of sounds, people, routines and verbs. Furthermore, attention correlates with vocabulary
size and with the categories of nouns, adjectives, adverbs and function words. Consequently, aspects of temperamental
easiness including high attention and high levels of positive emotionality are associated with better language productivity.
According to the parent data, negative emotionality and motor activity are significantly negative correlated with several
measures of language (see Table 4). Thus, aspects of temperamental difficulty are associated with lower scores on measures
of vocabulary size and composition.
Taken together, these results indicate that children who are perceived by both their teachers and their mothers as being
easy show better language abilities compared to children who are perceived as difficult.
These results are constrained by a possible source bias because temperament and language questionnaires were com-
pleted by either parents or day-care teachers. For this reason, we also correlated temperamental dimensions evaluated by
day-care teachers with the vocabulary scores obtained by the parents’ compilation of the PVB. In fact, while the temperament
questionnaire has been developed to be compiled by both parents and teachers, the PVB was originally developed only for
parents. The correlations show that attention continues to be associated with the language measures (correlations range
from r = .326 to r = .470, with an exception for the sounds category).
Table 5 reports the correlations between temperament and language measures from the TPL. In the day-care teach-
ers’ evaluation, four temperament dimensions—social orientation, inhibition to novelty, motor activity and attention—are

Table 3
Means and standard deviations for QUIT temperament dimensions.

QUIT (M ± DS) Teachers Parents Normative data

24–30 24–30 12–36

Social orientation 4.08 ± .71 4.31 ± .63 4.24 ± .65


Inhibition to novelty 2.97 ± .83 2.75 ± .68 2.78 ± .79
Motor activity 3.01 ± .96 3.55 ± .68 3.47 ± .73
Positive emotionality 4.5 ± .83 4.73 ± .58 4.69 ± .57
Negative emotionality 3.02 ± 1.14 3.08 ± .87 3.09 ± .88
Attention 4.26 ± .72 4.30 ± .59 4.36 ± .64

Note: QUIT, Questionari Italiani del Temperamento.


76 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

Table 4
Correlations between temperament dimensions and language measures (teachers/parents).

PVB QUIT

Social orientation Inhibition to novelty Motor activity Positive Negative Attention


emotionality emotionality

T P T P T P T P T P T P

Sounds .06 −.10 −.24 −.18 −.01 −.15 .34 *


.10 −.19 −.25 .39 *
.32*
Nouns .24 −.20 −.23 −.12 −.16 −.33* .26 .18 −.12 −.40* .42* .23
People .26 −.20 −.17 −.16 −.13 −.29 .33* .12 −.11 −.36* .43* .20
Routines .29 −.06 −.21 −.13 −.18 −.25 .31* .18 −.14 −.23 .47* .31
Verbs .24 −.22 −.23 −.11 −.19 −.36* .33* .14 −.16 −.35* .46* .12
Adjectives .25 −.23 −.21 −.15 −.20 −.34* .30 .12 −.13 −.34 .45* .09
Functors .30 −.23 −.13 −.05 −.19 −.34* .21 .08 −.12 −.31 .42* .15
Adverbs .29 −.20 −.16 −.08 −.18 −.35* .21 .08 −.14 −.29 .44* .17
Total .26 −.22 −.22 −.12 −.08 −.36* .29 .16 −.14 −.39* .45* .21

Note: QUIT, Questionari Italiani del Temperamento; PVB, Primo Vocabolario del Bambino; T, teachers; P, parents.
*
p < .001.

correlated with scores obtained by the children on TPL test. In particular, correlations between attention and the TPL scales
are all statistically significant.
In the parents’ evaluation, two temperament dimensions, negative emotionality and motor activity, correlate significantly
with scores obtained by the children on TPL test. Both production and comprehension, vocabulary and syntax, negatively
correlate with negative emotionality. Only expressive syntax negatively correlates with motor activity.
These results confirm the association that was found between temperament and language skills evaluated by using the
PVB. However, because language and temperamental measures were not recorded by the same informant, the associative
link between the two is even more meaningful.
In the day-care teachers’ assessment, attention is significantly associated with TPL scores, whereas in the parents’ assess-
ment, negative emotionality and TPL scores are negatively related. Both the day-care teachers’ and parents’ data highlight
the link between language ability and child temperamental characteristics, which are indicative of temperamental easiness.
In the case of day-care teachers, these characteristics are high levels of attention and social orientation, while for parents
these characteristics are low levels of both negative emotion and motor activity.

2.4. Day-care teachers

2.4.1. Children’s temperamental profiles in day-care teachers’ evaluation


Three temperamental profiles were highlighted by using K-means cluster analysis on day-care teachers data, replicating
previous research (Usai et al., 2009; see Table 6). Kruskal–Wallis test verifying differences between the clusters showed
significant differences for each temperamental dimension (all p < .001).
In the solution obtained from day-care teachers, the first profile (28 children) is characterised by high levels of inhibition
to novelty and negative emotionality. Social orientation is in the normal range, suggesting that these children are at ease
in familiar situations and with familiar people. Motor activity and positive emotionality is moderate, whereas attention is
close to the mean. We have defined this profile as “inhibited”: in general, these children are uneasy in new situations and
cry frequently. On occasion, they may show very intense emotions, usually fear, when faced with novelty. These children
approach new situations or people by watching attentively from a sheltered position.

Table 5
Correlations between temperament dimensions (teachers/parents) and TPL scores.

TPL QUIT

Social orientation Inhibition to novelty Motor activity Positive Negative Attention


emotionality emotionality

T P T P T P T P T P T P

Vocabulary scale
Production .12 −.12 −.07 −.17 −.22* −.22 −.08 .13 −.15 −.38** .31** .07
Comprehension .18 −.12 −.10 .05 −.11 −.11 −.07 −.04 −.01 −.35** .25* −.03
Syntactic scale
Production .31** −.14 −.17 −.11 −.19 −.28* .01 .05 −.05 −.26* .45** .10
Comprehension .23* −.16 −.23** −.05 −.04 −.08 .09 .01 −.07 −.33** .26* −.04

Note: QUIT, Questionari Italiani del Temperamento; TPL, Test del Primo Linguaggio; T, teachers; P, parents.
*
p < .05.
**
p < .01.
V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82 77

Table 6
Central points of the final clusters (Z points) on QUIT compiled by teachers.

QUIT Cluster 1 (N = 28) Cluster 2 (N = 45) Cluster 3 (N = 30)

Social orientation .24 .39 −.79


Inhibition to novelty 1.04 −.66 .02
Motor activity −.20 −.34 .70
Positive emotionality −.54 .60 −.42
Negative emotionality 1.08 −.39 −.42
Attention −.32 .77 −.86

The largest group within our sample was the second profile, which included 45 children who may be defined with
temperamental characteristics typical of an Italian preschooler. Indeed, this profile is characterised by the proximity to the
mean values of all the temperamental dimensions: good social orientation, low inhibition to novelty, medium-low motor
activity and negative emotionality, medium-high positive emotionality and attention.
The third profile was present in 30 children who are characterised by low levels of attention and high motor activity. Social
orientation and emotionality are low, whereas inhibition to novelty is close to the mean value. We have defined this profile
“inattentive”; indeed, these children are easily distracted and unfocused, demonstrating a poor capacity to modulate their
motor activity. They show difficulty in establishing satisfactory relationships with peers, perhaps because of their excessive
motor activity and low attention, which makes them impulsive and less able to effectively modulate their own behaviour
in relation to the environment.

2.4.2. Teachers’ clusters cross validation and generalizability of the cluster solution
We replicated the cluster solution by randomly splitting the full sample into two halves, samples A and B. Using
Kruskal–Wallis test, the three clusters were differentiated on most temperamental dimensions within each sample (p < .001).
Only motor activity was not significantly different between clusters in sample B.
The children were assigned to the new clusters obtained in the two halves and the two solutions were then compared for
agreement with the full sample solution. Cohen’s s were .75 (p < .001) and .77 (p < .001), for sample A and B, respectively,
and the mean  was .76. The three-cluster solution was regarded as adequate (mean  values greater than .60 are considered
acceptable; Barbaranelli, 2002).
To evaluate the generalizability of this cluster solution, the full sample of children was split into two age groups (24–27
months and 28–30 months) and the clustering procedure was applied to each group. The children were assigned to the new
clusters obtained in the two halves and the two solutions were then compared for agreement with the full sample solution
by using Cohen’s . For the teachers’ solution, Cohen’s ␬s were .84 (p < .001) and .66 (p < .001), for the 24–27 month and 28–30
month sample, respectively. We conclude that the same temperamental profiles can be identified from 24 to 30 months.

2.4.3. Associations between temperamental profiles identified by day-care teachers and language measures (PVB and TPL)
Firstly, we compared PVB scores of each profile identified by day-care teachers, to explore the relationship between
temperament and language development.
The Kruskal–Wallis test performed on average production of words and on vocabulary composition (average proportion
of individual semantic categories) (Table 7) indicates significant differences between the three groups.
Comparing the “typical” profile to each of the other two (i.e., “inhibited” and “inattentive”) on vocabulary size and compo-
sition, specific differences were highlighted. The Mann–Whitney test indicated that children with a “typical” temperament
have larger vocabulary compared to the “inattentive” profile (Z = −3.661, p < .001, mean rank of 26.42 and mean rank of
45.06 for “inattentive” and “typical” profiles, respectively). The composition of “typical” children’s vocabulary is more lex-
ically mature because it is made up of a larger number of verbs, adjectives and adverbs compared to “inattentive” children
(Graph 1). Alternatively, the categories of sounds, people and routines, which are the most primitive vocabulary components,
prevail in the semantic repertoire of “inattentive” children.
Similarly, the comparison between “inhibited” and “typical” profiles indicates that children with “typical” temperament
have a larger vocabulary (Z = −2.616, p < .01, mean ranks of 28.41 and 41.65 for “inhibited” and “typical” profile, respectively).

Table 7
Comparison between the temperamental profiles identified by teachers on the PVB vocabulary composition (Kruskal–Wallis test; the values are expressed
in mean ranks).

Cluster N Total Vocabulary composition

Sounds Nouns People Routines Verbs Adjectives Functors Adverbs

Inhibited 28 45.32 54.25 51.86 55.32 48.30 44.88 47.25 48.70 58.52
Typical 45 64.20 43.31 49.30 42.75 46.06 63.60 62.20 54.86 54.94
Inattentive 30 38.63 60.95 54.40 60.77 62.47 39.93 39.77 49.18 39.90

H 15.006 6.676 .537 7.258 5.937 13.411 11.097 1.004 6.796


p .001 .036 n.s. .027 .051 .001 .004 n.s. .033
78 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

Z=-2.482 Z=-2.532 Z=-2.609 Z=-3.365 Z=-3.133 Z=-2.211


50 p<0.05 p<0.05 p<0.01 p<0.001 p<0.01 p<0.05

45

40

35

Mean Ranks
30

25

20

15

10

0
Sounds Nouns People Routines Verbs Adjectives Functors Adverbs
Typical Inattentive

Graph 1. Comparison between “typical profile” and “inattentive profile” on the vocabulary composition.

“Typical” children used significantly more verbs (Z = −2.630, p < .01, mean ranks of 28.38 and 41.67 for the “inhibited” and
“typical” profiles, respectively) and produced more adjectives (Z = −2.162, p < .05, mean ranks of 29.82 and 40.75 for the
“inhibited” and “typical” profiles, respectively).
Finally, the only significant difference between “inhibited” and “inattentive” groups was in adjective production
(Z = −2.301, p < .05; mean ranks of 34.77 and 24.58 for the “inhibited” and “inattentive” profiles, respectively).
Regarding the distribution of risk factors for language delay extracted from the PVB (reduced vocabulary size, absence of
morphological ability and absence of combinatory language), the “inhibited” and “inattentive” groups showed the highest
proportions of children at risk of language delays. As for vocabulary size, 54% and 53% of children with “inhibited” and
“inattentive” profile, respectively, were at risk. Conversely, 75% of the children in the “typical” profile showed no vocabulary
size risk (2 (2) = 8.360, p < .05). The same pattern was shown for morphological ability; the groups of “inattentive” and
“inhibited” children showed high percentages of risk (70% and 61%, respectively), whereas only 36% of children in the
“typical” profile did not have morphological ability yet (2 (2) = 9.037 p < .05). We did not discover a statistical difference
in combinatory language between groups; however, the groups of “inattentive” and “inhibited” children showed higher
percentages of children at risk (47% and 43%, respectively) compared to the “typical” profile (25%).
Finally, we compared the TPL scores of each temperamental profile (Table 8).
The “typical” profile had the highest scores on the receptive syntax scale compared to the groups of “inhibited” and
“inattentive” children. We found a nearly significant difference between profiles on the expressive syntax scale; the “typical”
profile had higher scores than the groups of “inhibited” and “inattentive” children. No differences were found between groups
on the vocabulary scale.
The comparison between “typical” and “inattentive” profiles indicates that children with “typical” temperament had the
highest scores on the receptive and expressive syntax scale (receptive: Z = −2.313, p < .05; mean ranks of 36.45 and 25.67;
expressive: Z = −2.281, p < .05; mean ranks of 31.22 and 21.18, for the “typical” and “inattentive” profiles, respectively).
No differences were found between “typical” and “inhibited” profiles on TPL scores. Similarly, no differences were found
between “inhibited” and “inattentive” profiles.

2.5. Parents

2.5.1. Children’s temperamental profiles in parents’ evaluation


Four temperamental profiles were highlighted by using K-means cluster analysis on parents’ data. A 3-cluster solution
was tested in order to replicate the data found with teachers. Nevertheless, the validation of the 3-cluster solution, performed

Table 8
Comparison between the temperamental profiles identified by teachers on the TPL scores (Kruskal–Wallis test, the values are expressed in mean ranks).

Cluster Vocabulary scale Syntax scale

Production Comprehension Production Comprehension

Inhibited 38.59 42.09 36.13 40.92


Typical 46.26 52.91 42.60 52.12
Inattentive 37.02 44.89 28.30 37.10

H 2.631 2.961 5.829 6.029


p n.s. n.s. .054 .049
V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82 79

Table 9
Central points of the final clusters (Z points) on QUIT compiled by parents.

QUIT Cluster 1 (N = 24) Cluster 2 (N = 31) Cluster 3 (N = 21) Cluster 4 (N = 26)

Social orientation −.10 −.15 −.90 .97


Inhibition to novelty .93 −.49 −.44 .08
Motor activity −.46 −.67 .36 .93
Positive emotionality −.77 .72 −.64 .32
Negative emotionality .96 −.55 −.57 .26
Attention −.57 .81 −.36 −.14

by randomly splitting the full sample into two halves (samples A and B), indicates that two cluster solutions did not confirm
the results reported in the full sample solution. Furthermore, Cohen’s s were .62 and .12 for samples A and B, respectively,
and the mean  was .37, indicating poor agreement. The results suggest that the three-cluster solution found within teachers’
ratings is not an acceptable model for parents’ data.
Instead, the four-cluster solution proved to be adequate (see Table 9). Using Kruskal–Wallis test, the four clusters were
significantly differentiated for the temperamental features considered by the QUIT (all p < .001).
Two profiles—cluster 2 and cluster 1—are similar to those identified in the teachers’ evaluation as “typical” and “inhibited”.
Cluster 3, which included 21 children, was characterised by lower levels in most dimensions, in particular, social orientation
and positive/negative emotionality. We have defined this group of children as “introverted”. Whereas children with inhibited
temperament tend to be timid with situation that are novel or unfamiliar but show a normal social orientation, children
with “introverted” temperament tend to be low in sociability and reluctant to get involved in social activities. Furthermore,
inhibited children show a high negative emotionality associated with a low positive emotionality that suggests a prevalence
of expression of negative feelings, while “introverted” children are low on both positive and negative emotionality indicating
an attitude to be cautious to express their feelings.
Cluster 4 (26 children) was characterised by high levels in social orientation and motor activity. We have defined this
group of children as “extroverted”; indeed, these children show high interest for activities that involve other people.

2.5.2. Parents’ clusters cross validation and generalizability of the cluster solution
We replicated the four-cluster solution by randomly splitting the full sample into two halves, samples A and B. Using
Kruskal–Wallis test, in each sample the four clusters are differentiated on all temperamental dimensions (p ranged from .001
to .003), except social orientation in sample B. Assigning children to the new clusters, the comparisons of the two solutions
with the original showed acceptable values of agreement. Cohen’s s were .75 (p < .001) and .54 (p < .001) for samples A and
B, respectively, and the mean  was .65 (mean  values greater than .60 are considered acceptable; Barbaranelli, 2002).
To evaluate the generalizability of this cluster solution, the full sample of children was split into two age groups (24–27
months and 28–30 months) and the clustering procedure was applied to each group. The children were assigned to the new
clusters obtained in the two halves and the two solutions were then compared for agreement with the full sample solution
by using Cohen’s . For parents’ solution, Cohen’s s were .55 (p < .001) for the 24–27-month sample and .94 (p < .001) for
28–30-month sample. We conclude that the same temperamental profiles can be identified from 24 to 30 months, though
the four factor solution seems to be more adequate for older children.

2.5.3. Associations between temperamental profiles identified by parents and language measures (PVB and TPL)
To identify the relationship between temperamental profiles identified by parents and language development, we com-
pared PVB scores of each profile.
The Kruskal–Wallis test on the average production of words indicated significant differences between the four groups
(H(3) = 13.149, p < .01). The “typical” profile had the highest scores (mean rank = 64.16) compared to the groups of “intro-
verted” (mean rank = 47.58), “inhibited” (mean rank = 42.91), and “extroverted” children (mean rank = 38.66).
The only significant difference between the four clusters in vocabulary composition was sound proportion (H = 9.847,
p < .05; “typical”: mean rank = 37.65; “inhibited”: mean rank = 55.98; “introverted”: mean rank = 47.92; and “extroverted”
mean rank = 59.76).
Comparing each profile (“inhibited”, “introvert” and “extrovert”) to the “typical” one on vocabulary size and composition,
specific differences were highlighted. The Mann–Whitney test indicated that children with a “typical” profile have larger
vocabulary compared to the “inhibited” (Z = −2.536, p < .05, mean rank of 31.53 and mean rank of 20.61 for “typical” and
“inhibited” profiles, respectively), compared to “introverted” children (Z = −1.958, p < .05, mean rank of 29.27 and mean rank
of 20.93 for “typical” and “introverted” profiles, respectively), and compared to “extroverted” (Z = −3.502, p < .05, mean rank
of 35.35 and mean rank of 20,00 for “typical” and “extroverted” profiles, respectively).
On the vocabulary composition, the “inhibited” children used significantly more routines if compared with “typical”
(Z = −2.211, p < .05, mean rank of 30.52 and mean rank of 24.50 for “inhibited” and “typical” profiles, respectively). Vice
versa, the “introverted” children produces more people names (Z = −2.508, p < .05, mean rank of 32.50 and mean rank of
21.81 for “introverted” and “typical” profiles, respectively). The categories of sounds and routines prevail in the reper-
toire of “extroverted” if compared with “typical” (sound Z = −2.258, p < .05, mean rank of 33.98 and mean rank of 24.08 for
80 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

“extroverted” and “typical” profiles, respectively; routines Z = −2.991, p < .01, mean rank of 35.76 and mean rank of 22.65 for
“extroverted” and “typical” profiles, respectively).
No differences were found between the “inhibited” and “introverted” profiles on vocabulary size and composition and
between the “inhibited” and “extroverted” profiles. The only one difference between “introverted” and “extroverted” children
was on the proportion of nouns (Z = −2.170, p < .05, mean rank of 26.80 and mean rank of 18.25 for “introverted” and
“extroverted” profiles, respectively).
Regarding the distribution of risk factors for language delay extracted from the PVB, the only difference between groups
was reduced vocabulary size. For this index, the group of “introverted” children had the highest proportion of children at
risk (38%), whereas 97% of “typical” children showed no risk (2 (3) = 10.122, p = .018).
Finally, we compared the TPL scores of each temperamental profile. “Typical” children had the highest receptive
vocabulary scores (TPL; H = 10.280, p < .05; mean rank = 50.85) compared to the groups of “inhibited” (mean rank = 28.81),
“introverted” (mean rank = 41.88) and “extroverted” children (mean rank = 39.18).

2.5.4. Parent and day-care teacher agreement for cluster assignment


The agreement of children assignment to cluster solutions between parents and day-care teachers was poor. Among the 42
children who met day-care teachers’ classification for the “typical” cluster, only 17 of these cases met the same classification
for parents (40.5%). The agreement for “inhibited” cluster assignment was equally poor and the two types of informants
agreed on only 10 cases. Finally, among the 28 children who met day-care teachers’ classification for “inattentive” cluster,
eight cases were classified as “inhibited” by parents (28.6%), whereas 5 were classified as “typical” (17.9%), six children were
classified as “introverted” (21.4%), and 9 children were classified as “extroverted” (32.1%).

3. Discussion

The results of this study support previous research by illustrating a link between temperament dimensions and language
development (e.g., Bates, Bretherton, & Snyder, 1988; Dixon & Smith, 2000; Matheny, 1989; Morales et al., 2000; Mundy &
Gomes, 1998; Slomkowski et al., 1992). Furthermore, the present research replicates the results of our previous study (Usai
et al., 2009) by identifying a relationship between specific temperamental profiles (identified by teachers) and different
levels of language ability and extends these results to parent-derived temperamental profiles.
As for teachers’ data, we identified the same temperamental profiles already found in our previous study: a “typical”
profile that grouped most children; an “inhibited” profile of children who show intense negative emotional manifestations
and difficulties in coping with novelty; and an “inattentive” profile, identifying children with relatively poor attention spans
and a high level of motor activity. The results show that the temperamental profiles derived from teachers’ reports may be
generalized to younger and older children.
Regarding the relationship between temperament and language, according to the teachers, children in the “typical” profile
communicate with a wider and more mature vocabulary, with a larger number of verbs, adjectives and adverbs compared to
the other groups. Furthermore, they show the lowest conditions of risk for two of the indices taken from the PVB (vocabulary
size and morphological ability) and obtained the highest scores in both receptive and productive syntax assessed using a
standardized test.
The “inattentive” profile is the least competent group of children and is distinguished by a semantic repertoire in which
the most primitive vocabulary components (sounds, people and routines) prevail. Though this group of children shows on
the whole the poorest language performance in both the PVB teachers’ report and in the TPL standardized test of language
development, it does not differ in a significant way from the “inhibited” profile.
As compared with our previous study, we confirm a better language performance in children with a “typical” temper-
amental profile, also using a standardized language measure in addition to the teachers’ language report. Nevertheless,
differently from the previous study, we did not find the “inattentive” profile to be at greater risk for language acquisition
than the “inhibited”, suggesting that both a poor attentive control and high levels of inhibition may be associated with a
more immature language development.
As suggested in our previous study, the reasons to explain the poorer performance of “inattentive” and “inhibited”
children may be different. On the one hand, children with poor attentive control may be less able to modulate their attention
effectively in such way as to maximise the acquisition of words especially in the presence of distractions (Dixon, Salley &
Clements, 2006). A poor attentive control associated with high motor activity may be considered particularly troublesome
in educational settings, since attentive control plays an important role in many cognitive processes and its good functioning
is considered important for learning. In fact, in a previous study (Usai, Viterbori, & Alcetti, 2007), we found that preschool
children with an “inattentive” temperamental profile showed poorer school readiness skills at the end of kindergarten
compared to “typical” and “inhibited” children. On the other hand, “inhibited” children could be exposed to a smaller variety
of linguistic stimulation as a consequence of their attitude to avoid new situations and to show intense negative emotional
manifestations. In other words, adults would be less inclined to enter a relation and dialogue with the child, reducing the
number of communicative interchanges. We suggest that to be sociable, quiet and not preoccupied by novelty could increase
the probability to enter a good relation with adults and peers and, as a consequence, to be more often involved in learning
and social activities.
V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82 81

The associations between parent-derived temperament profiles and language skills are less definite. The cluster analysis
performed on parent reports did not completely confirm the temperamental constellations found on teacher reports. In
addition to “typical” and “inhibited” groups, parents identified an “introverted” group characterised by slightly lower scores
in most temperamental dimensions, in particular social orientation; and an “extroverted” group of children who showed
high levels of social orientation and motor activity. The results show that the temperamental profiles derived from parents’
reports may be generalized to younger and older children, though this solution seem to be more adequate for older children.
The children who were categorised in the “typical” profile by parents showed the best linguistic skills (the largest vocab-
ularies, both productive and receptive, as well as the most mature semantic production). Only few differences were found
among the other three temperamental profiles.
Both teachers and parents evaluated “typical” children as having the highest linguistic abilities, though they do not
completely agree in the children assignment to this cluster.
A bulk of literature has shown low levels of agreement between parents and teachers regarding children’s behaviour.
A meta-analysis by Achenbach, McConaughy, and Howell (1987), for example, found that the correlations between parent
and teacher reports were low to moderate, with an average of .27. Different reasons may explain this discrepancy between
parents’ and teachers’ evaluations. Some authors have suggested that children may display discrepant behaviours in different
situations (Rothbart & Bates, 1998). In particular, the behaviour of very young children may be quite sensitive to specific
physical and relational contexts; that means that children might behave differently in different settings and with different
people. Additionally, different settings may prompt different types of behaviours: for example, in day-care centres peer-to-
peer social activities are stimulated by the presence of many children, whereas adult-child vertical interactions are more
frequent at home.
An additional reason to expect moderate agreement between teacher and parent ratings is that parents and teachers may
have different perceptions of behaviour (Reid & Maag, 1994). Parents of pre-schoolers might be less sensitive in assessing
their child’s attention span due to the absence of external criteria to evaluate this specific behaviour. Furthermore, parents
do not have the advantage of comparing the behaviour of their own child with that of a large number of other children,
which is the case for teachers (Re & Cornoldi, 2009).
These results suggest the importance to analyse both language skills and the behavioural manifestations of temperament
during mother–child and teacher–child interaction to better understand whether the discrepancies we found in parents’
and teachers’ evaluations may reflect the specific characteristics of the relationship.

References

Achenbach, T. M., McConaughy, S. H. & Howell, C. T. (1987). Child/adolescent behavioral and emotional problems: Implications of cross-informant corre-
lations for situational specificity. Psychological Bulletin: 101., 213–232.
Aksan, N., Goldsmith, H. H., Smider, N. A., Essex, M. J., Clark, R., Hyde, J. S., et al. (1999). Derivation and prediction of temperamental types among preschoolers.
Developmental Psychology: 35., 958–971.
Axia, G. (1993). TPL. Test del Primo Linguaggio. Firenze: Organizzazioni Speciali.
Axia, G. (2002). QUIT. Questionari italiani del temperamento. Trento: Erickson.
Barbaranelli, C. (2002). Evaluating cluster analysis solutions: An application to the Italian NEO personality inventory. European Journal of Personality: 16.,
S43–S55.
Bates, E., Bretherton, I. & Snyder, L. (1988). From first words to grammar. New York: Cambridge University Press.
Bates, E., Marchman, V. A., Thal, D., Fenson, L., Dale, P., Reznick, J. S., et al. (1994). Development and stylistic variation in the composition of early vocabulary.
Journal of Child Language: 21., 85–123.
Bloom, L. (1993). The transition from infancy to language: Acquiring the power of expression. Cambridge, England: Cambridge University Press.
Buss, A. & Plomin, R. (1984). Temperament: Early personality traits. Hillsdale, N.J.: Erlbaum.
Caselli, M. C., Bates, E., Casadio, P., Fenson, J., Fenson, L., Sanderl, L., et al. (1995). A cross-linguistic study of early lexical development. Cognitive Development:
10., 159–199.
Caselli, M. C. & Casadio, P. (1995). Il Primo Vocabolario del Bambino. Milano: Franco Angeli.
Caselli, M. C., Casadio, P. & Bates, E. (1999). A comparison of the transition from first words to grammar in English and Italian. Journal of Child Language: 26.,
69–111.
Caselli, M. C., Pasqualetti, P. & Stefanini, S. (2007). Parole e frasi nel “Primo Vocabolario del Bambino”. Nuovi dati normativi fra 18 e 36 mesi e Forma breve del
questionario. Milano: Franco Angeli.
Caspi, A. & Silva, P. A. (1995). Temperamental qualities at age 3 predict personality traits in young adulthood: Longitudinal evidence from a birth cohort.
Child Development: 66., 486–498.
Chess, S. & Thomas, A. (1984). Origins and evolution of behavior disorder: Infancy to early adult life. New York: Brunnet/Mazel.
Cipriani, P., Chilosi, A. M., Pfanner, L., Villani, S. & Bottari, P. (2002). Il ritardo del linguaggio in età precoce: profili evolutivi e indici di rischio. (Vol. Eds.),
Indici di rischio nel primo sviluppo del linguaggio (pp. 95–108). Milano: Franco Angeli.
D’Odorico, L., Carubbi, S., Salerni, N. & Calvo, V. (2001). Vocabulary development in Italian children: A longitudinal evaluation of quantitative and qualitative
aspects. Journal of Child Language: 28., 351–372.
D’Odorico, L. & Fasolo, M. (2007). Nouns and verbs in the vocabulary acquisition of Italian children. Journal of Child Language: 34., 891–907.
Dixon, W. E. & Shore, C. (1997). Temperamental predictors of linguistic style during multiword acquisition. Infant Behavior and Development: 20., 99–103.
Dixon, W. E. & Smith, P. H. (2000). Links between early temperament and language acquisition. Merril-Palmer Quarterly: 46., 417–440.
Dixon, W. E., Salley, B. J. & Clements, A. D. (2006). Temperament, distraction, and learning in toddlerhood. Infant Behavior and Development: 29., 342–357.
Fenson, L., Dale, S. P., Reznick, J. S., Thal, D., Bates, E., Hartung, J. P., et al. (1993). The MacArthur Communicative Development Inventory: User’s guide and
technical manual. San Diego: Singular Publishing Company.
Gartstein, M. A. & Rothbart, M. K. (2003). Studying infant temperament via the Revised Infant Behavior Questionnaire. Infant Behavior and Development:
166., 1–23.
Goldsmith, H. H. (1996). Studying temperament via construction of the Toddler Behavior Questionnaire. Child Development: 67., 218–235.
Keogh, B. K. (1989). Applying temperament research to school. (Vol. Eds.), Temperament in childhood. Chichester, England: John Wiley & Sons, Ltd.
Keogh, B. K. (1994). Temperament and teachers’ views of teachability. (Eds.), Prevention and early intervention: Individual differences as risk factors for the
mental health of children (pp. 246–256). New York: Brunner/Mazel.
82 V. Garello et al. / Infant Behavior & Development 35 (2012) 71–82

Komsi, N., Räikkönen, K., Pesonen, A. K., Heinonen, K., Keskivaara, P., Järvenpää, A. L., et al. (2006). Continuity of temperament from infancy to middle
childhood. Infant Behavior and Development: 29., 494–508.
Kubicek, L. F., Emde, R. N. & Schmitz, S. (2001). Temperament, mental development, and language in the transition from infancy to early childhood. (Vol.
Eds.), Infancy to early childhood: Genetic and environmental influences on developmental change. New York: Oxford University Press.
Martin, R. P. (1989). Activity level, distractibility, and persistence: Critical characteristics in early schooling. (Eds.), Temperament in childhood (pp. 451–462).
Ltd, Chichester, England: John Wiley & Sons.
Matheny, A. P., Jr. (1989). Temperament and cognition: Relations betweens temperament and mental test scores. (Eds.), Temperament in childhood. New
York: Wiley.
Morales, M., Mundy, P., Delgado, C. E. F., Yale, M., Messinger, D., Neal, R., et al. (2000). Responding to joint attention across the 6- though 24-month age
period and early language acquisition. Journal of Applied Developmental Psychology: 21., 283–298.
Moritz Rudasill, K. & Rimm-Kaufman, S. E. (2009). Teacher–child relationship quality: The roles of child temperament and teacher–child interactions. Early
Childhood Research Quarterly: 24., (2), 107–120.
Mullola, S., Ravaja, N., Lipsanen, J., Hirstiö-Snellman, P., Alatupa, S. & Keltikangas-Järvinen, L. (2010). Teacher-perceived temperament and educational
competence as predictors of school grades. Learning and Individual Differences: 20., (3), 209–214. doi:10.1016/j.lindif.2010.01.008
Mundy, P. & Gomes, A. (1998). Individual differences in joint attention skill development in the second year. Infant Behavior and Development: 21., 469–482.
Rapkin, B. D. & Luke, D. A. (1993). Cluster analysis in community research: Epistemology and practice. American Journal of Community Psychology: 21., (2),
247–277.
Re, A. M. & Cornoldi, C. (2009). Two new rating scales for assessment of ADHD symptoms in Italian preschool children. A comparison between parent and
teacher ratings. Journal of Attention Disorders: 12., 532–539.
Reid, R. & Maag, J. W. (1994). How many fidgets in a pretty much: A critique of behaviour rating scales for identifying students with ADHD. Journal of School
Psychology: 32., 339–354.
Rothbart, M. & Bates, J. (1998). Temperament. In N. Eisenberg (Series Ed.) & W. Damon (Ed.), Handbook of child psychology : Vol. 3. Social, emotional, and
personality development (5th ed., Vol. 3, pp. 105–176). New York: Wiley.
Rothbart, M. K. (1981). Measurement of temperament in infancy. Child Development: 52., 569–578.
Rothbart, M. K. & Jones, L. B. (1998). Temperament, self-regulation, and education. School Psychology Review: 27., 479–491.
Slomkowski, C. L., Nelson, K., Dunn, J. & Plomin, R. (1992). Temperament and language. Relations from toddlerhood to middle childhood. Developmental
Psychology: 28., 1090–1095.
Stifter, C. A., Putnam, S. & Jahromi, L. (2008). Exuberant and inhibited toddlers: Stability of temperament and prediction to behavior problems. Development
and Psychopathology: 20., 401–421.
Usai, M. C., Garello, V. & Viterbori, P. (2009). Temperamental profiles and linguistic development: Differences in the quality of linguistic production in
relation to temperament in children of 28 months. Infant Behavior and Development: 32., 322–330.
Usai, M. C., Viterbori, P. & Alcetti, A. (2007). Temperamento e identificazione precoce delle difficoltà di apprendimento. Psicologia Clinica e dello Sviluppo:
2., 253–269.
Vagh, S. B., Pan, B. A. & Mancilla-Martinez, J. (2009). Measuring Growth in bilingual and monolingual children’s English productive vocabulary development:
The utility of combining parent and teacher report. Child Development: 80., (5), 1545–1563.
Van den Akker, A. L., Deković, M., Prinzie, P. & Asscher, J. J. (2010). Toddlers’ temperament profiles: Stability and relations to negative and positive parenting.
Journal of Abnormal Child Psychology: 38., (4), 485–495.
Viterbori, P., Zanobini, M., Scopesi, A. M. & Usai, M. C. (2006). Profili dello sviluppo linguistico precoce: analisi di alcuni indicatori di rischio. Psicologia clinica
dello sviluppo: 1., 141–162.
Volterra, V., Capirci, O., Caselli, M. C. & Vicari, S. (2004). Language in preschool Italian children with Williams and Down syndrome. (Vol. Eds.), Williams
syndrome across languages (pp. 163–185). Amsterdam: John Benjamins Publishing Company.

You might also like