Peabody Picture Vocabulary Test-III: Normative data for Spanish-speaking pediatric population

1Introduction

Vocabulary is generally defined as the knowledge a person has regarding words and their meaning (Butler et al., 2010). Two types of vocabulary are identified: expressive vocabulary, which alludes to the words that a person can produce, and receptive vocabulary, which refers to the words that are acquired visually or aurally; that is, words that can be learned by hearing someone or by reading (Burger & Chong, 2011).

Vocabulary development in children is fundamental for adequate school progress. Studies have shown that deficiencies in this aspect interfere with academic success resulting in reading comprehension difficulties (Pearson, Hiebert, & Kamil, 2007; Stahl & Fairbanks, 1986), and subsequent problems in the acquisition of new vocabulary. Studies also indicate that the vocabulary level attained by children is related to their overall linguistic and cognitive abilities (Feldman et al., 2005; Fernald, Perfors, & Marchman, 2006; Marchman & Fernald, 2008).

One of the most frequently utilized instruments in the measurement of vocabulary is the Peabody Picture Vocabulary Test-III (PPVT-III) (Pankratz, Morrison, & Plante, 2004; Stockman, 2000). The test was developed to measure receptive vocabulary and to quickly determine verbal aptitude in a wide range of ages (2 years and a half to 90 years old; Dunn, Dunn, & Arribas, 2010). The test involves presenting four images while the examiner says a word. The examinee must then indicate which image best represents the word said by the examiner (Dunn et al., 2010). The test advantages include quick administration time (between 10–15 minutes) and little oral communication from the examiner or the examinee, who can simply point to the answers. Accordingly, this test is particularly valuable for use with persons who are illiterate or those with aphasia, autism or cerebral palsy (Dunn et al., 2010).

The first edition was published in the United States of America (USA; Dunn, 1959) and was normed on Caucasian children. Later, in 1981, the PPVT revised (PPVT-R) was developed, which sought to reduce gender, religion or race bias. At that time, the norming sample included children as well as adults in the USA (2 years and a half to 40 years of age), and two parallel forms were created (Forms L and M). For the third edition, in 1997, a concerted effort was made to include minority populations in the USA, such as Asians, African Americans, Hispanics and Native Americans. At this time, the age range was extended to 90 years of age. The USA edition has two parallel forms (Forms A and B), but the version from Spain only has one. According to the authors of the manual, this decision was prompted by their understanding that in the Hispanic context, the psychologist rarely needs two parallel versions to evaluate one aptitude (Dunn et al., 2010). For the Spanish norming, a total of 2,550 persons representing 19 provinces in Spain were included (Dunn et al., 2010). Lastly, the most recent version was published in 2007, the PPVT-IV. However, currently there is no version of that edition for Spain or any Latin American country. For that reason, the PPVT-III was used in the present study.

The PPVT-III was developed with the aim of being a culturally valid instrument. For that reason, as was previously mentioned in this document, the norming samples were broader and included ethnic minority representatives. This resulted in a third edition, which is the least culturally biased when compared to other versions of the PPVT. Evidence of adequate performance of this test has been obtained in studies with ethnic minority participants. For example, Washington and Craig (1999) found that the PPVT-III was an appropriate instrument for use with African American children. Haitana, Pitama, & Rucklidge (2010) found the same with a sample of Maori children. Even so, there are researchers who argue that the inclusion of minority groups in the standardization sample does not free an instrument from cultural biases (Haitana et al., 2010). For example, a vocabulary test is based on the knowledge of word meaning, which can change depending on cultural experience (Stockman, 2000). This implies that word meaning can vary between people of different linguistic communities due to cultural nuances.

One must remember that the PPVT was initially created for use with English-speaking persons in the USA. Although it was translated into Spanish, and a Hispanic American version has been available since 1986 (Dunn, Lugo, Padilla, & Dunn, 1986), the standardization of the PPVT-III was carried out only in Spain. For that reason, the use of the Spaniard norms in Latin American countries may not be adequate, even with the commonality of the Spanish language. Not only is their cultural background different, so is the frequency with which a specific word is used in each country. For example, the word “calcetín” [sock] is known by most Spanish-speakers, but it is not a frequently used word in Colombia, where it is commonly referred to as “media”. Today, many professionals in Latin American countries utilize the PPVT-III version from Spain. However, despite its frequent use, norms for the PPVT-III are not available for Latin American countries. This makes it necessary to provide normative data for this population.

Furthermore, even though normative data is available for the population of Spain, it dates to the year 2003. Language changes overtime, and thus, the degree of difficulty of the PPVT-III stimuli may have transformed due to the frequency in the use of words by the current population. Words can change their meaning or can be substituted for others of increased present-day use. For example, several years ago the word “jolgorio” [revelry] was commonly utilized, but currently it is seldom used, and “fiesta ruidosa” [noisy party] is more commonplace. Thus, updated norms are also required for the population of Spain.

In view of the previously stated information, obtaining updated and adequate norms for each population is necessary. For that reason, the objective of the present study was to develop normative data for children and adolescents of nine countries in Latin America (Chile, Cuba, Ecuador, Guatemala, Honduras, Mexico, Paraguay, Peru, and Puerto Rico) and Spain.

2Method

3Results

The final multivariate linear regression models for the ten country-specific PPVT-III Total scores were significant (see Table 1). In all countries, the PPVT-III Total scores increased linearly as a function of age. The PPVT-III Total scores for Chile, Cuba, Ecuador, Honduras, Mexico, Peru, Puerto Rico, and Spain were also affected by a quadratic age effect. Children from Chile, Ecuador, Guatemala, Honduras, Mexico, Paraguay, and Spain who had parent(s) with a MLPE >12 years obtained higher PPVT-III Total score than children who had PPVT-III Total score with a MLPE ≤12 years. The child’s sex affected the PPVT-III Total scores for Chile, Ecuador, Guatemala, Mexico, and Spain, so that boys achieved higher score than girls. The amount of variance these predictors explained in PPVT-III Total scores ranged from 46.0% (in Guatemala) to 72.1% (in Cuba). The assumptions of multiple linear regression analysis were met for all final models. There was not multicollinearity (the VIF values were below 10; VIF ≤1.131; collinearity tolerance values did not exceed the value of 1) or influential cases (the maximum Cook’s distance value was 0.184 in a F _(3,212)  distribution which correspond to percentile 9).

4Discussion

The purpose of the present study was to obtain normative data for the PPVT-III for children and adolescents from nine Latin American countries (Chile, Cuba, Ecuador, Guatemala, Honduras, Mexico, Paraguay, Peru, and Puerto Rico) and Spain. The final regression models for the PPVT-III accounted for 46.0% and 72.1% of the variance.

There are very few normative data studies published for the PPVT-III, thus comparison with the existing literature is complicated. The age variable was significantly related to the Total scores of the PPVT-III, such that scores increased linearly as children become older. This gradual increase in performance as age advances is in line with the results of a normative study carried out in Spain (Dunn et al., 2010). The authors reported a gradual increase in the total score of the PPVT-III, with an increase being especially noticeable between the ages of 12 and 16, until stabilization begins at the age of 17, especially between 19 and 20 years of age (Dunn et al., 2010). Results of the present study showed a curvilinear relationship of age and relative scores on the PPVT-III (reference score was a 6-year-old boy with high-educated parents). This relationship indicates that relative scores increase as age advances, with the most prominent increase in younger children than older ones. For example, the difference in relative scores for a 6-year-old boy with high-educated parents (reference value) compared to a 7-year-old with the same characteristics ranges from 6.8% to 14% percent depending on country of origin, while the difference in relative scores for a 14-year-old with those same characteristics compared to a 15-year-old is only 1.8% to 4.5%.

This gradual increase in PPVT-III scores coincides with studies that have attempted to determine the development of vocabulary during childhood, adolescence, and adulthood. For example, Biemiller and Slonim (2001) estimated that English-speaking children in the second grade knew an average of 5,200 words and attaining 8,400 words by the fifth grade. According to Guerra, Williamson, & Lucas-Molina (2012), vocabulary continues to expand during middle childhood, reaching 40,000 words during adolescence (18 years). Sullivan & Brown (2014) also reported a progressive increase in vocabulary in people aged 16 to 42 years. In addition, vocabulary is quite resistant to decline during normal aging, which can remain stable and even improve over time (Harada, Love, & Triebel, 2013).

The MLPE was also associated with PPVT-III Total scores. Van der Elst, Hurks, Wassenberg, Meijs, & Jolles (2011) were the first to use MLPE as a variable associated with verbal fluency tests in their normative data study. Van der Elst et al. (2011) reported that children whose parent(s) had a higher educational level did better on these tests as compared to children whose parent(s) had a lower educational level. Normative studies on the PPVT-III have not developed norms using this variable but a number of studies have looked at its influence on performance. For example, educational level of parent(s) of African-American children was related to the score obtained by children in the PPVT-III, such that children whose parent(s) had completed high school or college, scored higher than children whose parent(s) had not finished high school (Washington & Craig, 1999).

The fact that parents’ educational level influences the score obtained by children in the PPVT-III is not surprising since research has shown that vocabulary level obtained by children depends to a large extent on the environment in which they developed. For example, children from higher socioeconomic backgrounds had a higher level of vocabulary than those from lower socioeconomic backgrounds (Hoff, 2003). Similarly, children’s scores on the vocabulary task of the British Ability Scales battery depended heavily on parental education (Becker, 2011). Thus, children whose parent(s) had a high educational level had a broader vocabulary than children whose parents had a lower educational level. In addition, this study showed that early schooling of children positively influenced the development of vocabulary, but only in those children whose parent(s) had a lower educational level. It seems that the way parent(s) interact with their children and the stimulation they offer (e.g. reading books) is helpful for language development as well as the socioeconomic level (Becker, 2011).

Finally, the variable gender was associated with performance in the PPVT-III, such as males achieves higher scores than females. Dunn et al. (2010) found that gender had a small influence on PPVT-III Total scores, thus only norms based on age were created.

Conflict of interest

None to report.