Abstract
Background: As a construct, temperament provides a framework for understanding differences among individuals in reaction to their life experiences. The measurement of the construct concerns both researchers and clinicians.
Objective: The purpose of this study was to examine whether the School-Age Temperament Inventory continued to demonstrate reliability and validity when retested with three existent samples of parent respondents.
Method: Sample 1 was a sociodemographically and racially heterogeneous group of 200 children from New England in the United States. Data for Sample 2 was provided by 589 mothers and fathers from the state of Georgia in the United States. In Sample 3, data was provided by parents (principally, mothers) of 1,391 adolescents from Australia. Orthogonal Procrustes rotations were conducted to examine the underlying structure of the inventory when it was contrasted with the results obtained in the original standardization of the tool.
Results: The total coefficients of congruence for the samples were .88 to .97, while those of the four factors ranged from .84 to .98. Across the samples, Cronbach alphas for the dimensions ranged from .80 to .92. Independent t-tests identified that boys were significantly more active and less task persistent than girls. However, regression analyses revealed that sex accounted for only 5% of the variance in task persistence and activity.
Discussion: The results provide substantial additional support for the reliability and validity of the School-Age Temperament Inventory. Recommendations for future research are offered which include exploring the role of temperament in contributing to developmental outcomes in children and examining cross-cultural samples.
As a stylistic individual characteristic, temperament is an integral facet of human development. Temperament is the consistent reaction style that an individual demonstrates across a variety of settings and situations, particularly those that involve stress or change (McClowry, 2003). An understanding of the construct of temperament can assist nurses and other clinicians to recognize why clients react differently to life experiences, plan clinical strategies that appropriately account for individual differences, and assess the effectiveness of various interventions.
Most theorists in the temperament field and clinicians who use it as a guiding framework for their practice acknowledge that temperament encompasses interactive biological and environmental components (Rothbart & Bates, 1998;Chess & Thomas, 1984). The complexity of these transactions is reflected in the measurement of temperament. Although a variety of methods exist, such as interviews and observational techniques, the most frequently used is Likert-type questionnaires because they are easy to administer and low in cost.
When gathering data on child temperament, parents, because they have multiple opportunities to observe their children in a variety of situations, are often asked to be informants. The reliability and validity of parental reports, however, remains a source of debate for several reasons (Rothbart & Bates, 1998). According to Kagan (1998), parental perceptions of their children's temperament can be biased by their own personality and behavioral expectations. Parents may also have a limited knowledge of normative growth and development on which to compare their child with others. Moreover, the expression of temperament and respondents' perception of it, are also influenced by age, ethnicity, race, sex, and socioeconomic status (Kohnstamm, 1989). Thus, temperament, although stable, is not static. Instead, it is influenced over time by the environment and by biological maturation as various genes are activated.
Parental reports of temperament are also controversial, because the questionnaires have been developed primarily with homogeneous samples of White, non-Hispanic, middle class families from the United States (US). In addition, many of the instruments have not undergone extensive psychometric evaluation. Instead, many researchers have assumed that questionnaires, once developed, are valid and can be used reliably in their studies. Nunnally and Bernstein (1994), however, assert that instrumentation validation is an unending process.
Large epidemiological cross-cultural samples with representative samples are needed to thoroughly explore the many compounded variables that potentially influence parental reports of child temperament. The principal disadvantages of such studies are the cost and time involved. A more modest approach is to examine whether the psychometric estimates of temperament instruments remain stable when completed by samples that differ from the one engaged in the development of the tool. Such studies can be low in cost when conducted on existent data sets. They also can provide additional evidence of reliability and validity of an instrument and focus future investigations. By examining the results in relation to the differences among the samples, hypotheses may be generated regarding the influence of compounding factors that can be best answered with large epidemiologically-based cross-cultural studies.
McClowry (1995) developed the School-Age Temperament Inventory (SATI) with a predominantly White, non-Hispanic middle class sample. The purpose of the present study was to re-examine the reliability and validity of the SATI when completed by three additional samples of parent respondents.