Synonyms

MSEL:AGS; Mullen Scales of Early Learning, American Guidance Service Edition

Abbreviations

ELC:

Early learning composite

Description

The Mullen Scales of Early Learning: American Guidance Service Edition is an individually administered, multidomain measure of early development designed for children from birth to 68 months of age. It was developed by Eileen M. Mullen, Ed.D., and published in 1995. The MSEL:AGS consists of five individual scales: four cognitive scales that cover the age range from birth through 68 months (visual reception, fine motor, receptive language, and expressive language) and one gross motor scale that is administered from birth to 33 months. The four cognitive scales can be combined to produce an Early Learning Composite (ELC) that is said to represent “g” or “general intelligence.”

The MSEL:AGS was based on the author’s information processing and neurodevelopmental theories that conceptualize intelligence “as a network of interrelated but functionally distinct cognitive skills” (Manual, p. 1). By assessing these individual areas of development independently, the examiner can identify specific areas of strength and weakness. From the author’s perspective, this developmental profile can be used to aid in making diagnoses and also in developing strategies for individualized intervention plans (Manual, p. 33). Brief descriptions of each scale are included below.

The visual reception scale consists of 33 items. It assesses the child’s processing of visually presented information and includes tasks that require visual discrimination, categorization, and visual memory skills. While the scale is not completely “nonverbal,” it was designed to minimize the amount of language needed to understand instructions and give a response. Items range from visually fixating and tracking moving objects in early infancy to matching shapes, pictures, letters, and words and remembering visual forms for older children.

The fine motor scale consists of 30 items and (similar to visual reception) also requires minimal receptive and expressive language skills. Fine motor items measure visual-motor planning and control, motor imitation, unilateral and bilateral manipulation of objects, and writing readiness skills. Individual items range from reflexively holding objects and bringing fist to mouth to stacking blocks, cutting with scissors, and copying shapes and letters with a pencil.

The receptive language scale consists of 33 items. This area focuses on the child’s ability to process linguistic input and assesses auditory comprehension and auditory memory skills. At the earliest age levels, the items assess the infant’s orienting to sounds and responding to voices. As the item difficulty increases, the young child is asked to follow verbal directions and respond to questions about spatial, color, size, length, and number concepts. Most items can be answered by pointing to a picture or handing over an object, but some of the more advanced items require brief spoken responses, such as answering general knowledge questions.

The expressive language scale consists of 28 items and involves the production of sounds/words and the use of auditory memory. Items range from vocalizing early vowel and cooing sounds in infancy to defining vocabulary words, repeating sentences, and solving verbal analogies.

The gross motor scale consists of 35 items. The manual states that this scale measures “central motor control and mobility” (p. 2) in a variety of positions that range from being held in the caregiver’s arms to the child being fully upright and independently walking, hopping, and running.

Time required for administration of the full MSEL:AGS varies by the child’s chronological age and ability level. The manual estimates administration takes approximately 15 min for 1-year-olds, 30 min for 4-year-olds, and 60 min for 5-year-olds. Directions include information on positioning infants up to 8–10 months of age and supplemental drawings. Some items can be scored based on parent report and therefore do not have to be directly observed. This is particularly helpful in expressive language items where it is often difficult to elicit vocalizations or spoken responses from young children in an unfamiliar setting. It is also permissible to vary the order of administration of both the scales and individual items within each scale and to readminister some items after a child has “warmed up.” For children born prematurely (at or before 36 weeks gestation), scores may be calculated based on the child’s corrected age up to 2 years.

The components of a complete MSEL:AGS include the Mullen test kit, stimulus booklet, record forms, and two manuals. The test kit contains most of the objects needed for item administration. However, some of these items lack the durability needed for young children (e.g., picture cards), and others present a potential choking hazard (e.g., small beads). The examiner must also supply a number of items not contained in the test kit, such as paper, cereal, large ball, colored tape or white chalk line, 6-in.-high bench, and stairs. Consequently, some preparation time is needed before each administration to insure that all the necessary materials are available.

The stimulus booklet contains black and white line drawings, primarily for use on visual matching, prewriting, receptive language, and expressive vocabulary tasks. On the record form, items are arranged in increasing order of difficulty for each domain. Arrows denote starting points for different ages, but examiners may use their judgment about which starting point to choose (e.g., the examiner may begin at an earlier start point for children with developmental delays). Brief descriptions are included for each item, along with cues for positioning infants and a scoring column. The MSEL:AGS Item Administration Book provides instructions for giving items in all five scales. For each item, this manual includes a list of needed materials, the recommended position for the younger infants (e.g., supported sitting, standing), administration directions, and scoring criteria. The instructions are generally understandable, particularly for professionals with experience in infancy and early childhood. The MSEL:AGS Manual provides information about the test’s theoretical background, scale descriptions, general administration, scoring procedures, standardization and psychometric data, and raw score conversion tables. There are also three case studies provided to illustrate the use and interpretation of the MSEL:AGS.

Scoring for each scale varies from 1 point for a correct response to 0 points for an incorrect response on some items – to items where multiple points can be earned for correctly completing component parts. Basal and ceiling rules are the same for each scale. A basal has been established for a scale when a child earns at least 1 point for each of three consecutive items beginning at a starting point. Testing continues until the child obtains scores of 0 points on three consecutive items. The sums of raw scores for each scale are used to compute derived scores that include T scores (mean = 50; SD = 10), percentile ranks, and age equivalents for each motor and cognitive scale and an Early Learning Composite (mean = 100; SD = 15) based on the four cognitive scales. Descriptive categories are also available (e.g., very high to very low).

Historical Background

The Mullen Scales of Early Learning: AGS Edition (1995) represents a revision of the original Mullen Scales (1992) and combines the Infant MSEL (1989) and the Preschool MSEL (1992) into a single test with norms that span the age range from birth through 68 months. The author, Eileen M. Mullen, Ed.D., developed the MSEL:AGS based on 30 years of clinical experience testing young children with developmental disabilities. She designed it to include multiple developmental domains based on the assumption that specific abilities mature at different rates in infants and young children. Viewed from this perspective, a test yielding only a global score would not provide information about patterns of cognitive strengths and weaknesses that can be used to plan interventions (Manual, p. 9). To arrive at the item sets for the current edition, item analyses were conducted using the Rasch one-parameter IRT model in order to obtain item difficulty estimates and indices of goodness of fit. A few items were dropped from the earlier versions, and the remaining items were sequenced by order of difficulty for the MSEL:AGS.

The structure of the individual scales was built on the author’s theory of information processing where gross motor learning is considered to be the foundation for conceptual development in both visual and auditory modalities (Manual, p. 7). Tasks from each of the four cognitive scales were analyzed in terms of the component processes required to perform them. Individual items in each scale are purported to target intrasensory (auditory or visual) processing or intersensory (auditory-visual) processing. More specifically, the visual reception and fine motor scales are designed to assess visual intrasensory processing, while the receptive language and expressive language scales are designed to assess auditory intrasensory processing. In addition, approximately half of the items on the receptive language scale are reported to assess auditory-visual intersensory processing. However, when one carefully analyzes the individual items, it is difficult to categorize them into purely one type of sensory processing since problem solutions often require multiple skills across multiple modalities. For example, solving the form-board task on the visual reception scale involves understanding the spoken directions (auditory), matching the shapes to their recesses (visual), using fine motor control to place the shapes into their respective holes (visual-motor), and so on.

The MSEL:AGS is the second most widely used measure of early development after the Bayley Scales of Infant and Toddler Development, Third Edition (2006) (Chawarska & Bearss, 2008, p. 55). Before the Bayley III (2006) was redesigned to include cognitive, language (receptive and expressive communication), and motor (fine and gross motor) scales to replace the previous more global mental and motor scales, the Mullen Scales was frequently chosen for research studies because investigators wanted to be able to assess individual developmental domains (Klin, Saulnier, Tsatsani, & Volkmar, 2005). In addition, the relatively long age span (birth to 68 months) was very helpful for longitudinal projects (e.g., studies of the development of baby siblings of children with autism spectrum disorders).

Psychometric Data

The Mullen Scales was standardized on a sample of 1,849 children ranging in age from 2 days to 69 months and grouped into 16 age groups ranging from 2-month age intervals from birth to 14 months to 5-month age intervals from 15 to 68 months. However, there were uneven numbers of children across the 16 age groups, varying from a low of 84 children at 11–12 months to a high of 156 children at 27–32 months. Standardization utilized 71 clinicians from a variety of disciplines and took place at more than 100 sites across the United States. It extended over an 8-year period and was conducted in two phases that covered different time periods and included different geographic regions. Phase 1 spanned June 1981 to February 1986 and included children from the Northeast. Phase 2 covered the period between December 1987 and April 1989 and included samples from the South, West, and North, and South Central areas of the United States. The standardization sample is described as “geographically diverse,” but it is not representative of the entire United States. For example, no children in the 1–14-month age range were recruited from the North/South Central region (Bradley-Johnson, 1997). For the total sample, there were 48.7% females and 51.3% males which were very close to the US population estimates in 1990. However, there was some variability among the 16 age groups, e.g., males were overrepresented at 5–6 months (59.8%) and underrepresented at 27–32 months (43.6%). Parental consent and information on demographic variables (age, sex, race/ethnicity, father’s occupation, and urban/rural residence) were obtained for all children in the sample. Settings for testing included kindergarten, day care, nursery, and homes. All children in the sample were from homes where English was the primary language. It should also be noted that children with known physical or mental disabilities were excluded from the standardization sample.

Chapter 6 of the manual provides information about the reliability and validity of the MSEL:AGS. In terms of reliability, information is provided about internal consistency, test-retest reliability, and interscorer reliability. As a measure of internal consistency, the manual presents split-half reliability coefficients for the 16 age groups of the standardization sample (Manual, p. 56). Modified split-half internal consistency coefficients were calculated for the five scales and the Early Learning Composite. The median values ranged from.75 to.83 for the scales and.91 for the ELC. However, four coefficients that were significantly lower than this range were omitted from the estimation of the median split-half coefficients. The author explained these omissions as being due to a ceiling effect on visual reception for the two oldest age groups and floor and ceiling effects on receptive language for the youngest and oldest age groups. With these same four age groups omitted, the standard error of measurement had medians ranging from 4.1 to 5.0 T score points for the scales and 4.5 standard score points for the ELC.

Test-retest reliability was reported in the manual for two samples based on administration of the original Mullen Scales of Early Learning. The first sample included 50 children from 1 to 24 months of age and the second sample included 47 children ranging from 25 to 56 months. The retest interval ranged from 1 to 2 weeks, with a mean test-retest interval of 11 days. For the younger sample, test-retest reliability was high for gross motor (.96) and ranged from.82 to.85 for the cognitive scales. For the older sample (which did not include children from 57 to 68 months), the median stability coefficients ranged from.71 to.79. Interscorer reliability was reported to be high (.91 to.99), but the number of scorers was not reported, and the sample only went up to 44 months.

The manual provides information about construct, criterion-related, and concurrent validity. Support for construct validity is provided by showing an increase in mean raw scores with increasing chronological age across all five scales. Intercorrelations and squared correlations were calculated for the five Mullen Scales for 5 age groups and the total standardization sample. Scales correlated with each other at low to moderate levels. This was interpreted as indicating both unique and shared variances. Exploratory factor analyses were also conducted for the five scales by 5 age groups. Only one factor was reportedly extracted at each age level and this was interpreted as supporting the Mullen ELC as a measure of “g” or general cognitive ability. However, establishing construct validity for this measure is problematic because there is a lack of supportive data for the author’s theory about intrasensory and intersensory processing leading to the identification of the child’s learning style and its subsequent implications for remediation (Bradley-Johnson, 1997).

Evidence for concurrent validity is based on three types of studies: (1) studies analyzed by the publisher correlating the MSEL:AGS with other measures such as the Bayley Scales of Infant Development (1969), Preschool Language Assessment (1979), and Peabody Developmental Motor Scales (1983), (2) studies reported in the manuals for the original Infant and Preschool Mullen Scales, and (3) studies reported in the research literature which primarily focused on differences between low-birth-weight and normal-birth-weight children. The MSEL:AGS correlated moderately well with the measures mentioned above – with higher correlations obtained when the specific Mullen scale more closely matched what the other instrument was purported to measure. For example, the receptive language scale on the Mullen correlated more highly with Auditory Comprehension on the Preschool Language Assessment (1979) than with the Bayley Scales (1969) Mental Development Index (that includes both verbal and nonverbal items).

Clinical Uses

The MSEL:AGS was designed for use by early childhood professionals working in a variety of settings, such as early intervention programs, preschool and kindergarten special education, universities, research laboratories, hospitals, rehabilitation centers, and private practice. The author described three major uses: (1) eligibility determination for early intervention and special education, (2) diagnostic assessment of children with special needs, and (3) individualized program planning (Manual, p. 5–6).

Under the Individuals with Disabilities Education Act (IDEA), infants and young children must meet specific eligibility criteria in order to qualify for early intervention or special education services. The MSEL:AGS includes three of the five areas of functioning that must be assessed (i.e., physical (including gross and fine motor skills), cognition, and communication), but additional assessment with another measure is needed to cover personal-social and adaptive/self-help skills.

Because the MSEL:AGS includes five individual scales, it is useful in diagnostic assessments where clinicians are trying to determine whether children are progressing within the typical range of development or whether they are demonstrating significant delays in one or more areas of development. Determining a profile of strengths and weaknesses is essential in diagnostic assessments of children suspected to have a wide variety of disabilities, such as specific speech-language impairments, motor impairments (e.g., cerebral palsy), autism spectrum disorders, and global developmental delays. Clinicians also use patterns of performance to formulate recommendations for skills to address in intervention. However, as noted in the validity discussion above, the modality-specific approach to individualized program planning proposed in the MSEL:AGS has not been substantiated.

In summary, the MSEL:AGS has been widely used in developmental assessments of infants and young children. Both clinicians and researchers have appreciated the multiple domain structure and relatively broad age range. Although it is regularly used for diagnostic evaluations and eligibility determination for services, children with physical and mental disabilities were excluded from the standardization sample. In addition, much of the data used for standardization was gathered in the 1980s, so the norms are outdated, and there are no current plans to renorm this measure.