de Morton Mobility Index

Abstract

The de Morton Mobility Index (DEMMI) is an objectively administered, performance-based clinimetric instrument designed to quantify the complete spectrum of mobility in older adults across acute, subacute, and community healthcare settings. Developed by Dr. Natalie A. de Morton and colleagues in 2007 through the rigorous application of Item Response Theory (IRT) and Rasch analysis, the DEMMI resolves longstanding psychometric deficiencies present in legacy functional assessments, such as the Barthel Index, the Functional Independence Measure (FIM), and the Timed Up and Go (TUG) test. These legacy instruments frequently suffer from significant floor and ceiling effects, ordinal scaling constraints, and poor sensitivity to subtle clinical changes. Composed of 15 hierarchically ordered performance items, the DEMMI evaluates functional motor capacity across five fundamental functional domains: bed mobility, chair transfers, static balance, ambulation, and dynamic balance.

Each item is observed and scored dichotomously (0/1) or trichotomously (0/1/2), producing an initial raw score ranging from 0 to 19. This raw score is systematically transformed via an empirically derived Rasch conversion algorithm into an interval-level latent mobility metric scaling from 0 (denoting total bedbound immobility) to 100 (denoting advanced, dynamic, and unassisted independent mobility). Extensive cross-validation across geriatric medical wards, orthopedic post-operative units, neurological rehabilitation clinics, and community ambulatory centers demonstrates robust psychometric properties: high inter-rater reliability (Intraclass Correlation Coefficient [ICC] > 0.89 to 0.94), minimal standard error of measurement (SEM ~ 4.2 to 4.8 points), a Minimal Clinically Important Difference (MCID) spanning 8 to 10 points, and a unidimensional structure free from differential item functioning (DIF) across demographic and diagnostic subgroups. As an open-access, equipment-minimal tool requiring under 15 minutes to administer, the DEMMI represents an international gold standard in mobility measurement for aging populations.

Keywords

de Morton Mobility Index, DEMMI, mobility assessment, Rasch analysis, older adults, geriatric rehabilitation, psychometrics, physical functioning, acute hospitalization, clinimetrics

Authors

The de Morton Mobility Index was conceptualized, operationalized, and psychometrically validated through a collaborative clinical research initiative directed by experts in physiotherapy, biostatistics, and geriatric clinical epidemiology:

• Natalie A. de Morton, PhD, BAppSc (Phty): Lead investigator and primary developer of the DEMMI. Dr. de Morton is an internationally recognized musculoskeletal and geriatric physiotherapy researcher based in Melbourne, Australia. Her programmatic research within the School of Physiotherapy at La Trobe University and subsequently at Southern Health (Monash Health), Eastern Health, and Cabrini Health focused on refining physical performance measurement in hospitalized older adults using modern psychometric methods.
• Megan Davidson, PhD, MAppSc, BAppSc (Phty): Associate Professor within the School of Physiotherapy and Faculty of Health Sciences at La Trobe University, Melbourne, Australia. Dr. Davidson specializes in health status measurement, outcome assessment design, and clinimetric validation in rehabilitation populations.
• Jennifer L. Keating, PhD, PGDipAdvManTher, BAppSc (Phty): Professor of Physiotherapy and Allied Health Research at the Department of Physiotherapy, School of Primary Health Care, Faculty of Medicine, Nursing and Health Sciences, Monash University, Victoria, Australia. Professor Keating has authored landmark texts on clinical research design, statistical methodology, and physical rehabilitation evidence-based practice.
• Dutch Translation and Cross-Cultural Validation Authors (2011): M. P. Jans, V. C. Slootweg, C. R. L. Boot, N. A. de Morton, G. van der Sluis, and N. L. U. van Meeteren (TNO Quality of Life, Leiden; VU University Medical Center, Amsterdam; and Academic Medical Center, Amsterdam, The Netherlands).

Purpose

The primary purpose of the de Morton Mobility Index is to deliver a unidimensional, interval-level, and clinically feasible instrument capable of measuring the full spectrum of physical mobility in older adults across the continuum of care. Historically, clinicians and clinical trialists relied heavily on composite activities of daily living (ADL) scales, such as the Barthel Index or the motor subscale of the Functional Independence Measure (FIM), to approximate patient mobility. While these legacy tools provide utilitarian summaries of general functional dependency, they conflate true physiological mobility with non-mobility tasks such as bowel/bladder continence, grooming, feeding, and dressing. Furthermore, common performance tests such as the Berg Balance Scale or the 10-Meter Walk Test exhibit substantial floor effects among acutely hospitalized, bedridden older adults who are unable to stand or walk, as well as distinct ceiling effects among high-functioning older adults returning to independent community living.

The DEMMI addresses these clinical and methodological limitations by establishing an evidence-based measurement bridge between bedbound dependency and advanced balance locomotion. It was engineered specifically to:

• Accurately quantify motor ability in acutely unwell older individuals admitted to hospital medical and surgical wards, where baseline mobility is frequently compromised by acute systemic illness, frailty, cognitive decline, or deconditioning.
• Track longitudinal recovery and functional trajectory with high sensitivity across diverse settings, including acute inpatient wards, subacute geriatric evaluation and management (GEM) units, inpatient rehabilitation facilities, residential aged-care communities, and outpatient physical therapy practices.
• Eliminate classical floor and ceiling artifacts, ensuring that patients with very low functional status (e.g., restricted to supine bridging or rolling) and patients with high functional status (e.g., performing dynamic multi-directional walking and jumping) are mapped onto the exact same linear continuum.
• Provide a statistically robust outcome measure for clinical trials and observational health services research by transforming ordinal performance ratings into a linear interval-scale metric (0–100) compliant with fundamental mathematical assumptions for parametric statistical hypothesis testing.
• Facilitate multidisciplinary discharge planning, target rehabilitation interventions, determine equipment prescription needs (e.g., gait aids), and inform safe community re-entry.

Psychological Construct

The construct measured by the DEMMI is physical mobility, defined in alignment with the World Health Organization’s International Classification of Functioning, Disability and Health (ICF) within the domain of ‘Activities and Participation’ (specifically under Chapter 4: Mobility, codes d410–d469: changing and maintaining body position, transferring oneself, and walking). Within the psychometric and physiological paradigm of the DEMMI, mobility is conceptualized as an integrated, latent motor capacity requiring the coordinated execution of musculoskeletal force generation, central neurological motor planning, vestibular and visual balance processing, biomechanical equilibrium, and postural adaptation.

Rather than treating functional movement as an arbitrary constellation of isolated tasks, the DEMMI operationalizes mobility as a unidimensional hierarchical continuum consisting of five interrelated functional tiers:

1. Bed Mobility

Bed mobility represents the foundational baseline of motor function. It measures an individual’s biomechanical capacity to alter position within a horizontal, gravity-supported recumbent plane. Tasks include supine pelvic elevation (bridging) and axial rotational torque generation (rolling onto one’s side). Successful execution requires coordinated trunk and pelvic core activation, gluteal muscular recruitment, and lower-extremity stabilizing force. In critically ill, frail, or immediate post-operative populations, bed mobility serves as the primary gateway to preventing immobility-related complications (e.g., pressure injuries, atelectasis, deep vein thrombosis).

2. Chair Transfers and Positional Transitions

This domain captures the transitional biomechanics between recumbent, seated, and upright postures. It measures the neuromuscular capacity to control center-of-mass translation against gravitational pull, including transitioning from a supine posture to an upright sitting posture at the bed’s edge, sustaining independent unsupported seated equilibrium, and rising from a seated position to a standing position (both with and without upper-extremity mechanical assistance). These activities require coordinated hip flexor activation, dynamic abdominal stabilization, quadriceps and gluteal concentric contraction, and anticipatory postural adjustments.

3. Static Balance

Static balance measures the sensorimotor integration necessary to maintain postural equilibrium over a progressively narrowed base of support without active ambulation. This domain evaluates closed-loop neuromuscular control during unsupported quiet standing, standing with the feet placed directly adjacent to one another (narrowed base of support), standing on toes (elevating the center of mass while narrowing foot surface area and engaging triceps surae force), and tandem standing (placing feet in a heel-to-toe linear configuration). These tasks challenge visual, somatosensory, and vestibular afferent pathways and demand rapid motor correction via ankle and hip strategies.

4. Ambulation and Gait Independence

Ambulation examines linear forward locomotor capacity, spatial displacement, and functional independence from external assistive technology. It assesses the patient’s capacity to navigate a standardized 10-meter level pathway. It systematically distinguishes between individuals who require assistive gait aids (such as rollator walkers, quad sticks, or standard canes) to generate base-of-support stability and those who possess the intrinsic musculoskeletal power, dynamic postural control, and neuromuscular synergy required to ambulate completely unassisted.

5. Dynamic Balance and Advanced Locomotion

Dynamic balance forms the apex of the DEMMI measurement hierarchy, assessing high-level multi-planar motor control, postural recovery, and eccentric-concentric explosive power. It includes picking up a pen from the floor (requiring coordinated forward trunk flexion, hip flexion, and vestibular adaptation without losing balance while retrieving a small object), stepping backwards (disrupting standard visual guidance, demanding intact somatic proprioception and reverse step cycling), and jumping (generating simultaneous bilateral lower-extremity propulsion, clearing the ground entirely, and absorbing kinetic ground reaction forces upon landing). These advanced items ensure that higher-functioning geriatric patients do not encounter artificial measurement ceilings.

Theoretical Framework

The architectural foundation of the DEMMI is rooted in Modern Psychometric Theory, specifically the mathematical and philosophical principles of the Rasch Measurement Model (Rasch, 1960). In traditional clinical index development governed by Classical Test Theory (CTT), total test scores are derived by linearly summing arbitrary numerical points assigned to disparate behavioral observations (e.g., 0 = dependent, 1 = partially dependent, 2 = independent). This practice relies on two fundamentally flawed psychometric assumptions: first, that ordinal raw ratings possess equal interval increments across the entire measurement range, and second, that all items contribute identically to the underlying construct regardless of their relative biomechanical difficulty.

The Rasch Measurement Model

Georg Rasch formulated a logistic latent trait model specifying that the probability of a specific patient successfully executing a specific clinical task is a logistic function of the difference between the person’s latent underlying ability ($ heta_n$) and the item’s inherent difficulty ($eta_i$):

$$P(X_{ni} = 1) = \frac{\exp(\theta_n – \beta_i)}{1 + \exp(\theta_n – \beta_i)}$$

When an item possesses polytomous response categories (such as the trichotomous items within the DEMMI), the model extends to the Partial Credit Model (Masters, 1982), wherein each category threshold ($ au_{ik}$) represents the transition point along the latent mobility continuum between consecutive score categories:

$$P(X_{ni} = k) = \frac{\exp \sum_{j=0}^k (\theta_n – \beta_i – \tau_{ij})}{\sum_{m=0}^{m_i} \exp \sum_{j=0}^m (\theta_n – \beta_i – \tau_{ij})}$$

Under this formal paradigm, several core axioms must be empirically satisfied:

• Unidimensionality: All items in the instrument must measure one, and only one, common latent construct—here, physical mobility. Performance variance cannot be driven by secondary orthogonal dimensions such as speech fluency, vision, or cognitive comprehension.
• Item Invariance and Local Independence: The calibrated difficulty of the mobility tasks must remain statistically invariant across patient samples, regardless of whether the assessment occurs in an orthopedic surgery ward, a stroke recovery unit, or an outpatient falls clinic. Furthermore, when controlling for the patient’s latent mobility ability, the response to one item must be statistically independent of the response to any other item.
• Interval Scaling Transformation: Raw, non-linear, bounded ordinal scores (0 to 19) are converted via log-odds ratios (logits) into a true linear continuous interval scale. This metric is subsequently calibrated to an intuitive 0 to 100 integer scale. On this transformed scale, a 5-point increase at the lower end (e.g., from 15 to 20) represents the exact same quantitative increment in functional mobility as a 5-point increase at the upper end (e.g., from 75 to 80).

Validity

The DEMMI has undergone rigorous empirical validation across diverse clinical specialties, demographic cohorts, and international healthcare environments. Published literature consistently demonstrates exceptional validity metrics:

Construct and Structural Validity

During the primary developmental validation studies conducted by de Morton et al. (2006, 2008) across acute medical wards, Rasch modeling confirmed that all 15 items demonstrated acceptable goodness-of-fit to the latent trait. Item infit and outfit mean-square (MnSq) statistics fell comfortably within the recommended psychometric bounds of 0.7 to 1.3, verifying that the items define a coherent, unidimensional hierarchy. Subsequent cross-cultural validation studies, including the Dutch validation by Jans et al. (2011) and German validations in subacute geriatric and orthopedic populations, confirmed that the hierarchical ordering of item difficulty remained perfectly stable across languages and clinical cultures.

Convergent and Criterion Validity

Convergent validity has been established by correlating DEMMI interval scores against established mobility, balance, and independence instruments:

• Barthel Index (BI): High-magnitude positive correlation ($r = 0.68$ to $0.84$), with the DEMMI exhibiting superior discrimination at both low and high extremes where the Barthel Index reaches floor and ceiling thresholds.
• Hierarchical Assessment of Balance and Mobility (HABAM): Strong positive correlation ($r = 0.81$ to $0.91$).
• Timed Up and Go (TUG): Strong inverse correlation ($r = -0.68$ to $-0.75$), indicating that patients scoring higher on the DEMMI require significantly less time to complete the functional stand-walk-turn-sit sequence.
• Functional Ambulation Category (FAC): Robust positive association ($
  ho = 0.74$ to $0.82$).
• Motor FIM: High positive correlation ($r = 0.76$ to $0.85$).

Known-Groups and Discriminant Validity

The DEMMI distinguishes reliably between distinct clinical cohorts with divergent functional levels. Studies reveal statistically significant score differences ($p < 0.001$) when comparing:

• Patients discharged home versus those transferred to inpatient subacute rehabilitation or residential aged-care facilities.
• Community-dwelling older adults without a history of falls versus recurrent fallers presenting to emergency departments.
• Patients requiring unassisted versus assisted ambulation devices.

Floor and Ceiling Effects

One of the primary psychometric achievements of the DEMMI is the virtual elimination of floor and ceiling effects within acute hospital and post-acute settings. While the Barthel Index typically exhibits ceiling effects exceeding 25% to 40% upon hospital discharge, and the Timed Up and Go test displays floor effects exceeding 30% upon acute admission (because many patients cannot stand or walk safely without physical assistance), the DEMMI maintains floor and ceiling effects below 2% to 5% across diverse inpatient populations.

Responsiveness and Minimal Clinically Important Difference (MCID)

The DEMMI displays exceptional sensitivity to longitudinal change during acute and subacute recovery. Effect size statistics (Cohen’s $d$) frequently exceed 0.80 to 1.15 in cohorts receiving active inpatient physiotherapy. Receiver Operating Characteristic (ROC) curve analyses and anchor-based estimation studies demonstrate that:

• The Minimal Clinically Important Difference (MCID) is approximately 8.0 to 10.0 points on the 0–100 interval scale. A change of 10 points signifies an unambiguous, clinically visible functional gain (e.g., progressing from requiring a walker with assistance to walking independently without aids).
• The Minimal Detectable Change at the 90% confidence level ($MDC_{90}$) is 8.9 points, while the $MDC_{95}$ is 9.7 to 10.4 points, confirming that an observed change of 10 points represents real functional progress beyond measurement noise.

Reliability

The DEMMI has demonstrated excellent reliability coefficients in repeated empirical trials conducted by both physical therapists and trained interdisciplinary healthcare professionals:

Inter-Rater Reliability

Because the DEMMI uses objective, standardized scoring criteria and unambiguous operational definitions, inter-observer variation is remarkably low. In the foundational validation studies by de Morton et al. (2006, 2008), paired physiotherapists scoring acute geriatric patients concurrently or sequentially within a two-hour window yielded an overall Intraclass Correlation Coefficient (ICC, model 2,1) of 0.89 to 0.94 (95% CI: 0.86–0.96). Subsequent studies in inpatient stroke rehabilitation and subacute geriatric evaluation units corroborated this finding, reporting inter-rater ICCs ranging between 0.91 and 0.96.

Test-Retest and Intra-Rater Reliability

Intra-rater test-retest reliability across brief temporal windows (24 hours or less, ensuring clinical stability before substantial therapeutic changes occur) has repeatedly produced ICC values between 0.86 and 0.93. Individual item kappa statistics ($kappa$) for categorical scoring agreement range from 0.65 to 0.92, reflecting substantial to almost perfect inter-rater concordance on specific items.

Standard Error of Measurement (SEM)

The Standard Error of Measurement (SEM) for the DEMMI across diverse acute and subacute clinical trials is consistently calculated at 4.2 to 4.8 points on the 100-point interval scale. Given this low error variance, clinicians can monitor individual trajectory changes with high statistical certainty, avoiding misclassification during rehabilitation discharge or transfer determinations.

Internal Consistency and Person Separation

Within the Rasch measurement paradigm, traditional Cronbach’s alpha coefficients are replaced by the Person Separation Index (PSI). Across acute and subacute cohorts, the DEMMI consistently demonstrates PSI values between 0.85 and 0.89, with corresponding Cronbach’s alpha equivalents exceeding 0.88. This establishes that the 15 items have high internal consistency and successfully stratify patient cohorts into at least three to four distinct functional mobility strata.

Factor Analysis

The internal structural architecture of the DEMMI was evaluated through both exploratory and confirmatory Rasch-based diagnostic methods rather than solely classical common factor analysis, given the ordinal, non-normal properties of performance assessment data.

Unidimensionality and Residual Analysis

A primary criterion of the Rasch model is strict unidimensionality. To confirm this assumption, a Principal Component Analysis (PCA) of the standardized Rasch residuals was conducted. The empirical findings demonstrated that:

• The primary Rasch latent dimension explained over 65% to 72% of the total variance in observed physical performance.
• The unexplained variance accounted for by the first contrast (the first residual component) had an eigenvalue of less than 1.8 to 2.0, well below the established threshold of 2.0 that would suggest the presence of a secondary orthogonal dimension.
• Item residual correlations were consistently below 0.20, confirming that the assumption of local item independence was robustly upheld.

Item Fit Statistics

Item fit was evaluated using information-weighted mean square (Infit MnSq) and outlier-sensitive mean square (Outfit MnSq) indices:

• All 15 items exhibited Infit and Outfit MnSq statistics within the range of 0.75 to 1.25 (well within the acceptable clinical parameter of 0.70 to 1.30).
• Standardized $Z$ scores (ZSTD) for item fit were contained between $-1.9$ and $+1.9$, demonstrating that no items suffered from significant underfit (which degrades measurement validity via unmodeled noise) or overfit (which introduces item redundancy).

Differential Item Functioning (DIF)

An essential psychometric requirement of the DEMMI was the absence of measurement bias across patient subgroups. Extensive Rasch differential item functioning (DIF) analyses were conducted across multiple demographic and clinical covariates:

• Sex: No statistically significant uniform or non-uniform DIF was detected between male and female participants across any of the 15 items ($p > 0.05$).
• Age Stratification: When comparing individuals aged 65–79 years to individuals aged 80 years and older, item calibrations remained stable, confirming age-neutral measurement.
• Diagnostic Classification: Items exhibited invariant measurement properties regardless of whether the primary medical etiology was acute cardiovascular, respiratory, orthopedic, or general medical illness.
• Cognitive Status: DIF evaluation across patients with intact cognition versus those with mild-to-moderate cognitive impairment (Mini-Mental State Examination [MMSE] < 24) showed no systematic bias, confirming the DEMMI’s validity in patients with mild dementia or delirium who can follow simple motor commands.

Instrument / Measurement Tool

• Type of Instrument: Performance-based, direct-observation clinimetric outcome measure.
• Target Population: Older adults across acute hospital wards, subacute geriatric rehabilitation, residential care, and community ambulatory centers.
• Administration Time: Approximately 10 to 15 minutes.
• Required Equipment:
  • Standard hospital or examination bed with an adjustable height mechanism.
  • Standard sturdy chair with armrests (seat height ~45 cm).
  • Stopwatch or digital timer.
  • Everyday object (standard ballpoint pen).
  • Clear, unobstructed 10-meter walking track with floor distance markers.
  • Patient’s customary walking aid (cane, crutches, or walker, if applicable).
• Number of Items: 15 hierarchically ordered performance items.
• Response Scale and Scoring Architecture:
  • Hierarchical scoring per item (Dichotomous 0/1 or Trichotomous 0/1/2): Bed items (0 = unable, 1 = able); Chair items (0 = unable, 1 = able; or 0 = unable, 1 = with arms, 2 = without arms); Static balance items (0 = unable, 1 = able <10s, 2 = 10s); Walking items (distance: 0 = <10m, 1 = 10m; aid: 0 = with aid or unable, 1 = without aid); Dynamic balance items (0 = unable, 1 = able). Total raw score ranges from 0 to 19, converted using Rasch analysis to a 0 to 100 interval score.
• Raw-to-Interval Conversion Table (Rasch Logit Calibration):
  • Raw Score 0 = DEMMI Score 0
  • Raw Score 1 = DEMMI Score 8
  • Raw Score 2 = DEMMI Score 15
  • Raw Score 3 = DEMMI Score 20
  • Raw Score 4 = DEMMI Score 24
  • Raw Score 5 = DEMMI Score 27
  • Raw Score 6 = DEMMI Score 30
  • Raw Score 7 = DEMMI Score 33
  • Raw Score 8 = DEMMI Score 36
  • Raw Score 9 = DEMMI Score 39
  • Raw Score 10 = DEMMI Score 41
  • Raw Score 11 = DEMMI Score 44
  • Raw Score 12 = DEMMI Score 48
  • Raw Score 13 = DEMMI Score 53
  • Raw Score 14 = DEMMI Score 57
  • Raw Score 15 = DEMMI Score 62
  • Raw Score 16 = DEMMI Score 67
  • Raw Score 17 = DEMMI Score 74
  • Raw Score 18 = DEMMI Score 85
  • Raw Score 19 = DEMMI Score 100
• Clinical Safety and Termination Rules: The test is discontinued immediately if the patient displays signs of acute physiological distress (severe dyspnea, chest pain, dizziness, diaphoresis) or if the examining clinician determines that proceeding would compromise patient safety.

Permissions & Fee and Test Year

The de Morton Mobility Index was developed and first published in 2007 (with extensive developmental studies published between 2006 and 2008) by Dr. Natalie A. de Morton and associates at La Trobe University and Southern Health, Melbourne, Australia. The instrument was deliberately established as an open-access, public-domain clinical tool to facilitate widespread international adoption across physiotherapy, nursing, geriatric medicine, and health services research.

Fee and Licensing: There are no licensing fees, royalties, or costs associated with the clinical, educational, or academic research utilization of the DEMMI. The scoring sheets, manual, and Rasch conversion charts may be freely reproduced and incorporated into electronic medical record (EMR) systems, provided that proper bibliographic citation is maintained and the authentic 15-item hierarchy and Rasch conversion metrics are not modified without authorization.

References

• de Morton, N. A., Davidson, M., & Keating, J. L. (2006). The de Morton Mobility Index (DEMMI) was valid and reliable for measuring mobility in acutely ill older medical patients. Journal of Clinical Epidemiology, 59(10), 1085–1093. https://doi.org/10.1016/j.jclinepi.2006.01.011
• de Morton, N. A., Davidson, M., & Keating, J. L. (2007). Validity, responsiveness and the minimal clinically important difference for the de Morton Mobility Index (DEMMI) in an older rehabilitation population. Australasian Journal on Ageing, 26(suppl 1), 60–61.
• de Morton, N. A., Davidson, M., & Keating, J. L. (2008). The de Morton Mobility Index (DEMMI): An essential health index for an ageing world. Health and Quality of Life Outcomes, 6, Article 63. https://doi.org/10.1186/1477-7525-6-63
• de Morton, N. A., Davidson, M., & Keating, J. L. (2010). Rasch analysis of the Barthel Index in the assessment of hospitalized older patients after admission for an acute medical condition. Archives of Physical Medicine and Rehabilitation, 91(1), 38–44. https://doi.org/10.1016/j.apmr.2009.08.148
• de Morton, N. A., Davidson, M., & Keating, J. L. (2011). Reliability of the de Morton Mobility Index (DEMMI) in an older acute medical population. Physiotherapy Research International, 16(3), 159–169. https://doi.org/10.1002/pri.486
• Jans, M. P., Slootweg, V. C., Boot, C. R. L., de Morton, N. A., van der Sluis, G., & van Meeteren, N. L. U. (2011). Reproducibility and validity of the Dutch translation of the de Morton Mobility Index (DEMMI) in older patients admitted to an acute medical ward. Physiotherapy Research International, 16(4), 200–210. https://doi.org/10.1002/pri.490
• Masters, G. N. (1982). A Rasch model for partial credit scoring. Psychometrika, 47(2), 149–174. https://doi.org/10.1007/BF02296272
• Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Copenhagen: Danish Institute for Educational Research.

Items of the Scale