Retrospective Measure of Behavioural Inhibition (RMBI)

Abstract

The Retrospective Measure of Behavioural Inhibition (RMBI) is an 18-item psychometric self-report questionnaire designed to assess childhood behavioural inhibition (BI) in adult populations retrospectively. Developed by Australian clinical psychologists and psychiatrists Gordon Parker and Gillian L. Gladstone at the University of New South Wales and the Black Dog Institute, the RMBI captures the temperament disposition originally conceptualized by developmental psychologist Jerome Kagan. Behavioural inhibition is characterized by heightened physiological reactivity, reticence, social withdrawal, and fearful non-approach behaviors in response to unfamiliar persons, environments, or novel stimuli during childhood (specifically ages 2 through 12). The RMBI measures phenotypic markers across two primary, empirically supported subdimensions: Social Inhibition / Low Sociability (fear of negative evaluation, reticence in peer groups, performance anxiety) and Fearful / Non-Approach Novelty Avoidance (clinging to caregivers, freezing, avoidance of novel activities, and hesitation in unfamiliar environments). Respondents evaluate each statement using a 3-point frequency scale (“no/hardly ever” = 0, “some of the time” = 1, “yes/most of the time” = 2), alongside an explicit non-scored option for items where the participant has no memory (“Do not remember at all”). Across non-clinical, clinical, and twin populations, the RMBI exhibits robust internal consistency (Cronbach’s α ranging between .84 and .89), high test-retest stability ($r > .80$), and confirmed factorial validity through both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Crucially, the RMBI demonstrates strong convergent and predictive validity, significantly correlating with current measures of adult behavioural inhibition (e.g., the Adult Measure of Behavioural Inhibition [AMBI]), trait neuroticism, introversion, and lifetime diagnoses of affective disorders, particularly social anxiety disorder (SAD) and major depressive disorder. It remains a standard benchmark in developmental psychopathology, psychiatric genetics, and longitudinal personality assessment.

Keywords

Retrospective Measure of Behavioural Inhibition, RMBI, Behavioural Inhibition, Temperament, Social Anxiety Disorder, Jerome Kagan, Gordon Parker, Gillian Gladstone, Adult Measure of Behavioural Inhibition, Psychometrics, Factor Analysis, Fearful Inhibition

Authors

The Retrospective Measure of Behavioural Inhibition was developed and validated by:

• Gillian L. Gladstone, Ph.D. — Clinical Psychologist and Research Fellow, School of Psychiatry, Faculty of Medicine, University of New South Wales (UNSW Sydney), and the Black Dog Institute, Prince of Wales Hospital, Randwick, New South Wales, Australia.
• Gordon B. Parker, AO, MBBS, MD, PhD, DSc, FRANZCP — Scientia Professor of Psychiatry, School of Psychiatry, University of New South Wales; Founder and former Executive Director of the Black Dog Institute; Consultant Psychiatrist, Prince of Wales Hospital, Sydney, Australia.

Additional normative and twin-study psychometric validation was conducted in collaboration with Philip B. Mitchell, Kay A. Wilhelm, Gin S. Malhi, and Thao-Vi Nguyen across the Black Dog Institute, the University of New South Wales, and the Australian Twin Registry.

Purpose

The primary clinical and empirical purpose of the Retrospective Measure of Behavioural Inhibition (RMBI) is to provide a reliable, cost-effective, and standardized self-report method for measuring an individual’s childhood temperamental style retrospectively once they have reached adulthood. In developmental psychology and psychiatric epidemiology, longitudinal assessment of behavioural inhibition traditionally necessitated prospective laboratory observations starting in infancy (e.g., Kagan’s infant motor reactivity and fear-induction batteries at 4, 14, and 21 months of age). While laboratory observational batteries remain a gold standard for pediatric research, their resource-intensive nature makes them impractical for large-scale epidemiological studies, adult clinical trials, psychiatric genetic paradigms, and retrospective case-control investigations.

Prior to the establishment of the RMBI, researchers studying adult affective, anxiety, and personality disorders lacked an unconfounded, specific retrospective index of temperamental inhibition. Existing retrospective inventories often conflated childhood shyness with broader traits such as generalized neuroticism, depressive rumination, or avoidant personality traits established in late adolescence. The RMBI addresses this gap by isolating clear behavioral, somatic, and environmental responses characteristic of early-to-middle childhood (specifically prior to puberty). These responses include hiding when unfamiliar visitors enter the home, clinging to caregivers in novel settings, hesitation to participate in group peer play, and somatic-emotional freezing when exposed to new stimuli.

In clinical practice, the RMBI serves an etiologic mapping function. When treating adults presenting with treatment-resistant depression, social anxiety disorder, panic disorder, or agoraphobia, clinicians utilize the RMBI alongside its adult counterpart (the Adult Measure of Behavioural Inhibition [AMBI]) to establish whether the adult pathology reflects an entrenched, neurobiologically underpinned temperamental trajectory (i.e., “trait inhibition”) or a late-onset state resulting from psychosocial stressors. In research settings, the RMBI has been extensively applied in psychiatric genetics and twin studies to estimate the heritability of internalizing traits and to trace how gene-environment interactions transform temperamental vulnerability into overt clinical anxiety and depressive episodes across the lifespan.

Psychological Construct

The psychological construct assessed by the RMBI is Childhood Behavioural Inhibition (BI). First operationalized empirically by Jerome Kagan and colleagues at Harvard University in the 1980s, behavioural inhibition refers to an innate, biologically based temperamental disposition observed in approximately 15% to 20% of healthy children. In the presence of unfamiliar people, objects, events, or social situations, behaviourally inhibited children characteristically show hesitation, motor freezing, withdrawal, wariness, and sustained physiological arousal, accompanied by prolonged latencies to vocalize or approach.

The RMBI conceptualizes behavioural inhibition as a multifaceted construct encompassing distinct behavioral, interpersonal, and autonomic manifestations across childhood. Psychometric and structural evaluations of the instrument delineate two primary core dimensions, with an optional third facet identified across selected exploratory studies:

1. Social Inhibition and Low Sociability

This dimension reflects the child’s chronic avoidance, social reticence, and distress within peer and classroom contexts. Children high on this dimension do not actively explore social environments, exhibit marked difficulty approaching unfamiliar peers on the playground, and experience intense fear regarding performance situations. Exemplars from the RMBI include:

• Standing back and watching other children play at school rather than initiating joint activity (Item 2).
• Finding it exceedingly difficult to approach and play with unfamiliar children (Item 9).
• Marked difficulty speaking or standing up in front of the class (Item 14).
• A subjective appraisal of oneself as an exceptionally “shy child” during early developmental years (Item 12).
• Low levels of extraverted, outgoing, and talkative behavior toward peers (Item 15, reverse-scored).

2. Fearful Non-Approach and Novelty Avoidance

This dimension operationalizes the universal, non-social manifestation of behavioural inhibition: the innate tendency to freeze, withdraw, or seek physical proximity to a secure attachment figure when confronted with environmental novelty or perceived threat. Rather than being restricted to social judgment, this facet taps autonomic and behavioural defense cascades triggered by unfamiliar stimuli, strange adults, or novel settings. Key indicators within the RMBI include:

• Freezing or becoming silent when visiting novel places with family (Item 16).
• Hiding, running out of the room, or clinging tightly to parents when unfamiliar visitors enter the house (Items 1, 6, and 18).
• Remaining physically tethered to parents/caregivers rather than exploring novel environments (Item 10).
• Severe reluctance to attend school on the first day or following holiday breaks (Item 3).
• Fear of domestic animals and other people’s pets (Item 8).

3. Risk Avoidance and Low High-Intensity Pleasure (Activity Modulation)

Often integrating with the non-approach dimension, this facet evaluates the child’s absence of sensation-seeking, risk-taking, and motor exploration. Inhibited children consistently display lower levels of physical risk-taking in sports or playground games (Item 13) and rarely wander off in public spaces (Item 7), reflecting an overactive threat-appraisal system that overrides typical developmental curiosity.

Theoretical Framework

The construction of the RMBI is anchored in the convergence of two major psychological frameworks: Kagan’s Developmental Neurobiology of Temperament and Jeffrey Gray’s Reinforcement Sensitivity Theory (RST).

Kagan’s Neurobiological Model

Jerome Kagan hypothesized that behavioural inhibition is rooted in a constitutional, genetically mediated hyperexcitability of the amygdala and its afferent and efferent projections to the sympathetic nervous system, the hypothalamic-pituitary-adrenal (HPA) axis, and the ventral striatum. When confronted with novel or mildly threatening sensory input, children with an inhibited temperament exhibit lowered activation thresholds within the central nucleus of the amygdala. This triggers an immediate sympathetic surge, evidenced physiologically by:

• Elevated and stable heart rates with diminished heart rate variability (respiratory sinus arrhythmia suppression).
• Elevated baseline morning salivary cortisol and hyper-reactive cortisol spikes following mild laboratory stressors.
• Prolonged pupillary dilation during cognitive challenge.
• Increased muscle tension in the larynx and vocal cords, manifesting as speech latency or mutism.

Gladstone and Parker (2005) designed the RMBI to retrospectively capture the exact behavioral equivalents of these amygdalar-driven responses (e.g., freezing, hiding, somatic apprehension, and refusal to vocalize) during developmental windows where this biological vulnerability is most acute.

Gray’s Reinforcement Sensitivity Theory (BIS/BAS)

The RMBI aligns directly with Jeffrey Gray’s neuropsychological model of personality, particularly the Behavioural Inhibition System (BIS). In Gray’s model, the BIS is an attentional and motivational system mediated by the septohippocampal system and the Papez circuit. It is activated by conditioned stimuli for punishment, novel stimuli, and innate fear triggers. Activation of the BIS produces:

1. The immediate interruption of ongoing motor behavior (“stop, look, and listen” or behavioral freezing).
2. An increase in autonomic arousal to prime defensive action.
3. A dramatic increase in attentional vigilance toward threat-relevant cues in the immediate environment.

Conversely, the Behavioural Approach System (BAS) governs appetitive motivation and reward-seeking. High scorers on the RMBI reflect a phenotypic profile characterized by a hyper-sensitized BIS and a hypo-reactive BAS: they actively eschew high-stimulation environments, avoid novel social engagements, and show virtually no risk-taking behavior during physical play.

Diathesis-Stress Trajectories in Adult Psychopathology

The theoretical framework of the RMBI posits behavioural inhibition not as a psychiatric pathology in itself, but as a robust developmental diathesis. According to the developmental psychopathology model formulated by Gladstone and Parker, when an infant or young child with high behavioural inhibition experiences environmental risks—such as overprotective or rejecting parenting, insecure attachment, or peer victimization—their temperamental hyper-reactivity crystallizes into stable cognitive schemas of personal vulnerability and social threat. Consequently, the RMBI captures the developmental antecedent that predisposes individuals to subsequent Axis I anxiety disorders, specifically Social Anxiety Disorder (SAD), Panic Disorder, Generalized Anxiety Disorder (GAD), and secondary Major Depressive Disorder (MDD).

Validity

Extensive psychometric investigations have established the construct, convergent, discriminant, and predictive validity of the Retrospective Measure of Behavioural Inhibition across diverse clinical and non-clinical populations.

Construct and Factorial Validity

Initial construct validation by Gladstone and Parker (2005) evaluated the instrument in a primary validation cohort comprising 204 adult participants (including university undergraduates and psychiatric outpatients presenting with mood disorders). Both exploratory and confirmatory factor analytic models confirmed that the RMBI measures a coherent construct distinct from simple general distress. Factor models demonstrated clear, robust loadings for items tapping social reticence, novelty wariness, and somatic freezing, confirming that the scale accurately maps onto the theoretical core of Kagan’s childhood BI construct.

Convergent Validity

The convergent validity of the RMBI has been substantiated through significant correlations with validated measures of temperament, personality, and clinical anxiety:

• Adult Behavioural Inhibition: The RMBI demonstrates strong, statistically significant correlations ($r = .55$ to $.68, p < .001$) with the Adult Measure of Behavioural Inhibition (AMBI), confirming temperamental continuity across developmental stages from childhood to adulthood.
• Carver & White’s BIS/BAS Scales: RMBI scores correlate positively with the Carver-White BIS scale ($r = .42$ to $.52$), corroborating its alignment with Gray’s behavioural inhibition system.
• Eysenck Personality Questionnaire (EPQ) & NEO-PI-R: The RMBI exhibits substantial positive correlations with Trait Neuroticism ($r = .40$ to $.48$) and moderate-to-strong negative correlations with Extraversion ($r = -.38$ to $-.51$), reflecting the expected low-sociability and high-vulnerability phenotype.

Discriminant Validity

Discriminant validity analyses show that the RMBI is psychometrically distinguishable from current depressive states and transient situational anxiety. While patients in an acute depressive episode show elevated state-dependent scores on many self-report inventories, longitudinal analyses by Gladstone et al. (2006) revealed that RMBI scores remain stable regardless of fluctuations in current Montgomery-Åsberg Depression Rating Scale (MADRS) or Beck Depression Inventory (BDI) scores, indicating that the retrospective recall of childhood inhibition is not an artifact of state-congruent cognitive memory bias (the “depressive recall bias”).

Predictive and Clinical Validity

In a seminal clinical study investigating lifetime anxiety disorders in a sample of psychiatric outpatients ($N = 312$), Gladstone et al. (2006) established the clinical predictive validity of the RMBI:

• Participants with a lifetime DSM-IV diagnosis of Social Anxiety Disorder (SAD) scored significantly higher on the RMBI ($M = 18.2, SD = 6.4$) than clinical controls without lifetime anxiety disorders ($M = 9.8, SD = 5.1; p < .001, d > 1.20$).
• Multivariate logistic regression revealed that elevated RMBI scores independently predicted a four-fold increase in the lifetime odds of developing social phobia ($ ext{OR} = 4.12, 95% text{ CI } [2.45, 6.94]$), even after controlling for sex, family history of depression, and concurrent neuroticism.
• In a large-scale twin study involving 1,848 adult twins from the Australian Twin Registry, Gladstone, Parker, and Nguyen (2006) utilized the RMBI to demonstrate that genetic factors account for approximately 45% to 52% of the variance in retrospective behavioural inhibition, with non-shared environmental factors explaining the remainder, aligning with heritability estimates derived from prospective infant studies.

Reliability

The RMBI possesses strong psychometric reliability across internal consistency, split-half, and temporal stability assessments:

Internal Consistency

Across validation and replication cohorts, the RMBI exhibits high internal consistency as quantified by Cronbach’s alpha (α):

• In the original instrument development sample ($N = 204$), Gladstone & Parker (2005) reported an overall Cronbach’s α of .88 for the total 18-item scale.
• The subscales exhibited strong internal consistency: the Social Inhibition / Low Sociability subscale yielded α values between .83 and .86, while the Fearful / Non-Approach Novelty Avoidance subscale yielded α values between .78 and .82.
• In the Australian Twin Registry cohort ($N = 1,848$), internal consistency remained robust at α = .85.
• In psychiatric outpatient cohorts presenting with unipolar and bipolar mood disorders, Gladstone et al. (2006) demonstrated an α of .87.

Test-Retest Stability

Temporal stability of the RMBI is essential given its retrospective nature. In a subsample of respondents assessed across an 8- to 12-week interval, the test-retest intraclass correlation coefficient (ICC) was $r = .84$ ($p < .001$), indicating that retrospective adult evaluations of childhood behavioural style remain exceptionally stable over time. Furthermore, longitudinal assessments over a 2-year interval in clinical samples yielded a retest stability of $r = .79$, confirming that recall is minimally distorted by episodic affective illness.

Factor Analysis

The latent structure of the RMBI was evaluated by Gladstone & Parker (2005) using both exploratory and confirmatory factor analysis procedures.

Exploratory Factor Analysis (EFA)

During initial scale development, the 18 items were subjected to principal components analysis (PCA) followed by Promax (oblique) and Varimax (orthogonal) rotations to allow for expected correlations between underlying temperamental dimensions. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy consistently exceeded .85, and Bartlett’s Test of Sphericity was highly significant ($p < .0001$), confirming correlation matrix factorability.

Scree plot inspections and parallel analysis supported a clear two-factor solution accounting for approximately 43.8% of the total variance:

• Factor 1: Social Inhibition and Fearful Hesitation (Variance explained: ~28.6%)
  Items loading heavily on this factor represent reticence in peer groups and fear of novelty in social settings. High-loading items include:
  • Item 9 (“At school, did you find it difficult to approach and play with new children?”) — loading: .78
  • Item 2 (“At school, did you tend to stand back and watch other children play?”) — loading: .74
  • Item 12 (“Do you consider that you were a shy child?”) — loading: .72
  • Item 14 (“Was it difficult for you to stand up in front of the class?”) — loading: .68
  • Item 15 (reversed: “Were you outgoing and talkative with other children?”) — loading: -.63
• Factor 2: Non-Approach, Clinging, and Novelty Freezing (Variance explained: ~15.2%)
  Items loading heavily on this factor evaluate autonomic threat reactions, proximity-seeking, and environmental non-approach. High-loading items include:
  • Item 18 (“When unfamiliar visitors came to your home, would you cling to your mother or father [or caregiver]?”) — loading: .76
  • Item 1 (“When unfamiliar visitors came to your home, would you hide or leave the room?”) — loading: .73
  • Item 6 (“When unfamiliar visitors came to your home, did you feel fearful or nervy?”) — loading: .70
  • Item 10 (“When you went on outings with your family to new places, would you spend most of the time next to your mother or father [or caregiver]?”) — loading: .67
  • Item 16 (“When you went on outings with your family to new places, would you become quiet or ‘freeze up’?”) — loading: .64

Confirmatory Factor Analysis (CFA)

Subsequent confirmatory factor analyses performed in independent community and clinical validation datasets confirmed superior fit indices for an correlated two-factor model over a unidimensional model:

• Chi-Square / Degree of Freedom Ratio ($\chi^2/df$): 1.84 (indicating excellent parsimonious fit)
• Comparative Fit Index (CFI): .94
• Tucker-Lewis Index (TLI): .93
• Root Mean Square Error of Approximation (RMSEA): .051 (90% CI [.042, .060])
• Standardized Root Mean Square Residual (SRMR): .048

The inter-factor correlation between Factor 1 and Factor 2 was estimated at $r = .52$, demonstrating that while these two dimensions capture distinct behavioral facets (social peer reticence vs. caregiver-clinging/novelty avoidance), they stem from a common overarching behavioural inhibition construct.

Instrument / Measurement Tool

• Instrument Name: Retrospective Measure of Behavioural Inhibition (RMBI)
• Authors: Gillian L. Gladstone & Gordon B. Parker
• Publication Year: 2005
• Administration Format: Paper-and-pencil self-report or computer-administered digital assessment
• Target Respondent: Adults (aged 18 and older) recalling their behavioral and emotional tendencies during early-to-middle childhood (ages 2 to 12 years)
• Completion Time: Approximately 4 to 6 minutes
• Item Count: 18 items
• Response Scale: 3-point categorical frequency scale plus a memory exclusion option:
  • no/hardly ever (scored 0 for direct items)
  • some of the time (scored 1 for direct items)
  • yes/most of the time (scored 2 for direct items)
  • Do not remember at all (coded as missing data / unremembered, excluded from the direct numeric sum)
• Reverse-Scored Items: Items 4, 5, 7, 11, 13, and 15 tap outgoingness, approach, high physical activity, and risk-taking. These items must be reverse-coded prior to computing the total or subscale score:
  • no/hardly ever = 2 points
  • some of the time = 1 point
  • yes/most of the time = 0 points
• Scoring and Interpretation:
  • Total Score Range: 0 to 36 points (assuming all 18 items are answered with memory recalled).
  • Total Score Calculation: Sum of all standard scored items (1, 2, 3, 6, 8, 9, 10, 12, 14, 16, 17, 18) plus all reversed scored items (4, 5, 7, 11, 13, 15).
  • Missing Data Rule: If more than 3 items are rated as “Do not remember at all” or left blank, the total score should not be computed. For 1 to 2 missing/unremembered items, mean item replacement or pro-rating across valid items is recommended.
  • Cut-off Scores:
    • Scores 0 – 11: Low Behavioural Inhibition (Uninhibited, proactive, extraverted childhood profile).
    • Scores 12 – 18: Average / Moderate Behavioural Inhibition (Normative childhood reticence).
    • Scores 19 – 36: High Behavioural Inhibition (Substantially elevated temperamental inhibition, marked vulnerability to social anxiety and affective disorders).

Permissions & Fee and Test Year

The Retrospective Measure of Behavioural Inhibition was initially published in 2005 in Psychiatry Research by Dr. Gillian L. Gladstone and Scientia Professor Gordon B. Parker. The instrument is copyrighted by the authors and the Black Dog Institute (Prince of Wales Hospital, Sydney, Australia).

Permissions and Research Access: The RMBI is made accessible for academic, clinical, and non-commercial research purposes without charging a license fee. Clinicians and academic investigators may reproduce and administer the scale within academic studies, clinical trials, or individual patient evaluations, provided that appropriate citation and attribution are given to the original 2005 publication. Commercial use, reproduction in for-profit software, or corporate training platforms requires formal written authorization from the authors and the Black Dog Institute. The instrument forms, along with its companion adult tool (the AMBI), can be accessed via the official repository of the Black Dog Institute.

References

• Gladstone, G. L., & Parker, G. B. (2005). Measuring a behaviourally inhibited temperament style: Development and initial validation of new self-report measures. Psychiatry Research, 135(2), 133–143. https://doi.org/10.1016/j.psychres.2005.03.005
• Gladstone, G. L., Parker, G. B., Mitchell, P. B., Wilhelm, K. A., & Malhi, G. S. (2006). Relationship between self-reported childhood behavioural inhibition and lifetime anxiety disorders in a clinical sample. Depression and Anxiety, 23(3), 135–142. https://doi.org/10.1002/da.20148
• Gladstone, G. L., Parker, G. B., & Nguyen, T.-V. (2006). Retrospective and contemporary behavioral inhibition in adult twins: Genetic and environmental influences. Journal of Anxiety Disorders, 20(7), 903–918. https://doi.org/10.1016/j.janxdis.2005.10.002
• Gray, J. A. (1982). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. Oxford University Press.
• Kagan, J., Reznick, J. S., Clarke, C., Snidman, N., & Garcia-Coll, C. (1984). Behavioral inhibition to the unfamiliar. Child Development, 55(6), 2212–2225. https://doi.org/10.2307/1129793
• Kagan, J., Reznick, J. S., & Snidman, N. (1987). The physiology and psychology of behavioral inhibition in children. Child Development, 58(6), 1459–1473. https://doi.org/10.2307/1130685
• Rosenbaum, J. F., Biederman, J., Gersten, M., Hirshfeld, D. R., Meminger, S. R., Herman, J. B., Kagan, J., Reznick, J. S., & Snidman, N. (1988). Behavioral inhibition in children of parents with panic disorder and agoraphobia: A controlled study. Archives of General Psychiatry, 45(5), 463–470. https://doi.org/10.1001/archpsyc.1988.01800290077010

Items of the Scale