Codependency test: what is psychometrically valid, what is pop-psychology

Codependency: concept vs clinical diagnosis

The query "codependency test" has held firm in top searches for the last decade — about 864 monthly impressions on the head-term, ~16,000 in the broader cluster including "codependent relationships". Yet the term "codependency" is a descriptive concept, not a clinical diagnosis. Understanding this distinction matters most to the person taking such a test: what you are searching for does not exist as a formal nosological entity.

Cermak (1986, Journal of Psychoactive Drugs) proposed including "Co-Dependent Personality Disorder" in DSM-III-R as an Axis II disorder. APA rejected the proposal. Since then, no DSM edition (III-R, IV, IV-TR, 5, 5-TR) and no ICD version (10, 11) has included codependency as a standalone diagnosis. Stafford (2001, Issues in Mental Health Nursing) in a conceptual and measurement review asked directly: "is it ethical to encourage an individual to accept that he or she is codependent before the construct has achieved a universal operational definition?" Twenty-five years later, that definition has still not formed.

Spann-Fischer Codependency Scale (SFCS): what the science shows

SFCS (Fischer, Spann & Crawford, 1991, Alcoholism Treatment Quarterly, 8(1):87–100) is the most-cited scale to date. 16 items, 6-point Likert, original Cronbach α ≈ 0.86, test-retest r ≈ 0.87. At first glance — acceptable psychometric reliability.

The problem lies in construct validity. Lindley, Giordano & Hammer (1999, J Clin Psychol, 55(1):59–64) found that "low self-confidence was the strongest predictor of codependency" on SFCS. That is, a large share of the variance SFCS attributes to "codependency" is in fact low-self-esteem variance. It overlaps with an independent construct that has its own validated instrument (RSES, see below). From measurement-theory standpoint, an instrument that duplicates another construct without adding information is a discriminant-validity problem.

Fuller & Warner (2000, Genetic, Social, and General Psychology Monographs, N=257 students) used SFCS alongside the Potter-Efron CA and found significantly elevated scores in respondents with a history of an "alcoholic, mentally ill, or physically ill parent". That is, SFCS is sensitive to a context of family dysfunction — but that is a family-context construct, not a unique "codependency" construct.

Regarding Russian adaptation: one academic adaptation exists — Berdichevskiy, Padun & Gagarina (2019, Clinical Psychology and Special Education, 8(1):215–234, available on PsyJournals.ru) confirmed the factor structure of a modified SFCS on N=227 students + N=38 cadets. This is peer-reviewed work, but the samples are narrow (young people without clinical complaints), and there are no clinical norms. The instrument is not yet ready for clinical use in Russian-speaking populations.

Four other instruments and their limitations

Beyond SFCS, several other widely cited scales exist. The psychometric status of each is summarised below.

• Holyoake Codependency Index (HCI) — Dear & Roberts 2000, Psychological Reports, 87(3):991–1002. 13 items, 3 subscales (External Focus, Self-Sacrifice, Reactivity), subscale α 0.73–0.84 (sample N=307), test-retest 0.88 at 3 weeks (Dear 2004). The strongest "classic" instrument psychometrically, but α sits at the lower bound of clinical acceptability (usually ≥ 0.80 is required) and there is no RU adaptation.
• Roehling-Gaumond Codependent Questionnaire (CdQ) — Roehling & Gaumond 1996, Alcoholism Treatment Quarterly, 14(1):85–95. 16 items, total α ≈ 0.86. Hungarian replication (Nagy et al. 2014) showed that the "enmeshment" subscale does not hold reliability (α < 0.70) and the proposed factor structure does not replicate in independent samples.
• Composite Codependency Scale (CCS) — Marks, Blore, Hine & Dear 2012, Australian Journal of Psychology, 64:119–127. 19 items, 3 factors (self-sacrifice, interpersonal control, emotional suppression). The strongest modern instrument psychometrically — it discriminates CoDA members from the general population. Internal consistency is good, but no RU adaptation.
• Friel Codependency Assessment Inventory (CAI) — Friel 1985, Focus on Family, 8(3):20–21 (a popular-press magazine). 60 items, widely cited in clinical practice. Critical fact: no peer-reviewed psychometric validation exists. The α=0.86 often attributed to CAI is not confirmed by any published peer-reviewed study. By strict psychometric criteria, CAI is not a validated instrument despite its widespread use.

The overall picture: none of the six most-cited instruments (SFCS, CdQ, HCI, CCS, Friel CAI, Hughes-Hammer CODAT 1998) is validated on a Russian clinical sample with established norms. Berdichevskiy 2019 SFCS is the only peer-reviewed RU adaptation, and that one is on non-clinical samples.

What to use instead: three valid pathways

If "codependency" patterns resonate, it is clinically and psychometrically more defensible to measure their components with validated instruments. Three pathways below, in the order they are typically needed.

• 1. Anxious attachment — ECR-R (Fraley, Waller & Brennan 2000, JPSP, 78(2):350–365). 36 items, two subscales — attachment anxiety and attachment avoidance, Cronbach α 0.93 and 0.94 respectively (Sibley, Fischer & Liu 2005). The "high anxiety + low avoidance" profile is phenomenologically close to what is popularly described as "codependency". Marks et al. 2012 and Ançel & Kabakçı 2009 (Turkish CODAT validation, N=400) both confirmed: higher codependency-scale scores associate with higher attachment-related anxiety. RU adaptation of ECR-R is by Chursina (2022/2023), Psychology in Russia: State of the Art, 16(3):222–232.
• 2. Self-esteem + interpersonal patterns. RSES (Rosenberg 1965) — 10 items, mean α across 53 nations ≈ 0.81 (Schmitt & Allik 2005, JPSP, 89(4):623–642). IIP-32 (Barkham, Hardy & Startup 1996, Br J Clin Psychol, 35(1):21–35) — 32 items (short form of Horowitz IIP), total α ≈ 0.87, eight octants of interpersonal problems. Lindley et al. 1999 showed that low self-esteem is the primary predictor of SFCS scores; measuring it directly is methodologically cleaner.
• 3. Co-occurring substance use in the family system. AUDIT (Saunders et al. 1993, Addiction, 88(6):791–804) — 10 items, cutoff ≥ 8 yields sensitivity 92% / specificity 94% for hazardous alcohol use. DAST-10 (Skinner 1982; Yudko et al. 2007 review, J Subst Abuse Treat, 32(2):189–198) — 10 items, mean α across the review ≈ 0.90. A large part of the "codependency" phenomenon is a reaction pattern to substance use in a close relationship. Knowing the formal severity of that substance use is methodologically critical.

Strategically: three validated instruments measuring known constructs with peer-reviewed norms yield more information about the clinical situation than a single scale with construct-validity scatter. Marks et al. 2012 and Ançel & Kabakçı 2009 empirically showed that codependency-scale scores are largely accounted for by exactly these three components.

Patterns worth paying attention to

The concept of "codependency" is useful as a description of a pattern, even if it does not exist as a diagnosis. Below are five phenomenological patterns that have historically been grouped under that term. If several resonate, that is not "I have codependency" — it is a signal to investigate clinically and with valid instruments.

• Self-worth depends on partner / significant-other approval; in their absence — a sharp drop in subjective worth
• Chronic difficulties with boundaries: hard to say "no", fear of abandonment overrides own interests
• Taking responsibility for others' emotions and actions — especially with a partner who has substance-use or mental-health problems
• A long-running "rescuer" role in relationships: trying to control the situation through caring for the other while suppressing own needs
• Emotional suppression: difficulty identifying and expressing own emotions, particularly negative ones (anger, fear, sadness)

Six markers of an invalid "codependency test"

Most online "codependency tests" in b17.ru, messengers, on Instagram pages of psychologists — are not psychometric instruments but entertainment quizzes with emotionally loaded interpretations. Six markers that distinguish them from a valid instrument.

• No source citation. A valid instrument always cites authors and year of publication (e.g., "Fischer, Spann & Crawford 1991"). A pop-quiz is usually anonymous or refers to a website rather than a peer-reviewed journal.
• No validation-sample data. "On what sample was the scale validated?" is a fundamental question. A valid instrument reports: N, demographics, country, clinical vs non-clinical context. A pop-quiz is silent on this.
• No psychometrically grounded cut-off scores. A valid instrument: "cut-off ≥ X yields sensitivity Y% / specificity Z%". A pop-quiz: "more than 70% 'yes' answers = severe codependency" — with no source for the 70%.
• 5–10 items with discrete "yes / no". SFCS has 16 items on a 6-point Likert. HCI has 13 items on a 6-point Likert. CCS has 19 items on a 5-point. The cruder the scale and fewer the items, the less psychometric capacity.
• Promise of a "fast result in 2 minutes". Psychometric completion requires conscious self-reflection. 30 seconds for 10 questions is not measurement — it is in-the-moment self-feeling.
• Emotionally loaded interpretation without caveats. "You are a typical codependent person, your life does not belong to you." A valid report: "score X falls in range Y, which corresponds to Z. This is not a diagnosis; clinical evaluation requires a specialist consultation".

When you need a clinician

Independent work with "codependency" patterns — through structured programs (CoDA, ACoA), books (Beattie, Mellody, Norwood) — can be a useful first step. But these programs do not substitute for clinical assessment when one is indicated. The 5-criterion decision tree covered in the companion article applies directly here.

Concrete grounds for clinician contact: patterns affect functioning (work, relationships, physical health, basic self-care); there is a substance-use disorder in the family system — own or close person; self-driven change attempts produce no shift over 3+ months of regular work; there are comorbid symptoms of depression or anxiety requiring clinical evaluation in their own right (PHQ-9 ≥ 10, GAD-7 ≥ 10). Any suicidal ideation calls for immediate clinician contact rather than self-help (see the previous article on red flags).

What a clinician does differently from a pop-quiz: operationalises the complaint through several validated instruments, explores family and relationship history, identifies specific clinical entities (e.g., anxious attachment, dependent personality features, childhood PTSD, current substance use in the system), and proposes an intervention — typically evidence-based, most often CBT, ACT, schema-focused therapy, or systemic family therapy depending on context. This is not "a diagnosis of codependency" — it is clinical analysis of components.

Sources / Источники
Fischer J.L., Spann L., Crawford D. (1991). Measuring codependency. Alcoholism Treatment Quarterly, 8(1), 87–100. · Cermak T.L. (1986). Diagnostic criteria for codependency. J Psychoactive Drugs, 18(1), 15–20. · Stafford L.L. (2001). Is codependency a meaningful concept? Issues Ment Health Nurs, 22(3), 273–286. · Lindley N.R., Giordano P.J., Hammer E.D. (1999). Codependency: predictors and psychometric issues. J Clin Psychol, 55(1), 59–64. · Dear G.E., Roberts C.M. (2000). The Holyoake Codependency Index. Psychol Rep, 87(3 Pt 1), 991–1002. · Dear G.E., Roberts C.M. (2005). Validation of the Holyoake Codependency Index. J Psychol, 139(4), 293–313. · Roehling P.V., Gaumond E. (1996). Reliability and validity of the Codependent Questionnaire. Alcoholism Treatment Quarterly, 14(1), 85–95. · Marks A.D.G., Blore R.L., Hine D.W., Dear G.E. (2012). Development and validation of a revised measure of codependency. Aust J Psychol, 64, 119–127. · Ançel G., Kabakçı E. (2009). Psychometric properties of the Turkish form of the Codependency Assessment Tool. Arch Psychiatr Nurs, 23(6), 441–453. · Fuller J.A., Warner R.M. (2000). Family stressors as predictors of codependency. Genet Soc Gen Psychol Monogr, 126(1), 5–22. · Fraley R.C., Waller N.G., Brennan K.A. (2000). An item response theory analysis of self-report measures of adult attachment. J Pers Soc Psychol, 78(2), 350–365. · Sibley C.G., Fischer R., Liu J.H. (2005). Reliability and validity of the Revised Experiences in Close Relationships scale. Pers Soc Psychol Bull, 31(11), 1524–1536. · Schmitt D.P., Allik J. (2005). Simultaneous administration of the Rosenberg Self-Esteem Scale in 53 nations. J Pers Soc Psychol, 89(4), 623–642. · Barkham M., Hardy G.E., Startup M. (1996). The IIP-32: a short version of the Inventory of Interpersonal Problems. Br J Clin Psychol, 35(1), 21–35. · Saunders J.B., Aasland O.G., Babor T.F., de la Fuente J.R., Grant M. (1993). Development of the Alcohol Use Disorders Identification Test (AUDIT). Addiction, 88(6), 791–804. · Yudko E., Lozhkina O., Fouts A. (2007). A comprehensive review of the psychometric properties of the Drug Abuse Screening Test. J Subst Abuse Treat, 32(2), 189–198. · Berdichevskiy A.A., Padun M.A., Gagarina M.A. (2019). Validation of a modified Spann-Fischer Codependency Scale. Clinical Psychology and Special Education, 8(1), 215–234. · American Psychiatric Association (2022). DSM-5-TR. · World Health Organization (2022). ICD-11.