Personality Implicitify Research Team

From Anna Freud's Ego to the DSQ-88: How Defense Mechanisms Became Measurable

Defense mechanisms have an unusual standing in psychology. They are one of the few psychoanalytic constructs that the empirical wing of the field has, after a century of skepticism, mostly decided to keep. The reasons for that are worth understanding, because they explain why a self-report instrument like the Defense Style Questionnaire (DSQ-88) earns its place in a multimethod personality battery rather than being dismissed as a translation exercise from the consulting room.

Anna Freud and the inventory of ego defenses

Sigmund Freud's writings on defense are scattered across his career and never organized into a clean list. The taxonomic work was done by his daughter. Anna Freud's Das Ich und die Abwehrmechanismen (1936; English translation as The Ego and the Mechanisms of Defense, 1937) collected ten core defenses — repression, regression, reaction formation, isolation, undoing, projection, introjection, turning against the self, reversal, and sublimation — and added the framing claim that has organized everything since: defenses are normal operations of the ego, not pathological symptoms in their own right. They become problems when they are rigid, when they are deployed in situations that do not call for them, or when they are organized around immature solutions that the person never updated.

That last point is the durable one. Anna Freud opened the door to a developmental view of defense, in which a person's characteristic defensive repertoire reflects how far the personality has progressed in its capacity to handle internal conflict and external stress. Subsequent theorists ran with the idea, but for several decades the literature remained almost entirely qualitative. Defenses were inferred from clinical material, written about in case reports, and taught in supervision. Nobody had a defensible way to measure them.

Vaillant: the longitudinal turn

The first major empirical traction came from George Vaillant's analysis of the Grant Study of Harvard sophomores from the late 1930s onward. Vaillant's Adaptation to Life (1977) and the subsequent Wisdom of the Ego (1993) used decades of longitudinal data — interviews, biographical material, observed outcomes — to argue for a hierarchy of defense maturity. Defenses sorted, in his analysis, into roughly four levels: psychotic (denial of external reality, distortion), immature (projection, acting out, passive aggression, hypochondriasis), neurotic (repression, displacement, reaction formation, intellectualization), and mature (sublimation, humor, anticipation, suppression, altruism).

Two findings from the Grant Study became canonical. First, the maturity of a person's habitual defense style at age 30 predicted physical health, occupational success, and relationship satisfaction at age 65 — the prospective relationship was on the order of correlations seen for any major personality variable. Second, defense maturity changed over the lifespan, and the change was directional: most people moved from neurotic-and-immature toward more mature defenses, with the steepest movement in midlife. The construct was real, it was prospectively predictive, and it was developmentally meaningful.

What the Grant Study could not provide was a measurement procedure usable outside a forty-year longitudinal investment. Vaillant's defense ratings required clinically trained coders working from rich biographical material. For the construct to enter the broader personality literature, somebody had to put it into a questionnaire.

Bond and the DSQ

That somebody was Michael Bond. Bond, working in Montreal in the early 1980s, designed the Defense Style Questionnaire as a deliberate translation of the Vaillant-style hierarchy into self-report. The original Defense Style Questionnaire (Bond, Gardner, Christian, & Sigal, 1983, Archives of General Psychiatry, 40, 333–338) used 88 items, each a first-person statement reflecting the conscious derivative of an unconscious defensive operation. The respondent did not have to identify the defense; they had to endorse the surface attitude or behavior that the defense produced.

The crucial methodological move is that the items were not derived rationally from the defense list. They were generated, given to clinical samples, and then refined empirically: items that did not load coherently with their intended defense or did not discriminate clinical from non-clinical samples were dropped. The result was a questionnaire whose items could be aggregated into scores for individual defenses and, more importantly, into higher-order defense styles derived from factor analysis of the data rather than from the theoretical hierarchy.

Bond's original three-factor solution, replicated repeatedly across the next two decades, sorted defenses into three styles whose names map loosely onto Vaillant's:

Mature. Humor, sublimation, suppression, anticipation. Defenses that reduce conflict without distorting reality and without externalizing the cost.
Neurotic. Reaction formation, idealization, undoing, pseudo-altruism. Defenses that dampen anxiety by shifting how the conflict is represented internally.
Immature. Projection, denial, splitting, passive aggression, acting out, somatization, dissociation, and others. Defenses that protect the self by externalizing the conflict — onto another person, onto the body, onto the situation.

The DSQ-88 is the long form of this instrument. Shorter versions (DSQ-72, DSQ-60, DSQ-40, DSQ-36) drop items to reduce administration time but retain the same factor structure. The 1993 update by Andrews, Singh, and Bond (Journal of Nervous and Mental Disease, 181, 246–256) is the most-cited validation reference and the source of the modern scoring conventions used in the DSQ-88 measure described on this site.

What the DSQ does and does not measure

Two things are worth being clear about.

First, the DSQ measures conscious derivatives of unconscious defensive operations. It does not, and cannot, measure the operations themselves directly. A person who relies heavily on projection will, on average, endorse items that reflect the attitudes projection produces ("when I'm having problems, it's usually because of someone else"); they will not, except rarely, endorse the operation in self-aware form. The convergent validity of the DSQ with clinician-rated defense ratings (Perry's Defensive Functioning Scale, Vaillant-style longitudinal coding) is moderate — typically in the .30 to .50 range across studies — and that ceiling reflects this fundamental limitation rather than a flaw in any single instrument. It is the same ceiling implicit-explicit motive measures hit, for the same reason.

Second, the DSQ has held up unusually well to the standard cross-cultural and cross-instrument tests. The three-factor structure replicates across English, French, German, Spanish, Chinese, and Japanese samples. The Mature factor reliably correlates positively with measures of well-being, life satisfaction, and adaptive functioning. The Immature factor reliably correlates with personality disorder diagnoses, with shorter time in employment, and with treatment dropout. The Neurotic factor sits in the middle, behaving like a marker of inhibited but functional anxiety regulation. None of these correlations are huge in absolute terms, but they are robust and they are theoretically interpretable.

DSQ in a multimethod profile

The natural place for the DSQ in a contemporary personality assessment is alongside a measure of relational organization (CCRT, attachment), a measure of interpersonal style (the interpersonal circumplex), and an implicit-motive measure (PSE or another content-coded narrative instrument). Each of these tells you something the DSQ does not. The DSQ tells you something they do not: it tells you what the person's first-line solution is when the conflict gets activated.

Two combinations are particularly clinically informative. A high Immature score in someone whose CCRT shows a strong wish-frustration pattern points toward externalization of an interpersonal disappointment that has not been worked through. A high Mature score in someone whose attachment scores are insecure-anxious is a relatively favorable prognostic combination: the person is anxious, but they are using anticipation, humor, and suppression rather than projection or splitting to regulate the anxiety. These cross-instrument readings are not derivable from any single questionnaire, including the DSQ; they are the kind of inference a multimethod battery is designed to support.

The honest caveats

The DSQ inherits the limits of any first-person questionnaire. It is vulnerable to defensive responding (the meta-irony of defense measurement is well-recognized). It is vulnerable to acquiescence and to social desirability, although most published versions include controls. The Mature factor is sometimes contaminated by simple positive-affect endorsement, and clinicians should be alert to high Mature scores that look more like mood-congruent self-presentation than actual defense maturity. The Immature factor is the most diagnostically useful in clinical samples; the differentiation between Mature and Neurotic is more stable in non-clinical research.

None of this means the construct is unsound. It means the construct is hard, and the instrument that measures it is doing so under known constraints. The DSQ has been around for forty years, has been translated into more than a dozen languages, has accumulated a respectable validation literature, and continues to be used in psychotherapy outcome research. That is a reasonable record for a self-report measure of an originally clinical construct.