What is differential identification theory?
Daniel Glaser's differential identification theory (1956) corrects Sutherland: you do not need to mix with offenders to learn to offend, it is enough to identify with someone, real or imaginary, from whose perspective crime seems acceptable. It may be an idol, a fictional character or a group admired from afar.
A boy watches a gangster film and walks out of the cinema differently. Another memorizes his idol's gestures until they become his own. Most stop there, in harmless fantasy, but some cross a line and act from the skin of the character they admire: they kill, harass or destroy as they believe he would. Daniel Glaser gave that leap a name: we do not offend only because of the bad company we keep, but because of the figures we identify with.
Differential identification at a glance
- Author
- Daniel Glaser
- Year
- 1956
- Family
- Processes · Social learning
- Idea
- We offend by identifying with models (real or imaginary) from whose perspective crime seems acceptable
Admired model (real or imaginary) → selective identification → adopting his perspective ⇒ the crime seems acceptable// OriginThe correction to Sutherland
Sutherland's differential association held that crime is learned in interaction with others: the more contacts with definitions favorable to crime, the more likely offending becomes. It explained the boy surrounded by criminals, but not the one who offends without having been raised among them, nor the fan who imitates an idol he has never touched. Glaser closed that gap in 1956: what matters is not only whom you associate with, but whom you identify with.
His formulation: a person pursues criminal behavior "to the extent that he identifies himself with real or imaginary persons from whose perspective his criminal behavior seems acceptable". The decisive word is "imaginary". The model can be on a screen, in a book or in a legend, as long as it becomes a reference figure. Thus a loner without a single criminal friend can turn criminal by identifying with a dark hero, a movie killer or a fallen idol.
The differential identification thesis
You do not need a flesh-and-blood criminal mentor: wanting to be the gangster you admire is enough.
Identifying is not just imitating
Identification runs deeper than imitation. To imitate is to copy a behavior; to identify is to adopt a perspective, to see the world and oneself as the model would. Glaser also stressed that identification is selective: from all possible models, we choose those that confirm what we already wish to be. The fan does not become violent "because of" his idol; he chooses an idol that legitimizes the violence already latent in him.
From admiration to crime, step by step
- 1ExposureAn attractive model appears, real, media-based or fictional: the idol, the movie gangster, the famous killer
- 2Selective identificationThe subject chooses, among all models, the one that legitimizes what he already wished to be
- 3Adopting the perspectiveHe stops merely imitating: he sees the world, and sees himself, as the model would
- 4Acting outHe acts "from the skin" of the character, for whom that conduct seems acceptable, even admirable
Whoever identifies adopts an identity rather than copying a behavior; that is why he does not feel he is doing something alien.
// In fictionFrom idol to costume: identifications that turned into crime
el rey por una noche is the pure case: a nobody who identifies to the point of delirium with a television idol and kidnaps in order to become him ("better to be king for a night than schmuck for a lifetime"). el Enigma takes identification into modern territory: a loner builds a persona, the Riddler, cheered on by an online community, and from behind that mask commits what Edward Nashton would never have dared to commit. Ghostface is the theory turned meta joke: a killer who has learned to kill from horror movies, quotes their rules and sees himself as the villain of a slasher.
el fan embodies parasocial identification with a sports idol: an imaginary, one-sided and total relationship that, when it breaks, turns into possessive violence. And Red Mist identifies with fictional villains themselves: a rich kid who, unable to be the hero, decides to be the supervillain he has seen in comics, costume included.
Identifications with an imaginary model, from fame to the supervillain. Click to open each case file.
// The mediaDo the media create criminals?
Differential identification lies at the heart of criminology's oldest and most manipulated debate: do violent media manufacture offenders? The media expand the catalogue of models, and for a vulnerable minority a character can become the reference figure that legitimizes the acting out. But watching violence does not "program" anyone: millions watch the same films and do not kill. The medium does not create the desire; it offers a form and a justification to a desire that was already looking for a model. Billy Loomis puts it this way: "movies don't create psychos; movies make psychos more creative". Cinema did not give him the urge to kill; it gave him the script and the alibi of being "inside a movie".
// The crackNeither moral panic, nor "idols are innocent"
The theory has two edges. The first is moral panic: blaming a video game, a song or a film for every crime and censoring culture as if it were the cause. Glaser does not say that, because identification is selective and rare, and the vast majority of fans do not offend. The second edge is the opposite: denying all influence, as if cultural models did not matter. For the vulnerable subject, the reference figure can indeed be the missing push.
Nor does identification excuse anything: that Pupkin believed himself a star or that Red Mist dressed as a supervillain does not lessen their responsibility. It explains from where they acted; it does not absolve them of what they did.
The link between Sutherland and Bandura
Differential identification holds a pivotal place in the history of criminal learning. Glaser bridged Sutherland's differential association, which required direct contact with offenders, and Bandura's social learning theory, which would prove in the laboratory that merely observing a model is enough to imitate it. His idea remains a widely cited framework in the debate on media effects and copycat crimes.
- 1947Sutherland consolidates differential association: crime is learned in direct interaction.
- 1956Glaser publishes "Criminality Theories and Behavioral Images" and adds identification with imaginary models.
- 1961–63Bandura and the Bobo doll experiments: the model is imitated merely by observing it.
- 1977Bandura formalizes social learning theory, a direct heir of this line.
// Why it mattersOffering better mirrors
If crime can be born of an identification, prevention does not lie in banning the bad models, which is impossible and counterproductive, but in competing with them: offering real reference figures, bonds and belonging that make it unnecessary to seek identity in a destructive idol or in an online community that cheers on hatred. In other words, better mirrors: mentors, causes and communities that offer status and identity without demanding blood in return. And in the age of screens and algorithms, it has never been so easy to find the community that celebrates the worst models.
Lights and shadows
- Closes a real gap in Sutherland: it explains the solitary offender, without criminal company, who identifies with a distant or fictional model.
- Anticipates Bandura's social learning and remains the most useful framework for the debate on violent media and copycat crimes.
- The idea of selective identification avoids determinism: the medium offers form and alibi to a desire that already existed.
- It has a concrete preventive reading: offer better mirrors instead of banning models.
- It is hard to falsify: "he identified with a model" can explain almost anything after the fact without predicting anything.
- It does not say why some make the toxic identification and others, exposed to the same thing, do not.
- It lends itself to moral panic and to justifying the censorship of fiction.
- It leaves out the structural causes (poverty, exclusion) that make some young people more likely to seek identity in destructive models.
Three key takeaways
- Glaser corrects Sutherland: we offend not only because of the company we keep, but because of the figures, real or imaginary, from whose perspective crime seems acceptable.
- To identify is to adopt a perspective, and it is selective: we choose the idol that legitimizes what we already wished for. The media do not "create" criminals, but they offer models and alibis.
- The two edges are moral panic and denying all influence. Prevention lies in offering better mirrors, not in banning models.
Sources · further reading
- Glaser, D. (1956). “Criminality Theories and Behavioral Images”. American Journal of Sociology, 61(5).
- Sutherland, E. H. & Cressey, D. R. (1978). Criminology. Lippincott.
- Bandura, A. (1977). Social Learning Theory. Prentice-Hall.
- Surette, R. (2015). Media, Crime, and Criminal Justice. Cengage.
// Test the theoryWhich of these villains does KRIMINIS analyse with Differential identification theory?
How it differs from its neighbours
Theory questionsFrequently asked questions
What is differential identification theory?
Daniel Glaser's differential identification theory (1956) corrects Sutherland: you do not need to mix with offenders to learn to offend, it is enough to identify with someone, real or imaginary, from whose perspective crime seems acceptable. It may be an idol, a fictional character or a group admired from afar.
Who developed differential identification theory, and when?
The sociologist Daniel Glaser, in 1956, in the article “Criminality Theories and Behavioral Images” (American Journal of Sociology, 61(5)). He proposed it as a correction to Edwin Sutherland's differential association, which in 1947 had established that crime is learned through direct interaction with other people.
Is differential identification still valid? What is it criticised for?
Yes, it is current and very influential. Critics say it is hard to falsify, since “they identified with a model” can explain anything after the fact, and that it does not explain why some people form a toxic identification while others exposed to the same thing do not. It also lends itself to moral panics about games or films and leaves out structural causes such as poverty.
Do violent media create criminals, according to Glaser?
Not in any simple sense. Media widen the range of models to identify with, and for a vulnerable minority a character can legitimise the step into action. But identification is selective: we choose models that confirm what we already want to be, and millions watch the same films without offending. Media give shape to a desire; they do not create it.