IX Coach

The Critiques of Adult Development Stage Theories, Fairly Stated

The main objections to stage theories of adult development, stated in the critics’ terms, with what the evidence still supports

Are stage theories of adult development scientific?

It depends on the theory and on the claim. Some stage measures have a long validation record with known limits, and critics have pressed on scoring, on the samples behind stage descriptions, on sex and culture bias, and on whether one newer model was publicized as validated before its research report was available. In the 2017 exchange over STAGES, the critique raised ten points by the 2020 count of Tom Murray, who co-authored the 2020 research summary with O’Fallon (the response itself refers to eleven issues), one of them that the model was prematurely promoted as validated. The developers accepted that timing point, and by Murray’s 2020 account said nine of the ten rested on misconceptions. On this page’s reading, the reviewers’ own conclusions are more qualified than either “proven” or “nonsense,” and the page states each objection in its author’s terms.

Are stage theories of adult development scientific? The answer depends on which theory and which claim. Some measures have a long record of testing; others, on this page’s reading of the sources, rest mainly on studies by their developers. The critics do not object to the same thing. Some question how stages are scored. Some question whether the people in a stage study can stand in for everyone else. Some question whether a model was described as validated before its research was public. This page states those objections in the terms their authors used, reports what the reviewers themselves concluded the evidence still supports, and turns the objections into guardrails for reading a stage claim.

Three bodies of work carry most of the debate. Lawrence Kohlberg’s stages describe moral reasoning, from reasons based on punishment and self-interest up to reasons based on universal principles. Jane Loevinger’s stages describe ego development, the frame of reference through which a person organizes their sense of themselves and the world; she measured it with the Washington University Sentence Completion Test (WUSCT), in which a person finishes 36 open sentence beginnings and raters score the endings against a scoring manual. The STAGES model, developed by Terri O’Fallon, is a newer framework that builds on the work of Loevinger and of Susanne Cook-Greuter, a researcher in the same tradition.

A stage label is easy to hear as a fact. On this page’s reading, a stage description is a hypothesis about how people reason, and how far that hypothesis reaches depends on who was studied, how their answers were scored, and who checked the scoring. The practices below are checks drawn from those objections. None of them can place you or anyone else at a stage, and none has been tested as an intervention.

実践法

これらの実践のつながり

Every practice on this page works from one question that the critiques of stage theories share: what evidence stands behind a claim about where a person, or a group, sits on a developmental sequence, and how far does that evidence reach? The practices check who did the validating, whether the people studied resemble the people being described, whether a scoring scheme counted the reasons a person actually gives, and whether a critic’s own conclusion keeps the model standing. None of them decides whether the stages are real. They keep a stage claim no wider than the evidence beneath it.

How well supported is Loevinger’s sentence completion test of ego development?

Loevinger (1979) reported internal consistency of about .85 and called the construct-validity evidence substantial, but said it falls short of clear proof that the stages form a fixed sequence. She judged the test adequate for research when it is administered and scored with sufficient care, and not for diagnosing an individual without other data. Manners and Durkin (2001) reviewed the construct, predictive and discriminant validity of the theory and the test, and their abstract concludes there is substantial empirical support for its conceptual soundness.

Do Kohlberg’s moral stages hold across cultures and between men and women?

The answer depends on which part of the claim. Snarey (1985) reviewed 45 cross-cultural studies and reports striking support for the underlying assumptions behind Kohlberg’s claim of universal moral development, and also major caveats about the range and general applicability of the stages and biases in favor of complex urban and middle-class populations. On sex differences, Walker (1984) found only a few inconsistent differences in childhood and adolescence, and reported that adult studies confounded sex with education and occupation. Jaffee and Hyde (2000) found small average differences in moral orientation and said their findings do not offer strong support for the claim that care is used mainly by women and justice mainly by men.

Has the STAGES model been validated, and who checked the validation?

Its validation claims have been disputed in print, and the developers have partly conceded the point. In a 2017 exchange, a critique by Cook-Greuter, Wilber and Sharma raised ten points, by Murray’s 2020 account, one of which was that the model was prematurely promoted as validated. The developers’ response said it did not agree that validity was misrepresented, but accepted that it was promoted prematurely, because the statistical report was not yet available for colleagues to scrutinize. By Murray’s account the response said nine of the ten points rested on misconceptions (the response itself counts eleven), and it also names a business rivalry running in both directions. The same response says STAGES research “is in its infancy, with just a tiny amount of empirical validation compared with Loevinger’s WUSCT.” The 2020 validation study’s first author is the model’s developer, and in 2023 Merckoll argued that the two studies cited as validating STAGES scoring do not hold up under scrutiny.

Can a stage score tell me who I am?

Not on its own, by the measures’ own cautions. Loevinger (1979) judged the sentence completion test not suitable for individual diagnosis without other data. In the STAGES research, Murray and O’Fallon (2020) report retest data from 115 clients: of 143 retests, 38% stayed the same, 50% rose and 11% fell. The authors note that many of the test-takers had been in a program that taught developmental models, and that they may have learned vocabulary that led to higher verbal-textual scores (scores on the written wording of the answers) without advancing deeper development. The authors read the 89% who stayed the same or rose as strong evidence for a developmental construct, while noting that the 11% who fell could reflect rater error, test-retest variability or real regression under stress. The practices here treat a score as one reading to be discussed, not as a label.

Do the studies of moral stages in children prove the stages apply to adults?

Those studies were of children, and they describe children more directly than adults. Walker (1980) assessed fourth- to seventh-grade children and found transitions to stage 3 reasoning only among children who had the hypothesized prerequisites. Blatt and Kohlberg (1975) gave children in schools twelve hours of moral discussion over twelve weeks, and the experimental group tended toward higher moral thinking than controls. A reader who wants to apply a stage claim to adult life can ask which ages and settings the evidence came from.

どれを選べばよいか

If a stage model has been described to you as proven, start with who validated it and whether the report is public. If you are trying to place yourself or someone else, prefer a scored instrument to a self-chosen description, and collect more than one observation before you place anyone. If you are applying a description across culture, class or sex, ask who the original sample was and what else differed between the groups. If you are reading a critique, read the support clause next to the objection. If the stage claim comes from research on children, treat it as a claim about children until you have evidence from adults.

これらのアプローチでは解決できないこと

This page is a reading of the critiques, and it does not rule on whether any stage model is scientific. Loevinger’s measure has a long record, but she herself said the evidence falls short of clear proof of sequence. On this page’s reading, much of the support cited for these models is correlational and drawn from particular samples. The newer STAGES validation was publicized before its research report was available, by the developers’ own account, the developers say the research is in its infancy, and the first author of the 2020 validation study is the model’s developer. The 2023 critique concerns scoring rather than the theory. Snarey and Walker reviewed older literature, and the Gilligan argument is reported through secondary sources because her book was not read for this page. None of the eight practices has been tested as a way to change anyone’s stage.

IX Coach でこれを実践する

IX Coach で実践する

よくある質問

Are stage theories of adult development scientific?

It depends on the theory and on the claim. Some stage measures have a long validation record with known limits, and critics have pressed on scoring, on the samples behind stage descriptions, on sex and culture bias, and on whether one newer model was publicized as validated before its research report was available. In the 2017 exchange over STAGES, the critique raised ten points by the 2020 count of Tom Murray, who co-authored the 2020 research summary with O’Fallon (the response itself refers to eleven issues), one of them that the model was prematurely promoted as validated. The developers accepted that timing point, and by Murray’s 2020 account said nine of the ten rested on misconceptions. On this page’s reading, the reviewers’ own conclusions are more qualified than either “proven” or “nonsense,” and the page states each objection in its author’s terms.

「Close the reading and come back to the room」はどのように機能するのか

Shame tends to pull attention into self-judgement and makes the body feel small and exposed, which is the opposite of the state needed to read carefully. Turning to the room and the body interrupts that ranking and brings you back to the ground you are actually standing on. Only then does reading about stages become something you can weigh, rather than something that places you.

「Reach a person if the shame is more than a reading can carry」はどのように機能するのか

Shame about being behind can grow into a sense that there is no point in continuing, and that feeling is a reason to bring in another person rather than rank yourself alone. A clinician can look at the shame, the reading and your life together, which a page of theory cannot. Saying the feeling out loud to someone often helps it loosen its grip.

「Check who validated a stage model before you repeat its claims」はどのように機能するのか

The 2017 exchange over the STAGES model turned on how validity was described and when the supporting report was public. The critique by Susanne Cook-Greuter, Ken Wilber and Beena Sharma, as quoted in the developers’ response, said the model was “prematurely promoted as validated.” By Murray’s 2020 account the critique raised ten points in all. The response does not agree that validity was misrepresented, but accepts that it was promoted prematurely, because the statistical report was not yet available for colleagues to scrutinize; Murray (2020) reports that the response judged nine of those ten points to rest on misconceptions; the response itself counts eleven issues. A validation claim is only as checkable as the report behind it, so asking who ran the study, whether it is in print, and who else has examined it tells a reader how much weight the claim can carry.

「Prefer a scored assessment to reading yourself into a stage」はどのように機能するのか

Loevinger built her test on free-written sentence endings scored by raters working from a scoring manual, because a person’s frame of reference shows most clearly when they project it rather than choose from options. Her 1973 faking studies found that people could deliberately lower their scores, mostly down to the Self-protective stage (an early stage in her scheme), and that almost half of the men told to fake higher could raise theirs, usually by no more than half a step. Her 1979 review judged the test adequate for research but not for individual diagnosis unsupported by other data. Reading stage descriptions and choosing the one that fits skips the scoring step the test’s author treated as central. On this page’s reading, a self-chosen label can reflect the stage a person wants to be, the vocabulary of the descriptions, or the most flattering option, and nothing in the choice separates these.

「Do not place someone at a stage from one conversation or one moment」はどのように機能するのか

In the Moral Judgment Interview (MJI), a conversation about three moral dilemmas, Bataglia, Morais and Lepre (2010) describe the predominant stage as proposed from the answers to the follow-up questions after the dilemmas. Loevinger (1979) judged the test unsuitable for individual diagnosis without other data. The distance between one observation and a stage label is where a model that is reasonable for research can turn into a way of sorting people. Several observations, kept apart from the label, keep the claim no wider than the evidence.

「Check whether the people in a stage study resemble the people you are describing」はどのように機能するのか

Snarey (1985) reviewed 45 cross-cultural studies that tested Kohlberg’s claim of universal moral development. He reports striking support for the underlying assumptions, and also identifies major caveats about the range and general applicability of the stages, with biases in favor of complex urban societies and middle-class populations. A description built from one kind of sample can carry that sample’s sense of what counts as a good reason. Asking who was studied is how a reader tests whether a description travels to the people in front of them.

「Read a gender gap against the schooling and jobs of the people measured」はどのように機能するのか

Walker (1984), in a critical review of sex differences in moral reasoning, found only a few inconsistent sex differences in childhood and adolescence. He reports that studies suggesting men show higher moral development in adulthood confound sex with differences in education and occupation (mix the two together, so their separate effects cannot be read off), and that a meta-analysis (a statistical pooling of many studies) supported the conclusion that the overall pattern is one of nonsignificant sex differences. Where schooling and work are confounded with sex, a gap between groups cannot be read as a sex difference until those factors are held fixed.

「Ask whether a moral scoring scheme counts care and emotion as reasons」はどのように機能するのか

Bataglia, Morais and Lepre (2010) report the critique, citing Carol Gilligan (1993), that Kohlberg’s theories concern only a male morality based on justice and devalue the affective dimension and emotions. Gilligan’s own book was not read for this page, so the claim rests on that report. Jaffee and Hyde (2000), in a meta-analysis (a statistical pooling of many studies), note that Gilligan (1982) argued for two modes of moral reasoning, justice and care, and tested whether either is used predominantly by one sex. They found small differences, favoring women on care and men on justice, and said these findings do not offer strong support for that sex-predominance claim. On this page’s reading, a scheme that counts only justice-based reasons would score a care-based reason as absent rather than as less developed, so the practice asks what a framework counts.

「Hold a critique next to the support the reviewers still found」はどのように機能するのか

Loevinger’s own 1979 review ends with support and a stated limit: its construct validity (whether the test measures the concept it is built to measure) is substantial, but falls short of clear proof of sequentiality (that the stages come in a fixed order). The 2001 review by Manners and Durkin examined the theory and the sentence completion test for construct, predictive (whether scores forecast the outcomes they should) and discriminant (whether the test separates the concept from related measures) validity, and its abstract reports substantial empirical support for the conceptual soundness of both. Keeping both halves of each verdict in view keeps a reader from treating a critique as a dismissal, or a supportive review as a proof.

「Check whether a stage claim’s evidence came from children or from adults」はどのように機能するのか

Several of the studies cited on this page were run on children. Walker (1980) assessed fourth- to seventh-grade children, exposed them to stage 3 reasoning in a brief role-play, and found transitions only among children who had reached the hypothesized prerequisites. Blatt and Kohlberg (1975) ran a classroom experiment in which children received twelve hours of moral discussion over twelve weeks, and the experimental group tended toward higher moral thinking than controls. Evidence about how children move through reasoning stages says less directly about adult reasoning, and the ages of the sample decide how far a claim reaches.

How well supported is Loevinger’s sentence completion test of ego development?

Loevinger (1979) reported internal consistency of about .85 and called the construct-validity evidence substantial, but said it falls short of clear proof that the stages form a fixed sequence. She judged the test adequate for research when it is administered and scored with sufficient care, and not for diagnosing an individual without other data. Manners and Durkin (2001) reviewed the construct, predictive and discriminant validity of the theory and the test, and their abstract concludes there is substantial empirical support for its conceptual soundness.

Do Kohlberg’s moral stages hold across cultures and between men and women?

The answer depends on which part of the claim. Snarey (1985) reviewed 45 cross-cultural studies and reports striking support for the underlying assumptions behind Kohlberg’s claim of universal moral development, and also major caveats about the range and general applicability of the stages and biases in favor of complex urban and middle-class populations. On sex differences, Walker (1984) found only a few inconsistent differences in childhood and adolescence, and reported that adult studies confounded sex with education and occupation. Jaffee and Hyde (2000) found small average differences in moral orientation and said their findings do not offer strong support for the claim that care is used mainly by women and justice mainly by men.

Has the STAGES model been validated, and who checked the validation?

Its validation claims have been disputed in print, and the developers have partly conceded the point. In a 2017 exchange, a critique by Cook-Greuter, Wilber and Sharma raised ten points, by Murray’s 2020 account, one of which was that the model was prematurely promoted as validated. The developers’ response said it did not agree that validity was misrepresented, but accepted that it was promoted prematurely, because the statistical report was not yet available for colleagues to scrutinize. By Murray’s account the response said nine of the ten points rested on misconceptions (the response itself counts eleven), and it also names a business rivalry running in both directions. The same response says STAGES research “is in its infancy, with just a tiny amount of empirical validation compared with Loevinger’s WUSCT.” The 2020 validation study’s first author is the model’s developer, and in 2023 Merckoll argued that the two studies cited as validating STAGES scoring do not hold up under scrutiny.

Can a stage score tell me who I am?

Not on its own, by the measures’ own cautions. Loevinger (1979) judged the sentence completion test not suitable for individual diagnosis without other data. In the STAGES research, Murray and O’Fallon (2020) report retest data from 115 clients: of 143 retests, 38% stayed the same, 50% rose and 11% fell. The authors note that many of the test-takers had been in a program that taught developmental models, and that they may have learned vocabulary that led to higher verbal-textual scores (scores on the written wording of the answers) without advancing deeper development. The authors read the 89% who stayed the same or rose as strong evidence for a developmental construct, while noting that the 11% who fell could reflect rater error, test-retest variability or real regression under stress. The practices here treat a score as one reading to be discussed, not as a label.

Do the studies of moral stages in children prove the stages apply to adults?

Those studies were of children, and they describe children more directly than adults. Walker (1980) assessed fourth- to seventh-grade children and found transitions to stage 3 reasoning only among children who had the hypothesized prerequisites. Blatt and Kohlberg (1975) gave children in schools twelve hours of moral discussion over twelve weeks, and the experimental group tended toward higher moral thinking than controls. A reader who wants to apply a stage claim to adult life can ask which ages and settings the evidence came from.

研究

関連する概念