Infantilization is a form of abuse in which a competent adult or young adult is treated like a child (Garber, 2011; Koops, 2012; Levy, 1943; Sharlin & Polansky, 1972). Viewed narrowly in the context of parenting, infantilization has been defined as a “parent’s inability to tolerate a child’s age-appropriate growth toward healthy independence” (Garber, 2011, p. 327). When competent individuals are treated like incompetent children, a number of negative outcomes are possible, among them a loss of sense of control over one’s life, increased depression and anxiety, anger directed toward caregivers or authority figures, lower self-esteem, suicidality, and damage to relationships (Garber, 2011; Kakihara et al., 2010; Robertson et al., 2016; Wang et al., 2007). Infantilization generally occurs with populations that lack power: older adults, women, communities of color, LGBTQIA+ individuals, adolescents, individuals with disabilties, and so on. Those who infantilize the powerless often have good intentions; infantilization is not necessarily malicious (Garber, 2011; LeMoyne & Buchanan, 2011; Schiffrin et al., 2014).

Infantilization by population

Older Adults: Negative stereotypes towards older adults are pervasive in our society. Agesist stereotypes represent older adults as being incapable of caring for themselves and cognitively impaired or slow, even when they are competent (Cherry et al., 2016; Chrisler et al., 2016; Higashi et al., 2012; Rittenour & Cohen, 2016; Williams & Warren, 2009). These assumptions may be intensified by caregivers who witness increased need for care with age. As a result, many older individuals, especially those confined to retirement homes or medical centers, are monitored excessively, are often excluded from decision making and even subjected to patronizing, sing-songy speech (Chrisler et al., 2016; Harwood et al., 1993; Higashi et al., 2012; Williams & Warren, 2009). Abuse of older adults may be “hostile (e.g., beliefs that older adults are incompetent, frustrating, and use too many societal resources) or benevolent (e.g., beliefs that older adults are warm, frail, and need to be cared for gently)” (Chrisler et al., 2016, p. 94) and can be so severe in some cases that it results in “utter despair, welcoming death” (Chane & Adamek, 2015, p. 71). Treating older adults as if they are incompetent children can often work as a self-fulfilling prophecy, causing older people to become weaker and more dependent (Allen et al., 2015), and such treatment, in turn, may lead to cognitive performance deficits (Hehman & Bugental, 2015) and poorer physical and emotional health (Wong & Waite, 2017). Because women are often subject to “benevolent sexism” (Chrisler et al., 2016, p. 94) and because they comprise the majority of older adults, these outcomes are especially likely to occur with older women (Chrisler et al., 2016; Montminy, 2005).

Young adults. Recent research has looked at the infantilization of young adults – adolescents and so-called “emerging adults” (Arnett, 2000). Although young adults can benefit to some extent from guidance and support, parents, educators and employers still tend to infantilize them, often considering them to be inherently incompetent, immature, risk-prone and untrustworthy, no matter what their actual capabilities might be (Bailey et al., 2015; Epstein, 2010; Kakihara et al., 2010; Raufelder & Kulakow, 2021). Epstein (2010) reported that adults ages 18 and over predicted that adolescents (ages 13 to 17) would score 48% on a comprehensive test of adult competencies (Epstein, 2010, p. 159); adolescents actually scored 90.6% on average – just 2.5 percentage points below the mean score for adults 18 and over (Epstein, 2010, p. 157).

Among young adults, infantilization is associated with lower self-esteem, dependency, depression, norm-breaking, feelings of incompetence, and trouble forming trusting relationships later in life (Epstein, 2010; Kakihara et al., 2010; Lee et al., 2015; Lindell et al., 2017; Schiffrin et al., 2014; Sharlin & Polansky, 1972). In recent years, concern has grown about parents who exert too much control over the lives of their offspring long past the childhood years – a parenting style sometimes called “helicopter parenting” or “hovering” (Assor et al., 2020; Cline & Fay, 1990; LeMoyne & Buchanan, 2011; Lindell et al., 2017; Savard et al., 2013). Parenting of this sort is associated with a number of unintended negative consequences. Some studies have found that college students whose parents exercised excessive control over their lives were taking more medication for depression or anxiety, had lower levels of well-being, and had lower perceived autonomy (LeMoyne & Buchanan, 2011; Marano, 2008; Schiffrin et al., 2014; Skenazy, 2009). A controlling parenting style might also be contributing to the artificial extension of children well past the teen years (Arnett, 2000; Epstein, 2013; LeMoyne & Buchanan, 2011; Savard et al., 2013).

Women. The perception and treatment of women varies enormously from culture to culture. Therefore, it is important to discuss the different cultural contexts in which infantilization of women occurs. Iran and Sudan each have eleven laws, and Jordan, Qatar, and Saudi Arabia each have ten laws restricting women’s rights (Htun et al., 2019). Saudi Arabia recently became the last country in the world to grant women the right to drive automobiles (Hubbard, 2017), but even in progressive Western cultures, women are often infantilized in various ways. In the United States, women were, until recently, largely excluded from many professions – the military, police work, the clergy and so on (Coleman, 2017; Gertz, 2017; Pellerin, 2015; Scarborough, 2016; Welch, 1982). Until January 2016, women were prohibited from ground combat roles in all branches of the U.S. military. (Coleman, 2017; Gertz, 2017; Pellerin, 2015). Nearly 10 percent of military positions had previously been off-limits to females due to the 1948 Combat Exclusion Policy, including serving as Army Rangers, Green Berets, Navy SEALs, Air Force parajumpers, and Marine Corps infantry (American Civil Liberties Union, 2017; Gertz, 2017; Pellerin, 2015; Scarborough, 2016). Often, the restrictions placed on women have been based on faulty ideas about the potential competence women have to perform as well as men – or even about women’s proper “place” in society as mothers and homemakers. A study by Ford and Harding (2010) on how women were treated at management conferences concluded that women at such conferences were often “subjected to processes of infantilization and seduction” (p. 503), were “persuaded that they lack intelligence and are irrational” (p. 517), and were told that they “should, in effect, get back to the kitchen” (p. 517). Discriminatory treatment of women in the workplace based on faulty stereotypes is unfortunately still prevalent in many countrries around the world to this day.

Outside the workplace, women across the lfiepsan and in various realms of life continue to face discriminatory treatment as well. Older women are victims of violence by spouses more often than older men (Montminy, 2005). As drivers, women are still subject to the “lingering stereotypes that women are less equipped to successfully maneuver an automobile” (Skinner et al., 2015, p. 64; see Albert, 1999; Berger, 1986) and to “the paternalistic view that women need protection” (Skinner et al., 2015, p. 64). A perception of incompetence is especially common when people are evaluating women of communities of color, an issue that will be discussed in the next section (Heilman et al., 1992). While women have made enormous strides over the past century, recent research continues to show that women are often treated as inferiors, at least in part because of unwarranted assumptions that are made about their inherent incompetence.

Communities of color and persons with disabilties: Prejudice against communities of color can be expressed in different ways; at times, it is expressed as infantilization – when, for example, a black male adult is addressed in a demeaning way as “boy” (Ash v. Tyson Foods & Inc., 2006). In popular culture, black women, too, “are infantilized [and] viewed as play things who are endlessly available and childlike” (Dagbovie-Mullins, 2013, p. 746). Sparked by Francis Galton’s (1892) investigations of individual differences in mental and physical capabilities, by the early 1900s, “racial psychology” was an accepted area of mainstream research psychology (Abramson & Lack, 2014; Viteles, 1928; Yerkes, 1921). Unfortunately, this specialty largely reinforced mainstream misconceptions and prejudices about the inferiority of communities of color. Feelings of superiority among the mainly white, male researchers were often expressed as a form of infantilization. Galton himself was clear on this point: “The mistakes the negroes made in their own matters were so childish, stupid, and simpleton-like, as frequently to make me ashamed of my own species” (Galton, 1892, p. 328). Even the distinguished psychologist Robert Yerkes showed paternalistic tendencies in his analysis of racial data on intelligence. Quoting the findings of his research associates at the U.S. Army’s Camp Dodge, Yerkes wrote, “‘it took all the energy and enthusiasm the examiner could muster to maintain the necessary attention, as there was a decided disposition for the negroes to lapse into inattention and almost into sleep’” (Yerkes, 1921, p. 705).

Even today, communities of color and LGBTQIA+ individuals in various professions experience both prejudice, often expressed as doubts about their expertise, and paternalism, a term derived from pater, the Latin word for father (Baker, 2015; Glass, 2001; Schroeder et al., 2017). As a result, these groups functioning in white-male-dominated cultures often have trouble being recognized as authorities (Moore, 1996). In the classroom, the legitimacy and authority students attribute to a teacher can be diminished when students learn that a teacher is gay (Moore, 1996). LGBTQ disadvantages in the workplace extend across multiple dimensions. In the STEM field, members of the LGBTQ community are more likely to face harassment and social marginalization in the workplace than non-LGBTQ colleagues (Cech & Waidzunas, 2021). They are also more likely to experience career limitations and devaluation of their professional expertise (Cech & Waidzunas, 2021).

Regarding people with disabilities, by definition, the individual with a disability cannot perform in all the ways able-bodied people can. It is hardly surprising, therefore, that people with disabilities are often infantilized in various ways. Offering a sight-impaired individual help in crossing a street can be seen as admirable, but unwanted helping can sometimes be perceived as offensive and infantilizing. In a recent study of attitudes toward disability, researchers concluded, “Sometimes expressions of warmth may reflect empathetic concern, but may also disguise paternalistic intentions, and the imposition of unwanted helping and control over those stereotyped as legitimate dependents like children, elderly people, and disabled people” (Nario-Redmond et al., 2017, p. 331). An early study of attitudes toward individuals with disabilities found that they “were viewed as lacking interactive skills, as more dependent on others for help” (Weinberg, 1976, p. 122). Thinking that members of any community of color are inferior in some way can often lead to infantilization. As Schroeder et al., (2017) put it, “Thinking of others as relatively mentally incapable, perhaps more like children than like adults, can lead people to treat others as relatively childlike” (p. 1123).

Mainstream adults. One unfortunate modern trend is the infantilizaion of adults through manipulation by marketers. The tendency for adults to want to revert to childhood is driven in part by marketers who tap into people’s fear of the ravages of aging – and, of course, the fear of death itself. Such marketing efforts are documented in Benjamin Barber’s, 2007 book, Consumed: How Markets Corrupt Children, Infantilize Adults, and Swallow Citizens Whole, and are noted in recent articles (e.g., Bernardini, 2013; McHugh, 2018). According to one source, marketers have concluded that the best way to get people to buy things they don’t necessarily need is to treat them “like spoiled or anxious children” – to promote “a prolonged lifetime of buying ever more toys” (Swift, 2007). This kind of manipulation extends even to the political realm – to “the purposeful dumbing down of politics to cater to a series of petulant and self-centered impulses” (Swift, 2007).

Although the present study doesn’t look specifically at the role that marketing plays in infantilization, we note that people can be infantilized at any point in their lives; infantilization is not restricted to young people. It is also not always a form of abuse. Treating people as if they are incompetent young children can come in different forms, can be done for different reasons, and can have a variety of different outcomes.

Is infantilization worth studying?

Various researchers over the years have singled out infantilization for study because it identifies a form of mistreatment that is both unique and can cause considerable damage. But why does treating someone like a child cause any damage at all? Shouldn’t it feel good to be coddled and protected, and isn’t the “inner child” worth preserving in adults? In Man-Child: A Study of the Infantilization of Man (Jonas & Klein, 1970), a psychiatrist and anthropologist teamed up to argue, echoing Freud, that, compared with other adult mammals, human adults are weak and childlike, quickly reverting to specific behavior patterns of early childhood when they play with actual children. So what’s wrong with infantilization?

Jonas and Klein were speaking about the natural tendencies of human adults. The problem arises when infantilization is imposed. Levy’s (1943) book, Maternal Overprotection, sheds some light on this issue. In the chapter entitled “Infantilization,” Levy looks at cases like the following: “a mother who helps her thirteen-year-old son dress; …another who still butters bread and gets water for a twelve-year-old; … another who punishes a thirteen-year-old son by putting him to bed in the afternoon” (Levy, 1943, p. 53). He also compares how long breast feeding continues in different ethnic groups. Over the course of the book, he shows that infantilization can lead to specific deficits and forms of psychopathology: prevention of social maturity, unhealthful dependence on the mother, poor decision making, difficulties in forming friendships, sleep disturbances, bowel and bladder problems, problems in forming romantic bonds – even sexual impotence and “a notable lack of interest in sports” (p. 198).

All of these problems have one basic feature in common: They are based on a deeply embedded belief that someone is incapable and incompetent; in each instance, potential and ability are suppressed. The engendering of feelings of inadequacy is, we believe, a special feature of infantilization that makes it worthy of study in its own right. What’s more, the modern emergence of helicopter parenting, documented in recent books such as Hara Marano’s (2008) A Nation of Wimps and Lenore Skenazy’s (2009) Free-Range Kids, along with determined efforts by today’s marketing professionals to infantilize mainstream adults (Barber, 2007; Swift, 2007), makes the study of infantilization even more pertinent now than it was in Levy’s day. The recent rise of authoritarianism and populism in Western countries is relevant here; new forms of infantilization occurring at the societal level might be contributing to such trends (see Discussion).

Measuring infantilization

As far as we are aware, validated tests that measure infantilization comprehensively do not exist. Neither the Mental Measurements Yearbook (Carlson et al., 2017) nor the PsycTESTS databaseFootnote 1 list any tests that claim to measure “infantilization” or related concepts, such as “juvenilization” or “paternalism.” Inspired by Levy’s (1943) book, Sharlin and Polansky (1972) wrote about developing a 42-item scale rating the infantilization that can occur in mother–child interactions, but they did not publish the full scale or any relevant validation data. Robey et al. (2006) modified the Implicit Association Test (Greenwald & Banaji, 1995) in order to determine whether staff serving people with disabilities “associated disability with child-like features” (Robey et al., 2006, p. 441). They called their test the “infantilization IAT” (p. 441).

Perhaps the most ambitious attempt to measure infantilization – at least in the context of parenting – was with the incorporation of “babying scales” into the Fels Parent Behavior Rating Scales, originally developed by Horace Champney in 1937 (Baldwin et al., 1949; Champney, 1941; Champney & Marshall, 1939). These were checklist-type scales used by “trained raters” (Baldwin et al., 1949) who observed parent–child interactions in people’s homes. Some of the scales looked specifically at “babying,” “protectiveness,” and “solicitousness,” warning that parents might “underestimate the child’s level of ability—or, they may find it more convenient to control his [sic] behavior by keeping him [sic] in a relatively helpless state” (Baldwin et al., 1949, p. 8). An overprotective parent, according to the Fels researchers, “attempts to preserve the child in a state of cocoon-like obliviousness to everyday life; all conceivable hazards and risks are kept from his [sic] path” (Baldwin et al., 1949, p. 8), while the solicitous parent is “the anxious or panicky parent who always expects disaster and whose apprehensions are expressed directly to the child” (Baldwin et al., 1949, p. 9). Note that these scales were developed well before Levy’s (1943) book was published. Note how similar the concerns were to modern concerns about helicopter parenting (Assor et al., 2020; Cline & Fay, 1990; LeMoyne & Buchanan, 2011; Lindell et al., 2017).

A number of validated tests exist that measure some aspects of infantilization. Among them are tests that measure aspects of emotional abuse,Footnote 2 such as the Practical Adolescent Dual Diagnostic Interview (Estroff & Hoffmann, 2001), the Childhood Maltreatment Measure (Roos et al., 2016), the Checklist for Child Abuse Evaluation (Feindler et al., 2003; Petty, 1988), the Emotional Maltreatment Screening Questionnaire (Bailey et al., 2012), the Childhood Trauma Questionnaire – Short Form (Bernstein et al., 2003), and the Checklist of Controlling Behaviors (Lehmann et al., 2012). In addition, in the English language hundreds of tests exist that measure various aspects of physical abuse with different populations, along with foreign translations of some of these tests. Tests of this sort, which sometimes measure sexual abuse, harassment, or other forms of abuse, in addition to physical abuse, include the Comprehensive Child Maltreatment Scale (Higgins & McCabe, 2001), the Composite Abuse Scale (Hegarty et al., 1999), the Revised Conflict Tactics Scales (CTS2) (Straus et al., 1996; Vega & O’Leary, 2007), and the Parent–Child Conflict Tactics Scale (Straus et al., 1998). We are also aware of the Paternalism Scale (Cicirelli, 1990), which measures belief in paternalism, and the Ambivalent Sexism Inventory (Glick & Fiske, 1996), which measures some aspects of paternalism.

There are also a number of parenting tests that assess, among other things, aspects of parenting that could be considered abusive or infantilizing, such as the Parenting Practices Measure (Abar et al., 2011), the Parent Questionnaire (Abar et al., 2012), the Perceived Parent–Child Relationship Scale (Fuligni & Eccles, 1993), the Quality of Parent–Child Relationship Index (Boney-McCoy & Finkelhor, 1995), the Young Caregiver of Parents Inventory (Pakenham et al., 2006), the Parental Control Measure (Mogro-Wilson, 2008), and the Home Literacy Environment Measure (He et al., 2014). Some tests look narrowly at one form of infantilization. The Expected Limits for Adolescent Drivers Measure (Laird, 2014) looks at parental restrictions on driving; the Parental Media Restriction Measure (Parkes et al., 2013) looks specifically at restrictions on media usage; and the Children’s Perceptions of Restricted Access Structured Interview (Fisher & Birch, 1999) looks specifically at parental restrictions on diet. The internet has complicated matters when it comes to testing, because it allows anyone to post a test of any sort, with the general public having no way to distinguish a scientifically validated test from an informal one., a dating website, has posted thousands of informal tests, and the total number of informal tests available to online users will likely continue to proliferate at an increasing rate. While preparing this essay, we easily identified five informal online tests that claim to measure emotional abuse, five such tests that claim to measure physical abuse, and three such tests that measure aspects of the controlling parenting style. Given the absence of standards for rating internet content – even critically important content such as news stories – it is not clear how internet users can be protected from potentially misleading and perhaps even harmful online tests in years to come.

Although the larger problem has no easy solution, we believe that internet users should at least have access to free, empirically validated online tests, and, in that spirit, the first author of the present study, working with various associates, has developed a number of such tests over the years in areas such as creativity (Epstein & Phan, 2012; Epstein et al., 2008, 2013a, 2013b), mental health screening (Epstein & Muzzatti, 2011; Epstein et al., 2017), sexual orientation (Epstein et al., 2012), parenting (Epstein et al., submitted), stress management (Epstein et al., submitted), motivation (Epstein et al., 2022), and relationship skills (Epstein et al., 2013a, 2013b, 2016).

The present paper focuses on a new test in this series, one that measured 15 different categories of infantilization, all of which were derived from empirical studies. With data collected from a large online sample of people from 153 countries (most from the U.S.), we were able to evaluate some aspects of the reliability and validity of the new test. We used these data to rank order the 15 areas of infantilization according to how well they predicted self-reported life outcomes, such as how depressed people were and how much control they felt they had over their lives, the goal being to determine which kinds of infantilization might do the most harm. Because the participant sample was so large, we were able to look at various demographic issues, asking questions such as: Which demographic groups are more infantilized, and to what extent does infantilization change over the life span? Because our data were collected over a period of 8.5 years, we were able to ask whether there was any trend in infantilization over time.


Study design

Our study employed a “concurrent study design” that provided convergent validity evidence with related measures, following the most recent guidelines of Standards for Educational and Psychological Testing, co-published by the American Educational Research Association, the American Psychological Association, and the National Council on Measurement in Education (American Educational Research Association, 2014). Specifically, we sought to measure the strength of the relationships between our test scores and the scores on answers to criterion questions. This design was concurrent because we obtained test scores and criterion measures at the same time, a strategy that avoids possible temporal confounds. Results from studies employing this design are considered especially robust when the pattern of relationships between test scores and criterion measures proves to be consistent across different demographic groups (American Educational Research Association, 2014, pp. 17–18).


The present study analyzed data obtained from a sample of 32,118 cases (after cleaning) collected online between October 25, 2011 and February 18, 2020. After the questionnaire was posted (at, no efforts were made to promote it, and we therefore had no control over the nature of the sample (see Discussion).

The raw data were cleaned as follows: Cases in which fewer than half the questions had been answered were removed, and so were cases in which participants indicated that they had an English fluency level under 6 (on a scale of 1 to 10, where 10 was the highest possible fluency). In addition, based on dates and times, as well as demographic information, we removed the data of participants who took the test more than once on the same day, keeping only data obtained on the first trial in which they answered more than half the questions.

Of those, 20,993 (65.4%) of our participants identified themselves as female, 9,991 (31.1%) as male, and 793 (2.5%) as other; gender was not given for 341 (1.1%) of our participants. The mean age of our participants was 27.6 (median = 24.0, range 12–90). A minimum-age cutoff of 12 was used given that the Gunning-Fog reading level of our test questions was 6.3. 12 is the typical age at which students in the U.S. have completed the 6th grade, so this age cutoff was allowed by the Institutional Review Board of the host institution.

22,017 (68.6%) of our participants identified themselves as White, 2,856 (8.9%) as Asian, 2,284 (7.1%) as Other, 2,138 (6.7%) as Black, 2,072 (6.5%) as Hispanic, and 249 (0.8%) as American Indian; information about race or ethnicity was not given by 502 (1.6%) of our participants.

3,769 (11.7%) of our participants said they lacked a high school education; 10,844 (33.8%) said they had earned a high school degree; 5,537 (17.2%) said they had earned an associate’s degree; 7,883 (24.5%) said they had earned a college degree; 2,999 (9.3%) said they had earned a master’s degree; and 745 (2.3%) said they had earned a doctoral degree; information about education level was not given by 341 (1.1%) of our participants.

22,822 (71.1%) of our participants identified themselves as straight, 1,527 (4.8%) as gay/lesbian, 3,452 (10.7%) as bisexual, 1,988 (6.2%) as unsure, and 1,237 (3.9%) as other; 1,092 (3.4%) of our participants did not provide information about their sexual orientation.

Due to a technical problem with our internet service provider, country of origin was not recorded for 9,230 of our participants (28.7% of our sample). 13,853 of our participants (60.5% of the total number of people for whom we had location information) were from the U.S.

Test instrument

Participants responded to six criterion questions concerning sense of control, happiness, depression, anger, and personal and professional success. Questions were answered on a 10-point Likert scale, as follows: (1) “How much control do you feel you have over your life?” (scale from No control to Complete control). (2) “How much success have you had lately in your personal life?” (scale from Low to High). (3) “How much success have you had lately in your professional life?” (scale from Low to High). (4) “How happy and fulfilled are you?” (scale from Very unhappy to Extremely happy). (5) “How depressed have you felt in recent months?” (scale from No depression to Severe depression). (6) “How angry have you felt in recent months?” (scale from No anger to Extreme anger).

The study employed a new questionnaire called the Epstein-Dumas Infantilization Inventory (EDII), which consisted of 142 items in 15 different categories, presented to all participants in the same random order. The questionnaire was only available in English. The 15 categories were derived from a review of relevant literature available when the questionnaire was being developed. See Table 1 for a list of the categories, sample items, and an updated list of literature relevant to each of the categories. The delineation of these categories was based on the judgment of the researchers.

Table 1 15 Categories of Infantilization Assessed Using the EDII


Participants were first given basic information about infantilization – defined as being “treated like a child” – in different populations. They were told that, “the lower your score, the better” and that “mainstream adults in the U.S. generally score quite low,” that “active-duty U.S. Marines score about 24 percent,” that “incarcerated felons score… about 35 percent” and that “highly restricted children… score about 70 percent.” Participants were then informed that, “ When competent people are treated like children, they often react with depression or anger. Knowing how much you are infantilized can be an important first step toward improving your life.” Following the instructions, participants were asked basic demographic questions, along with the six criterion questions. Participants were then asked, “Who, if anyone, is trying to control your life?”

These background questions were followed by the test itself, which began as follows: “Over the past year, has anyone restricted your activities in any of the following ways? Click only on the statements with which you agree.” 142 items followed, and any number of them could be selected. Item 143 allowed people to type in a restriction that was not included in the previous items. See Table 1 for a sampling of items in each of 15 categories of restriction. The score was determined by computing the percentage of items (out of 142) that had been selected.

Upon completing the test, the participant clicked “Submit” and was shown his or her score on a percentage scale from 0 to 100 that indicated how that score compared with those of various groups (noted above). A brief explanation of the results was included notifying participants that the higher their score is the more they are infantilized. Instructions on how to contact the researchers were also included. Participiants were told, “If you're worried about the amount of control you're subjected to, consider making changes in your life that will remove controlling influences. If you're feeling depressed or angry, a qualified counselor or therapist can help you cope and also help you make changes in your life.” Websites where participants could research counselors and therapists were then listed and information regarding where participants could receive referrals for mental health problems was included.


Reliability and validity

Internal consistency reliability was relatively high: Cronbach’s alpha was 0.96, and the Guttman split-half value was 0.90. Because the test was administered over the internet and because we preserved the anonymity of our participants, we were not able to retest participants and thus could not calculate test–retest reliability.

Test scores were appropriately correlated (negatively correlated for positive outcomes, such as sense of control, and positively correlated for negative outcomes, such as anger) with answers to criterion questions about sense of control over one’s life (Spearman’s ρ = − 0.37, P < 0.001), anger (ρ = 0.26, P < 0.001), happiness (ρ = − 0.27, P < 0.001), depression (ρ = 0.24, P < 0.001), personal success (ρ = − 0.24, P < 0.001), and professional success (ρ = − 0.19, P < 0.001).Footnote 3 Note that the questionnaire measured only self-reported restrictions; we had no independent measures of these restrictions.

Descriptive statistics

Gender. An effect was found for gender (MMale = 23.3 [SD = 19.2], MFemale = 27.0 [19.0], MOther = 33.9 [21.4], Kruskal-Wallis H = 476.7, P < 0.001), and, in a pairwise comparison, females proved to be significantly more infantilized than males (MMale = 23.3 [SD = 19.2], MFemale = 27.0 [19.0], Mann-Whitney U = 9.1 × 107, P < 0.001, d = 0.19), and those identifying as “other” under gender proved to be significantly more infantilized that males and females combined (MOther = 33.9 [21.4], MMale/Female = 25.8 [19.1], U = 9,457,059.5, P < 0.001, d = 0.40).

Table 2 shows scores on each of the 15 subscales, broken down by gender. Significant gender differences were found in 14 of the 15 subscales – in all but Restrictions on Risky Behavior. In the pairwise comparison between males and females, mean score differences were greatest in the Emotional Abuse category, with females being subjected to more abuse. In the pairwise comparison between “others” and males and females combined, again, mean score differences were greatest in the Emotional Abuse category, with others being subjected to more abuse.

Table 2 Mean Infantilization Scores by Gender

Education and race. Effects were found for level of education completed (MNone = 33.2 [22.0], MHighSchool = 28.5 [20.0], MAssociate’s = 26.0 [18.5], MCollege = 22.3 [17.0], MMaster’s = 19.4 [15.6], MDoctorate = 18.7 [15.5], H = 1,424.5, P < 0.001; ρ = − 0.21, P < 0.001), with total infantilization scores decreasing as educational level increased (Fig. 1), and for race/ethnicity (MAmericanIndian = 28.4 [21.3], MAsian = 25.5 [19.6], MBlack = 25.7 [18.1], MHispanic = 26.9 [19.0], MOther = 28.3 [20.6], MWhite = 25.8 [19.1], H = 44.23, P < 0.001). In a pairwise comparison, non-whites proved to be significantly more infantilized than whites (Mnon-White = 26.6 [19.5], MWhite = 25.8 [19.1], U = 1.0 × 108, P < 0.01, d = 0.04), although both the mean difference (0.8) and the effect size were small (d = 0.04).

Fig. 1
figure 1

Mean total infantilization score by education level completed. The more education people had, the lower their infantilization score. 95% confidence intervals are shown

Sexual orientation. An effect was found for sexual orientation (MStraight = 24.3 [18.2], MGay = 27.5 [21.1], MBisexual = 31.7 [21.4], MUnsure = 30.2 [20.8], MOther = 32.9 [21.5]), H = 644.8, P < 0.001) and, in a pairwise comparison, non-straight participants (gay, bisexual, unsure, and other combined) proved to be significantly more infantilized than straight participants (MStraight = 24.3 [18.2], MNon-straight 30.7, [21.3], U = 7.7 × 107, P < 0.001, d = 0.32). People in non-labeled sexual orientation categories (other and unsure combined) proved to be more infantilized than people in conventionally labeled sexual orientation categories (MStraight/Gay/Bisexual = 25.4 [18.9], MOther/Unsure = 31.2 [21.1], U = 3.7 × 107, P < 0.001, d = 0.29).

Age and country. Figure 2 shows mean total scores as a function of age. Infantilization appears to decrease with age (ρ = − 0.23), although it may increase again in old age. There were too few people in the oldest age groups in the present study to identify an upward trend with any degree of confidence. Using the median age of our participants, 24.0, to split our sample into two groups, we again found an effect for age (M12-24 = 29.8 [20.4], M25-90 = 22.1 [17.1]), U = 95,763,581.0, P < 0.001, d = 0.41), and we were able to look at age effects in each of our 15 infantilization categories (Table 3). Significant age effects were found for all 15 categories, with young people proving to be especially infantilized in four of those categories: Physical Restrictions (mean difference = 1.0), Emotional Abuse (0.7), Restrictions on Appearance (0.8), Violations of Privacy (0.7), and Restrictions on Education (0.7).

Fig. 2
figure 2

Mean total percentage infantilization score as a function of age. Infantilization appears to decrease with age, although it may increase again in old age. There were too few people in the oldest age groups in the present study to identify an upward trend with any confidence

Table 3 Total Infantilization Scores Above and Below Median Age

An effect was found for country, with participants from the United States proving to be more infantilized than participants from other countries around the world, although both the mean difference (0.7) and the effect size were small (MUnitedStates = 26.6 [19.6], MOther = 25.5 [18.9], U = 1.2 × 108, P < 0.001, d = 0.06).

Employment and therapy status. Effects were found for participant’s employment status, with unemployed participants more infantilized than employed participants (MEmployed = 23.2 [17.9], MNotEmployed = 29.1 [20.2], U = 8.9 × 107, P < 0.001, d = 0.31), as well as for therapy status, with people who were in therapy at the time they took the test more infantilized than people who were not in therapy (MInTherapy = 29.1 [19.9]), MNotInTherapy = 24.8 [18.8], U = 8.79 × 107, P < 0.001, d = 0.22).

Frequency of infantilization by category. By far, items in the Emotional Abuse category were selected more often than items in the other 14 infantilization categories (Table 4), and the most commonly selected item was “Criticized you,” which was selected 21,651 times. Figure 3 shows the mean frequencies of category selection in order from the most frequently selected category (Emotional Abuse) to the least frequently selected category (Restrictions on Medical or Mental Health Treatment). Items selected infrequently were generally applicable to relatively small numbers of participants. The two least frequently selected items were “Tried to make you join or quit a gang” (n = 126) and “Tried to prevent you from undergoing shock therapy” (n = 95). Participants could also type in their own items. The most commonly added item, phrased in slightly different ways by different people, was “diapering me” (n = 23).

Table 4 Number of Times Each Type of Infantilization was Selected, in Descending Order by Adjusted Total
Fig. 3
figure 3

Mean percentages of the number of items selected by each participant within each infantilization category, in order from highest to lowest values. 95% confidence intervals are shown

Sources of infantilization. Mothers were identified most frequently (that is, by 54.1% of our participants) as sources of infantilization, with fathers second (35.2%) (z = 48.3, P < 0.001) (Fig. 4). Because participants could select multiple sources of infantilization, we could not meaningfully explore the relationship between sources and types of infantilization.

Fig. 4
figure 4

The percentage of participants choosing each of 14 different sources of infantilization. Note that participants could choose multiple sources. 95% confidence intervals are shown

Regressions and factor analysis

Stepwise regression analyses showed that, of the 15 types of infantilization we measured, in single-component models Emotional Abuse emerged as the best predictor of negative outcomes for all six of our criterion variables (Table 5).Footnote 4 In multiple-component regression models, Violations of Privacy, Restrictions on Education, Restrictions on Social Relationships, Medical or Mental Health Restrictions, and Restrictions on Work tended to be somewhat predictive of negative outcomes (Table 6). Physical Restrictions was a relatively poor predictor, however. Notably, in a 13-component exploratory factor analysis, a component one can reasonably label “abusive speech” (corresponding closely to our Emotional Abuse category) proved to have the highest loadings (Table 7).

Table 5 Linear Regressions Prediciting Criterion Variables from Competency Scores (Best Predictor Standardized β)
Table 6 Linear Regressions Prediciting Criterion Variables from Competency Scores (Top Three Predictors Standardized β)
Table 7 Factor Loadings for the Test Items

Changes over time

Because our data were collected over a period of ten years, we looked for changes in the data over time. Table 8 shows relevant means for all years in which we had full years of data – that is, from 2012 to 2019. Over this period, there was a 14.2% increase in infantilization scores (from 22.5 to 25.7). Although we lack data for the full years of 2011 and 2020, it is notable that the total score in 2020 [26.0] was 30.1% higher than the total score from 2011 [20.0] (U = 6958.00, P < 0.001).

Table 8 Differences in Mean Total Score and Demographics by Year

At first glance, one might surmise that the increase in total scores can be accounted for by changing demographics. Between 2012 and 2019 (Table 8), the average age of participants decreased from 32.3 to 26.5, the percentage of participants who were employed decreased from 61.0 to 45.3, and the percentage of female participants increased from 34.8 to 66.5. However, a linear regression suggested that the percentage of variance in total scores accounted for by the seven variables listed in Table 8 was only 8.3 (P < 0.001), leaving open the possibility that our test might have been detecting a societal increase in the rate of infantilization occurring over the course of our study.


We believe that the current study sheds important new light on the nature and degree of infantilization that occurs across the life span. For one thing, it shows, surprisingly, that infantilization is not just something that happens with adolescents or young adults; rather, it shows that infantilization occurs throughout the life span, that it decreases with age, and that it might increase again in old age. Unfortunately, we had too few participants (only 90 people over age 70, out of 32,118 participants in total) to detect a possible increase in infantilization in old age with any degree of confidence – one obvious limitation of the current study that needs to be addressed in future research.

The present study lends support to what other researchers have found: that infantilization is common among vulnerable groups (e.g., Epstein, 2010; Ford & Harding, 2010; McNamara et al., 2016; Sharlin & Polansky, 1972; Skinner et al., 2015). In the present study, that concept was shown to apply to females, communities of color, individuals with disabilities, adolescents, the less educated, the unemployed, and people undergoing therapy. More specifically, we found that infantilization was more common for people identifying as “other” under gender or as “other” or “unsure” under sexual orientation. We found that the less education people had, the more they were infantilized (Fig. 1); this could be an age effect, but there is evidence that people with less education are indeed subjected to more criticism, negative attitudes, or abuse (Kuppens et al., 2018; Phillips & Hempstead, 2017).

Our most notable finding, we believe, is that Emotional Abuse was not only the type of infantilization people reported most frequently (Table 4, Fig. 3), it was also the best predictor of negative outcomes for all six of our criterion variables: self-reported estimates of sense of control over one’s life, anger, happiness, depression, personal success, and professional success (Table 5). Verbal abuse emerged as a strong component in our factor analysis (Table 7). Our study was correlational in design, however, so these results shed no light on the possible causal relationship between emotional abuse and negative outcomes. We believe this is one of those rare situations, however, in which it is reasonable to speculate that the variable in question probably does play a causal role (Azar, 2006). It would be unethical to conduct a true experiment to test that possibility; however, a wide variety of studies employing different methodologies and looking at different populations all point to emotional abuse as a likely cause of negative outcomes and even suggest that emotional abuse does more damage than physical abuse.

In a study that relied on retrospective reports of various forms of abuse that had been experienced by more than 6,000 adults, Turiano et al. (2017) found that memories of emotional abuse – but not of physical abuse – were predictive of lower levels of positive affect and higher levels of negative affect in adulthood. These findings, they argued, were not surprising because children exposed to emotional abuse tend to have underdeveloped emotion regulation skills and imitate the dysfunctional emotions of their parents (see Turiano et al., 2017, p. 289). In a study employing both quantitative and qualitative methods, along with statistical modeling techniques, Gavin (2011) found that emotional abuse was a better predictor of negative outcomes than physical abuse, perhaps, Gavin wrote, because “social or emotional pain, embarrassment, or ostracism can be relived time after time, whereas it is difficult to mentally re-create physical pain” (Gavin, 2011, p. 504). Based on an analysis of the content of interviews, Gavin speculated about the causal path that emotional abuse follows in reducing well-being in adults, noting that such abuse “can erode self-confidence, sense of self-worth, and self-concept” (p. 507). (For other viewpoints on the possible causal pathways that both emotional abuse and, more generally, infantilization, play in negative outcomes in adulthood, see Epstein (2010), Gibb and Abela, (2008), Jonas and Klein, (1970), Koops (2012), Levy (1943), and Mojallal et al. (2021).

Our own findings about the high predictive value of Emotional Abuse and the relatively low predictive value of Physical Restrictions (see our definition of Physical Restrictions in Table 1) support the findings of both the Turiano et al. (2017) and Gavin (2011) studies, and all three of these studies have important implications both for public policy and future research. As Sims (2008) has noted, emotional and verbal abuse attract relatively limited media coverage compared to the frequent reporting that is devoted to physical abuse. On its surface, physical abuse seems to be more damaging – mainly because the damage can often be directly observed by others – but emotional abuse might actually produce more lasting pain and dysfunction. This disparity in media coverage, Sims says, leaves victims of emotional abuse “unlikely to see their own experiences reflected in the stories of abuse they read” (p. 398) and might deprive them of greatly needed treatment resources, perhaps because they themselves are unable to recognize their experiences as a form of abuse.

Although our data suggest that residents of the U.S. experience more infantilization than residents of other countries, the design of our study does not offer any substantive insights regarding either cultural or country differences in infantilization or its outcomes. An intriguing study by Chao and Aque (2009) points to how important cultural differences can be in exploring a topic like infantilization. Their study found dramatic differences in how Asian immigrant and European American high school students perceived and reacted to controlling behavior by parents. Because different cultures instill very different views about the level of respect young people should show toward parents, grandparents and other adults (see Huang, 2013), controlling behavior by adults might be viewed by some young people as abusive and by others as protective and nurturing (Chao & Aque, 2009; cf. Andreoletti et al., 2015). Cultural differences of this sort apply not only to young people who are being asked about their current lives but also to older adults being asked to recall and interpret events from their childhood. The present study does not address these important cultural issues.

Our results suggest that infantilization has been increasing over time. In other words, to quote Sharlin and Polansky (1972) once again, societies might be encouraging more and more people to become “less competent and self-sufficient than [they] otherwise might be” (p. 93). Perhaps the increase in infantilization scores we observed is another facet of other changes that appear to be occurring in much of the world: increases in authoritarianism, populism, nationalism, abusive political rhetoric, internet abuse, and partisanship. (Ágh, 2018; De Castella, 2011; Fortson et al., 2007; Kasparov & Halvorssen, 2017; Velez & Lavine, 2017; Wadley, 2011; Walter, 2016; Wood, 2017). Other researchers have already noted the steady and somewhat dramatic increase in helicopter parenting that has taken place over the past decade (LeMoyne & Buchanan, 2011; Lindell et al., 2017; Marano, 2008; Schiffrin et al., 2014; Skenazy, 2009).

Limitations, strengths, and future directions

The main limitation of this study – the fact that it was conducted online with a convenience sample – is also, we believe, a strength in some respects. Internet studies allow researchers to collect data from large and diverse samples; in the present instance, from 32,118 people in 153 countries. Even without control over the sample, we think this sort of coverage is preferable to the traditional sample: sophomores from the subject pool at a single university (Sears, 1986). Our sample may not be representative of people in general. A test like ours is likely to attract test takers who have special concerns about infantilization; in other words, infantilization scores in the general population might be lower than the ones we found – by how much we cannot say. As noted earlier, because we used an online sample and took care to preserve the anonymity of our participants, we were unable to assess test–retest reliability.

On the positive side, even though our participants were self-selected, we have no reason to doubt the validity of the specific patterns we found in our data, such as the indicators we found that Emotional Abuse may be especially harmful, or multiple findings suggesting that vulnerable groups are more likely to be infantilized. Even if our online instrument attracted test takers who had special concerns about being infantilized, it is hard to imagine why self-selection of that sort would produce the clear and consistent patterns we found in demographic differences, factor loadings, the predictiveness of subscales, and so on.

Furthermore, there is a growing and impressive body of evidence demonstrating that people are more honest when they complete anonymous surveys online than they are when answering questions administered using any other survey methodology, especially when people are being asked questions about socially sensitive issues (e.g., Dillman et al., 2014; Durantet al., 2002; Dwight & Fiegelson, 2000; Joinson, 1999; Kaplan & Saccuzzo, 2009; Krumpal, 2013; Ong & Weiss, 2000; Robertson et al., 2017; Trau et al., 2013; Villarroel et al., 2006; cf. Reips, 2012). In this respect, it is possible that the EDII has measured infantilization more accurately than any assessment tool has previously.

As previously mentioned, age was another important limitation in our study. We may have seen a glimpse of increasing infantilization among older adults (Fig. 2), but we did not have enough participants over age 70 to draw any conclusions about people over that age with any degree of confidence. Given the extensive literature on infantilization among older adults (see Introduction), future research using the EDII or similar instruments should be conducted to explore this issue.

The validity of our results is limited by the fact that we relied on self-reporting for our criterion measures. Other research, conducted over a period of decades, has found that self-reported data are often good predictors of behavior, especially when test items pinpoint specific behaviors, as almost all our test items were designed to do (see Gnambs & Kaspar, 2015; Krumpal, 2013; Boyatzis, 1982; Smith & Smith, 2005; Spencer & Spencer, 1993; Wood & Payne, 1998). Whatever one’s reservations about the value and validity of self-report measures, self-report remains one of the most commonly used measures in the social and behavioral sciences (Harrison et al., 1996; Howard, 1994).

How comprehensive was our list of 142 types of infantilization? We addressed this issue by allowing our participants to type in ways they were infantilized that were not covered in our questions. Among the more common suggestions: diapering (mentioned 23 times), criticizing the way the participant parented his or her own children (9 times), and “gaslighting” (a trendy term suggesting that someone is trying to make you feel like you’re crazy, mentioned 4 times). As we continue to collect data on infantilization in future years, we might add items of this sort to revisions of the inventory; at the moment, however – based on the additions our participants suggested – we believe the current set of items does a good job of covering the wide range of ways in which people are infantilized.

At present, the EDII sheds no light on how to reduce or eliminate the infantilization many people experience in their lives, although our data might be helpful to people who have addressed this issue (e.g., Andreoletti et al., 2015; Garber, 2011; Koops, 2012). If, as our data suggest, infantilization has been increasing in recent years, and if this increase is related to what appear to be real increases in intolerance, authoritarianism, partisanship and other indications of breakdowns in interpersonal relationships, finding ways to fix such problems should be considered a pressing matter by both social scientists and policy makers.