Lower subjective status is associated with reduced satiation and satiety among children and adolescents: A laboratory study

Abstract

Objective:

Subjective status represents one’s perceptions of their social/socioeconomic standing compared to others. Low subjective status is associated with higher energy intake and body mass, independent of objective status indicators. Low subjective status could be blunting sensations of satiation/satiety, which may spur energy intake. However, there is limited research directly examining the role of subjective status on satiation and satiety, especially in children. We cross-sectionally examined whether subjective socioeconomic status (SSES) and subjective social status (SSS) are independently associated with satiation and satiety. We hypothesized that children/adolescents reporting lower SSES or SSS would report reduced satiation and satiety after energy intake.

Methods:

While fasted, children/adolescents (N=133, Age_Mean=13.7±3.0 years) consumed a standardized breakfast shake. Participants reported their satiation (difference in pre- and post-shake appetite ratings divided by percentage of shake consumed) and satiety (ratings of hunger across a 90-minute period following shake consumption).

Results:

Lower SSS was associated with reduced satiation (B=0.04, 95%CI: 0.0003, 0.08) and both lower SSS and SSES was associated with greater hunger across 90-minutes (SSS: B=−8.06, 95%CI: −12.94, −4.32; SSES: B=−6.57, 95%CI: −12.35, −1.52). Higher SSES was also associated with lower odds of an unsatiated, yet slowly increasing (OR=0.61, 95%CI: 0.42, 0.90) or decreasing (OR=0.63, 95%CI: 0.41, 0.96) hunger trajectory.

Conclusions:

Lower subjective status is associated with reduced satiation and satiety among children/adolescents. Blunting of these sensations in early life may help explain the broader relationships between low subjective status, excess energy intake, and higher body mass, as well as socioeconomic disparities in these outcomes.

1. Introduction

Childhood obesity is rising (Hu & Staiano, 2022), particularly among socioeconomically disadvantaged children. Among the most socioeconomically disadvantaged children, those born in the mid-2000s had a 15% higher cumulative incidence in obesity across primary school compared to children born in the early 1990s (Cunningham et al., 2022). This rise in disparities in childhood obesity may partially be explained by eating behaviors and dietary practices that promote excess energy intake. Many social determinants associated with children’s obesogenic eating behaviors, like low household SES (Kininmonth et al., 2021), discrimination (Raney et al., 2023), and social exclusion (Pink et al., 2024) involve subjective perceptions of disadvantaged social status compared to others. In this study, we examine associations between low subjective status and blunting of satiation (fullness experienced from food intake) and satiety (fullness maintained between meals) among children—processes that may motivate overeating and excess energy intake—using laboratory-based measures.

1.1. Subjective status

Subjective status refers to one’s perceived social standing relative to others (Adler et al., 2000; Singh-Manoux et al., 2005), and can be based on different domains of status. Subjective social status (SSS) is based on perceived social standing on domains like respect, popularity, or importance. Subjective socioeconomic status (SSES) is based on perceived social standing regarding access to socioeconomic resources and opportunities, such as income, educational attainment, or occupational prestige. These social and socioeconomic dimensions capture distinct aspects of perceived status and may have independent contributions to health outcomes (Galvan et al., 2023). Both forms of subjective status are based on one’s internalized perceptions or evaluations of their status relative to others and differ from objective status (actual respect one receives or their actual socioeconomic resources).

Subjective status may be experienced differently across developmental stages. Children demonstrate understanding of the concepts of social class and the value of having more money at age 5 to 6 (Rauscher et al., 2017). Around age 9, children exhibit stronger associations between subjective and objective socioeconomic status and ascribe status to desirable material possessions and money (Peretz-Lange et al., 2022; Rauscher et al., 2017). Children also tend to hold generally positive evaluations of their own social status (Mandalaywala et al., 2020), but SSES declines as children get older (Amir et al., 2019; Peretz-Lange et al., 2022). While childhood subjective status may be more dependent on perceived family or parental social class, subjective status in adolescence may be informed by a combination of family status and emerging sense of personal standing compared to peers (Goodman et al., 2001).

Measures of subjective status may have advantages over objective status when examining the role of one’s status on eating behaviors, especially from a developmental perspective. First, youth may vary in how accurately they can assess their family’s objective socioeconomic circumstances (Ridolfo & Maitland, 2011). Even in households facing food insecurity, children may be shielded from the negative effects of food insecurity by caregivers (McIntyre et al., 2003). Children’s subjective evaluations of family socioeconomic position may be more sensitive for gauging household socioeconomic conditions in such cases. Second, compared to adults, indicators of objective socioeconomic status (e.g., income, occupation) may be a less salient reflection of personal status among school-aged children, whose perceived status may be more informed by social factors like respect and acceptance relative to peers. Third, while indicators of objective SES typically focus on current conditions (e.g., current income), subjective status is based on a temporal ‘averaging’ of one’s past, present, and anticipated status (Singh-Manoux et al., 2005), which may better reflect perceived status across developmental stages.

Overall, subjective status is believed to represent a holistic and internalized evaluation of the sufficiency or inadequacy of one’s social and socioeconomic circumstances that is not captured by measures of objective SES (Singh-Manoux et al., 2005). Accordingly, a growing body of research has demonstrated that low subjective status predicts poorer health outcomes independent of objective socioeconomic status (SES) (Cundiff & Matthews, 2017; Singh-Manoux et al., 2005; Zell et al., 2018). Low subjective status is also associated with poor health outcomes among adolescents (Quon & McGrath, 2014).

1.2. Subjective status and eating behaviors

The relationship between low subjective status and health extends to eating behaviors and adiposity (see Cheon, Bittner, & Pink, 2025 for a review). Lower subjective status is associated with higher body mass index (BMI) and increased odds of overweight/obesity in adults (Cundiff & Matthews, 2017; Rahal et al., 2020; Tang et al., 2016). These patterns have also been observed among adolescents (Menon et al., 2019; Noh et al., 2014; Quon & McGrath, 2014; Tan et al., 2018), suggesting that subjective status may contribute to adiposity early in life. This relationship between subjective status and increased likelihood of overweight/obesity may be explained by low subjective status motivating behaviors that may lead to excess energy intake (Cheon et al., 2025). Lower subjective status and perceived disadvantage have been associated with poor dietary habits, such as lower intake of fruits and vegetables, greater consumption of foods high in fat and sugar, less compensation for larger meals, and selection of larger portion sizes (D’Hooge et al., 2018; Ghaed & Gallo, 2007; Rahal et al., 2023; Sim, Lim, Forde, et al., 2018; Wijayatunga et al., 2019; Zheng et al., 2024).

Experimental studies that manipulated subjective status have supported causal relationships between low subjective status, obesogenic preferences, and behaviors toward food. Exposure to experimental manipulations of low subjective status or perceived deprivation promotes higher preferences for palatable, energy-dense foods, greater energy intake (Bratanova et al., 2016; Cardel et al., 2016; Cheon & Hong, 2017), selection and consumption of larger portion sizes (Sim, Lim, Forde, et al., 2018), and higher taste-based sensitivity to the presence of energy in beverages (Cheon et al., 2018; Lim et al., 2020).

Although most of the prior research on subjective status and eating behaviors have focused on adults, these relationships have also been identified among adolescents. Among adolescents, lower SSES is associated with poorer dietary habits (Belardinelli et al., 2022; Elgar et al., 2016; Quon & McGrath, 2015). Among youth from lower SES households, lower SSES is associated with greater severity of hyperphagic behaviors (Smith et al., 2023). Experiences that threaten perceived social status have also been associated with potentially obesogenic eating behaviors and adiposity. For instance, among youths reporting distress from teasing, lower reported SSS was associated with greater tendencies to eat in the absence of hunger as well as higher BMI and adiposity (Cheon et al., 2024). Similarly, among children who report higher levels of social anxiety, greater stress response to social exclusion was associated with greater energy intake from a subsequent snack—a behavior predicting higher BMI approximately 1.5 years later (Pink et al., 2024).

1.3. Subjective status and blunting of satiation and satiety

One plausible yet understudied explanation for why low subjective status may motivate excess energy intake and preferences for energy-dense foods is through the disruption of satiation and satiety. Feelings of low subjective status could blunt satiation and/or satiety, spurring higher energy intake, selecting larger portion sizes, and desire to consume energy-dense snacks. An explanation for why lower socioeconomic position is associated with obesogenic food preferences and eating behaviors is heightened vulnerability to hunger among people experiencing socioeconomic disadvantage (Nettle, 2017). Given that experimental manipulations of low SSES stimulate appetite and energy intake independent of negative affect (emotional eating), Cheon and Hong (2017) speculated there may be a potential overlap between feeling deprived of social/socioeconomic resources and deprivation of food resources (e.g., hunger). Other authors have also proposed that desire for financial resources shares a bidirectional relationship with desire for food (Briers & Laporte, 2013). If this is true, low subjective status may also dampen satisfaction or satiety from food intake.

A study by Sim and colleagues (2018) provided preliminary support for this supposition. Male university students who completed an overnight fast completed a randomized cross-over trial in which they were exposed to either a manipulation of low SSES (or control condition) prior to evaluating and consuming an isocaloric chocolate milkshake. Compared to the control condition, participants in the low SSES condition exhibited increased circulation of the appetite-stimulating hormone ghrelin while evaluating the milkshake. While there was no effect of the SSES manipulation on hormones associated with satiety (pancreatic polypeptide and insulin), those in the SSES condition reported lower fullness approximately 40 minutes following milkshake consumption. Although not directly examining satiation/satiety, a pilot feeding intervention by Wijayatunga and colleagues (2019) provides indirect support that low subjective status may dysregulate those processes. Females (aged 20 or older) reporting lower SSES engaged in greater energy intake for the rest of the day following consumption of a large lunch consisting of 60% of their estimated daily energy requirement. Although the authors did not directly assess satiation or satiety, they speculated that this observed relationship between lower SSES and poorer compensation for large meals could be due to reduced sensations of satiety.

1.4. Current study

There have been limited direct tests of the relationship between low subjective status and satiation or satiety. Sim and colleagues (2018) found that low SSES was associated with reduced sensations of fullness after energy intake. However, due to the short interval between intake and re-assessment of hunger (40 minutes), it is not clear whether this may have been due to the potential effect of SSES on insufficient satiation or due to a gradual return of hunger during the interval. Disruptions to satiety, such as low satiety responsiveness, in childhood are associated with uncontrolled eating behavior and higher body mass later in life (Derks et al., 2024; Kininmonth et al., 2021). Understanding whether subjective status affects satiation and satiety is an important yet overlooked question in the study of social and developmental determinants of obesogenic eating behaviors.

Using previously collected data to examine the cross-sectional relationship subjective status shares with satiation and satiety among youths, the current study tested two hypotheses. First, we predicted that lower subjective status (SSS and/or SSES) would be associated with lower reductions in appetite (satiation) following consumption of a standardized shake beverage (Hypothesis 1). We also conducted exploratory analyses examining whether lower subjective status (SSS and/or SSES) would be associated with lower reductions in cravings following consumption of the standardized shake. Although prior research has suggested an experimental manipulation of perceived deprivation compared to others produced increased desire to eat foods that were subsequently presented (Sim, Lim, Forde, et al., 2018), there has been limited research directly linking subjective status to cravings for specific foods. Second, we predicted that lower subjective status (SSS and/or SSES) would be associated with greater return of hunger (lower satiety) across 90 minutes following consumption of the shake (Hypothesis 2). These hypotheses were tested among a sample of children and adolescents in a controlled laboratory setting.

2. Methods

2.1. Participants

This study involved secondary analysis of data from the Children’s Growth and Behavior Study (CGBS) (clinical trials number: NCT02390765). CGBS started in 2015 to explore the potential effects of environment and genes on behavior and health longitudinally. Children and their caregivers were recruited from the Washington, D.C. metro area using direct mail, flyers, and previous study enrollment/ineligibility. Eligible participants were cognitively capable, between the ages of 8–17 years old at baseline, in general good health, and had a BMI ≥ 5^th percentile for sex and age. Child assent and caregiver consent were also attained during baseline. Assessments were collected annually for up to 6 years. Participants completed the measures in the current analyses at the baseline visit and again at the Year 3 follow-up visit. However, subjective status was only measured at one of these visits. Thus, each participant’s data used for this study was based on the interval at which measures of subjective status were collected (SS Interval: either baseline visit or Year 3 follow-up visit). All study procedures were approved by the National Institutes of Health Review Board. All hypotheses and analyses were preregistered on the Open Science Framework,¹ with the exception of the latent profile analyses and analyses examining craving as an outcome (osf.io/d5k27).

At the time the time of analysis, 369 participants were enrolled in the study. Participants were excluded from analyses if they were missing data for the subjective status measure (n = 178), covariates (n = 58), or outcomes (Figure 1). Many participants were missing data for subjective status because this measure was introduced into the study at a later point after data collection had already begun. Since most of the missing data is due to changes in study protocol, missingness is largely due to some participants not being asked to complete the subjective status measure. Therefore, missingness is unlikely to be related to participants’ actual subjective status values and complete case analysis was used since systematic bias due to missing data was unlikely. The final sample (N = 133), based on participants who had data available for predictors (SSS and SSES) and covariates, was 53.40% male, 69.90% White, and had a mean age of 13.67 (SD = 3.02) years. Those with missing data on the satiation and satiety related outcome measures were retained, resulting in different sample sizes based on the outcome being analyzed: satiation quotient (n = 110), craving quotient (n = 108), and hunger ratings over 90 minutes² (n = 123).

Figure 1.

Overview of the numbers of participants included for analyses after those with missing data were sequentially excluded. The row of boxes at the bottom depict the final analytic sample size for each of the outcomes. CGBS: Children’s Growth and Behavior Study.

2.2. Measures

Subjective status.

Subjective status was measured using the youth version of the MacArthur Scale of SSS (Goodman et al., 2001). This measure includes one item to assess SSS and one item to assess SSES. These refer to one’s perceived social standing on personal social status (e.g., respect) or family socioeconomic resources, respectively.

The SSES item was presented in the following manner: “Imagine that this ladder pictures how American society is set up. At the top of the ladder are the people who are the best off—they have the most money, the highest amount of schooling, and the jobs that bring the most respect. At the bottom are people who are the worst off—they have the least amount of money, little or no education, no jobs, or jobs that no one wants or respects. Now think about your family. Please tell us where you think you family would be on this ladder. Full in the circle that best represents where you family would be on this ladder.”

The SSS item was presented in the following manner: “Now assume that the ladder is a way of picturing your school. At the top of the ladder are the people in your school with the most respect, the highest grades, and the highest standing. At the bottom are the people who no one respects, no one wants to hang around with, and have the worst grades. Where would you place yourself on this ladder? Fill in the circle that best represents where you would be on this ladder.” For each item, participants selected the rung (1 to 10) that best represents their subjective status, with higher scores indicating higher subjective status. Responses were treated as continuous variables.

Objective socioeconomic status.

Caregivers’ report of their educational and occupational background was measured using the Hollingshead Two Factor Index of SES (Hollingshead, 1975). Education and occupation scores ranged from 1 (graduate school/executive positions) to 7 (> 7 years of education/service workers). The education score was weighted by three and the occupation score was weighted by five. The two were added together to create a composite objective socioeconomic status (OSES) score. If both caregivers (rather than a single caregiver) completed this measure, education and occupation scores were averaged before computing the composite OSES score. This score was reverse scored so that higher scores represented higher OSES and treated as a continuous variable.

Pre- and post-shake appetite and craving ratings.

The satiation quotient was used to determine the change in craving and appetite ratings per percent of calorie consumed of the breakfast shake that children received during the lab visit (Green et al., 1997). Before and after consuming a breakfast shake (see Procedures for description of the shake), participants rated items assessing appetite and craving using 100-point visual analog scales labeled with “not at all” and “extremely” at the anchors. The appetite items included the following prompts: “How hungry do you feel right now?”, “How much food do you think you could eat right now?”, and “How full do you feel right now?” (reverse scored). The craving items included, “How strong is your desire to eat one or more specific foods?”, “How much do you crave one or more specific foods?”, “How strongly do you want to eat one or more specific foods?”, and “How much does your desire or craving to eat have power over you?”. Composite scores for appetite ratings were created by averaging pre-shake (α = 0.58) and post-shake ratings (α = 0.80), before computing a satiation quotient by subtracting mean post-shake from mean pre-shake ratings and dividing the difference by the percentage of the total shake (in kcal) the participant consumed. Likewise, we created composite scores for craving ratings by averaging pre-shake (α = 0.88) and post-shake ratings (α = 0.86), before computing a craving quotient by subtracting mean post-shake from mean pre-shake ratings and dividing by the percentage of the total shake (in kcal) the participant consumed (Fillon et al., 2021).

Hunger ratings across 90 minutes.

Participants rated their hunger on a 100-point visual analog scale across four timepoints: 0, 30, 60, and 90 minutes following consumption of the breakfast shake, which served as a measure of satiety. The scale used anchors that ranged from “not at all” to “extremely.” The pattern of these hunger ratings over time were analyzed using two approaches: In the first, area under the curve (AUC) was computed for hunger ratings across time representing the total magnitude of hunger reported over the 90 minutes following shake consumption. In the second approach, latent profile analysis (LPA), a person-centered approach, was used to identify and categorize participants into groups based on sharing similar trajectories of hunger ratings over the 90 minutes (Leffondré et al., 2004). Hunger ratings were analyzed using two different approaches because AUC does not consider the pattern or shape of hunger trajectories (e.g., increasing and decreasing patterns of hunger responses over time can have the same AUC). For this reason, LPA was used to classify and analyze the sample by clusters based on their trajectories of return of hunger over time. Yet, LPA does not directly account for total magnitude of hunger experienced across the 90-minute period, thus AUC was also used in analysis.

Body mass index.

BMI (kg/m²) was measured using height and weight measurements in triplicate from a stadiometer with a calibrated scale. These measurements were then converted and standardized into z-scores using age and sex according to the CDC guidelines (Kuczmarski et al., 2000).

Demographics.

Participant race was assessed with response options of White, Black or African American, Asian, and multiple/other races. The options for sex were male and female.

2.3. Procedures

Participants arrived at the lab at 8:30 AM after completing an overnight fast in which they were instructed not to eat anything after 10:00 PM. At the start of the lab visit, fasting was confirmed by the research team. Participants completed the pre-shake appetite and craving ratings. Next, participants were provided a breakfast shake prepared by the lab at 10:00 AM. Each shake’s energy content was standardized to approximately 21% of each child’s estimated daily energy needs, determined by the child’s age, height, body weight, and average physical activity level reported from the previous week using the International Physical Activity Questionnaire (Craig et al., 2003), as estimated based on individual energy need equations provided by the Institute of Medicine of the National Academies (Meyers et al., 2006). Participants were able to select one of three shake flavors (vanilla, chocolate, or strawberry), and the shake included 17% protein, 67% carbohydrates, and 16% fat. The shakes were presented in an opaque disposable cup with a lid and straw so that participants could not visually detect how much of the cup was filled by the shake. Participants were instructed to try to consume the entire shake and that they had several minutes (approximately 5 minutes) to complete the shake. They were also told they were free to stop consuming the shake without finishing it if they desired or felt ill. Immediately after consuming the shake, participants completed the post-shake appetite and craving ratings. These pre- and post-shake ratings of appetite and craving were used to compute the satiation and craving quotients. Participants were also asked to complete a separate hunger rating after consuming the breakfast shake. The hunger ratings were administered in 30-minute intervals following shake consumption (0 minutes/immediately, 30 minutes, 60 minutes, and 90 minutes post-shake), and were used to examine return of hunger (satiety) over time after they were prepared for AUC and LPA analyses. During this time, participants completed other questionnaires and measures that were not included in the current analyses.

2.4. Analysis

Analyses were conducted using SPSS version 29 (IBM, 2022). Independent t-tests and chi-squared tests were used to check for presence of systematic differences in characteristics or responses of participants based on the study interval (baseline visit or Year 3) that participants’ data was used for analysis. Spearman correlations were conducted to identify associations between study variables. Covariates included the visit that subjective status was measured since it was collected at either baseline or Year 3. OSES³, standardized BMI Z-score, child age (8–20 years old), sex (male or female), and race (Black, White, Asian, multiracial) were also used as covariates across analyses. These covariates were selected since some of them could be associated with differences across children in their energy requirements, how rapidly hunger is experienced after eating, and desire to eat (e.g., BMI Z-score, sex, age) (Carnell & Wardle, 2008; Webber et al., 2009), while other covariates allowed us to further isolate the unique effects of subjective status on our outcomes independent of demographic factors associated with health disparities (e.g., objective SES, race) (Cheon et al., 2025). Dummy coding was applied to race with ‘White’ as the reference group.

Our hypotheses were tested using multiple linear regression. SSES or SSS were entered as a predictor of satiation quotient, craving quotient, and AUC outcomes in separate models with BMI Z-score, age, sex, race, and objective SES as covariates. Unstandardized coefficients (B) are reported. Wild bootstrapping was used to estimate confidence intervals for linear regression coefficients, which provides robust estimates even when assumptions of non-normality and heteroscedasticity are violated (Liu, 1988).

The LPA analysis of hunger ratings over the 90-minute period employed a two-stage framework. Stage 1 classified groups of children with different trajectories. Stage 2 analyzed the associations of the identified subgroups with SSES and SSS. Stage 1 first identified 24 different derived features of the trajectories, such as range, mean-over-time, standard deviation, etc., as proposed by Leffondré and colleagues (2004). A factor analysis was then carried out to reduce the features, and thus to select a subset of important and non-redundant features. Finally, K-means clustering was performed based on the selected subset of the features, once a chosen number of clusters was decided. These analyses were implemented in R software, version 4.3.2 using the R traj package (Sylvestre & Vatnik, 2023). In Stage 2, multinomial logistic regression was applied to investigate the association between SSES and SSS (in separate models) and likelihood of belonging in each group compared with the reference group while adjusting for covariates. Odds Ratios (OR) of group membership (vs. reference) was reported. During Stage 2, “Asian” and “Multiracial” responses for race were merged into a single category “Multiracial” for better representation of race across the groups of hunger trajectories.

3. Results

Descriptive statistics are presented in Table 1. Correlations between variables are presented in Table 2. Overall, SSES and SSS were significantly positively correlated and SSES was significantly positively correlated with OSES. There were no other associations between SSS and SSES with other variables, except SSS being positively correlated with satiation quotient (greater satiation). When comparing characteristics of participants whose data came from baseline or Year 3 visits, there were no significant differences on any study variables (see Table 1), except participants whose data came from the Year 3 visit were older (Mean = 14.69, SD = 2.81) than those whose data came from baseline (Mean = 12.13, SD = 2.67), t(131) = 5.23, p < 0.001, d = 0.93.

Table 1:

Descriptive statistics of study variables among eligible participants in the Children’s Growth and Behavior Study.

Continuous variables	N	Minimum, maximum	Mean (SD)
Age (years)	133	8.00, 20.00	13.67 (3.02)
Subjective Social Status	133	2.00, 10.00	7.08 (1.67)
Subjective Socioeconomic Status	133	3.00, 10.00	6.76 (1.45)
Objective Socioeconomic Status	133	37.00, 80.00	63.47 (9.73)
BMIz	133	−1.76, 2.88	0.39 (0.98)
Satiety: Area Under the Curve	123	0.00, 249.95	92.35 (58.91)
Satiation Quotient	110	−0.72, 1.62	0.26 (0.29)
Craving Quotient	108	−0.55, 1.24	0.17 (0.25)
Categorical variables	N	Categories	Frequency (percentage)
Sex	133	Male	71 (53.40%)
		Female	62 (46.60%)
Race	133	Asian	10 (7.50%)
		Black	17 (12.80%)
		Multiple Races	13 (9.80%)
		White	93 (69.90%)
SS interval	133	Baseline	53 (39.80%)
		Y3	80 (60.20%)
Hunger Trajectory Group (based on latent profile analysis)	123	Fast inc. hunger	21 (17.10%)
		Slow inc. hunger	47 (38.20%)
		Decreasing hunger	28 (22.80%)
		Low stable hunger	27 (22.00%)

Fast inc. hunger: fast-increasing hunger group; Slow inc. hunger: slow increasing hunger group; BMIz: body mass index z-score; SSS: subjective social status, SS interval: study visit where data for the study was collected.

Table 2.

Spearman correlation coefficients (with sample size) between variables in the study.

	SSES	SSS	Hunger AUC	Satiation Quotient	Craving Quotient	Age	BMIz	Objective SES
SSES	1 (133)	0.29** (133)	−0.12 (123)	0.05 (110)	−0.03 (108)	−0.018 (133)	−0.09 (133)	0.24** (133)
SSS	0.29** (133)	1 (133)	−0.17 (123)	0.20* (110)	−0.14 (108)	−0.03 (133)	−0.07 (133)	0.04 (133)
Hunger AUC	−0.12 (123)	−0.17 (123)	1 (123)	−0.50** (103)	0.18 (101)	−0.21* (123)	−0.002 (123)	0.07 (123)
Satiation Quotient	0.05 (110)	0.20* (110)	−0.50** (103)	1 (110)	0.20* (108)	0.14 (110)	−0.19* (110)	−0.08 (110)
Craving Quotient	−0.03 (108)	−0.14 (108)	0.18 (101)	0.20* (108)	1 (108)	−0.04 (108)	−0.04 (108)	0.02 (108)
Age	−0.02 (133)	−0.03 (133)	−0.21* (123)	0.14 (110)	−0.04 (108)	1 (133)	0.12 (133)	−0.05 (133)
BMIz	−0.09 (133)	−0.07 (133)	−0.002 (123)	−0.19* (110)	−0.04 (108)	0.12 (133)	1 (133)	−0.14 (133)
Objective SES	0.24** (133)	0.04 (133)	0.07 (123)	−0.08 (110)	0.02 (108)	−0.05 (133)	−0.14 (133)	1 (133)

SSES: subjective socioeconomic status; SSS: subjective social status, Hunger AUC: area under the curve for hunger ratings over 90 minutes after consuming the shake.

^*p < .05.

^**p < .01.

Satiation and craving quotients.

There was a significant positive association between SSS and the satiation quotient (B = 0.04, bootstrapped 95% CI: 0.0003, 0.08). Higher (lower) reported SSS was associated with greater (smaller) reductions in appetite per percent of the shake consumed. There was no relationship between SSS and craving quotient (B = −0.02, bootstrapped 95% CI: −0.06, 0.02). No significant relationship was observed between SSES and the satiation quotient (B = 0.02, bootstrapped 95% CI: −0.01, 0.06) or craving quotient (B = −0.01, bootstrapped 95% CI: −0.04, 0.02) (see Table 3 for full results of regression models).

Table 3.

Results of linear regressions with subjective status (SSS or SSES) as predictors of satiation (satiation quotient) and satiety (area under the curve for return of hunger during 90 minutes after shake consumption). SSS and SSES were used as predictors of outcomes in separate models, indicated by the “subjective status” row.

		B (Bootstrapped 95% CI)
Outcome	Variables	SSS	SSES
Appetite Quotient	Age	0.03 (0.01, 0.04)	0.03 (0.01, 0.04)
	Black	−0.04 (−0.28, 0.20)	−0.03 (−0.27, 0.20)
	Asian	−0.14 (−0.34, 0.05)	−0.13 (−0.31, 0.18)
	Multiple races	−0.10 (−0.19, −0.08)	−0.10 (−0.23, −0.02)
	Sex	−0.01 (−0.13, 0.10)	−0.004 (−0.12, 0.13)
	OSES	−0.004 (−0.01, −0.003)	−0.004 (−0.01, −0.004)
	SS interval	−.15 (−0.28, −0.02)	−0.15 (−0.29, −0.01)
	BMIz	−0.06 (−0.12, 0.01)	−0.07 (−0.13, 0.004)
	Subjective status	0.04 (0.0003, 0.08)	0.02 (−0.01, 0.06)
Craving Quotient	Age	0.003 (−0.01, 0.02)	0.003 (−0.01, 0.02)
	Black	0.03 (0.18, 0.16)	0.02 (−0.17, 0.13)
	Asian	−0.04 (−0.15, 0.12)	−0.05 (−0.16, 0.10)
	Multiple races	−0.03 (−0.19, 0.12)	−0.03 (−0.20, 0.12)
	Sex	0.04 (0.06, 0.16)	0.035 (−0.08, 0.16)
	OSES	0.0004 (−0.01, 0.003)	0.0004 (−0.01, 0.004)
	SS interval	−0.09 (−0.21, 0.03)	−0.09 (−0.22, 0.04)
	BMIz	−0.02 (−0.08, 0.05)	−0.01 (−0.07, 0.05)
	Subjective status	−0.02 (−0.06, 0.02)	−0.01 (−0.04, 0.02)
Area Under the Curve	Age	−5.10 (−9.64, 0.08)	−4.72 (−9.28, 0.75)
	Black	32.36 (2.04, 23.05)	31.71 (1.50, 20.85)
	Asian	40.14 (4.70, 43.83)	39.24 (0.40, 44.28)
	Multiple races	41.97 (−1.92, 99.32)	37.60 (−3.15, 90.38)
	Sex	2.19 (−21.80, 30.18)	0.55 (−22.98, 29.47)
	OSES	0.77 (−0.35, 2.15)	0.91 (−0.20, 2.45)
	SS interval	4.74 (−23.14, 38.00)	4.02 (−24.25, 37.23)
	BMIz	−2.71 (−11.28, 5.42)	−1.94 (−10.26, 6.20)
	Subjective status	−8.06 (−12.94, −4.32)	−6.57 (−12.35, −1.52)

Covariates in these analyses were age, sex, race (dummy coded using White as the reference group for comparisons with Black, Asian, and Multiple Races), BMI Z-scores, objective SES, and the visit at which the data for the current analysis was collected (SS interval). B represents unstandardized regression coefficients. SSS: Subjective social status; SSES: Subjective social status. Sex is coded as 0 = male and 1 = female. SS Interval is coded as 0 = baseline visit and 1 = year 3 visit. Bolding indicates statistically significant association (95% CI does not include zero).

Area Under the Curve.

There was a significant negative association between SSS and AUC (B = −8.06, bootstrapped 95% CI: −12.94, −4.32). Higher (lower) SSS was associated with lower (greater) overall magnitude of hunger across the 90-minute period following shake consumption. However, unlike the satiation quotient, a significant negative association was also found when investigating the relationship between SSES and AUC (B = −6.57, bootstrapped 95% CI: −12.35, −1.52). Higher (lower) SSES was associated with lower (greater) overall magnitude of hunger across the 90-minute period (Table 3).

Latent Profile Analysis.

LPA was used to identify different groups within the sample who exhibited similar trajectories or patterns in their hunger ratings over the 90 minutes after shake consumption. Out of 24 derived features of hunger trajectories, a factor analysis, followed by K-means clustering, was used to establish four primary profiles or groups. Factor analysis first reduces 24 features to four features, retaining principal components with eigen values larger than 1. Next, K-means clustering was performed based on these four features, that originally identified five subgroups using scree plot. When reviewing the interpretability of these subgroups, there were two separate subgroups representing slowly increasing hunger trajectories (n = 15 and n = 32). These subgroups were merged into one factor for better interpretability and reduced redundancy. The process produced four primary groups which correspond to principal components that explained 87% of the variation of the 24 derived features of the hunger trajectories across participants. The first profile group (Figure 2A) exhibited the fastest return of hunger across the four time points and reported feeling the hungriest after consuming the breakfast shake. Group 1 consisted of approximately 17% of the total sample (n = 21). The second group (Figure 2B) was the largest and their hunger increased slowly across all four time points. Group 2 consisted of approximately 38% of the total sample (n = 47). The third group (Figure 2C) had a hunger trajectory that decreased across the four time points and consisted of approximately 22% of the total sample (n = 28). Lastly, the fourth group (Figure 2D) reported relatively low hunger across all time points with only a slight increase. Group 4 consisted of approximately 22% of the total sample (n = 27).

Figure 2.

Trajectories of hunger across 90 minutes following shake consumption derived from latent profile analysis. Solid black lines represent the median. Dotted lines represent the upper and lower bounds of the 95% confidence interval.

Separate multinomial logistic regression models were used to analyze the LPA and investigate whether there was a significant relationship between SSS or SSES and odds of being in one of the latent profile groups. Group 4 (low stable hunger) was used as the reference group in the analysis because this group exhibited a pattern that represents consistent satiety across time, to which other groups that exhibit patterns of returning or diminishing hunger could be compared. There was no significant relationship between SSS and likelihood of falling into any of the four groups. There was, however, a significant relationship between SSES and likelihood of being in Groups 2 (slow increasing hunger) and 3 (decreasing hunger), such that participants with higher SSES had lower odds of being in the slow increasing (group 2), OR = 0.61, 95% CI: 0.42, 0.90, and decreasing hunger group (Group 3), OR = 0.63, 95% CI: 0.41, 0.96 (Table 4).

Table 4.

Odds ratios and 95% confidence intervals (in parentheses) from multinomial logistic regression analyses of the associations between subjective status (SSS or SSES) and trajectories of hunger across 90 minutes following shake consumption. The low stable hunger group was the reference for comparison. SSS and SSES were used as predictors of outcomes in separate models, indicated by the “subjective status” column.

	Group	Subjective status	AGE	OSES	Black	Multiple races	Male	SSS interval
	Fast increasing	0.71	0.80	1.00	1.74	5.82	4.25	0.25
	hunger (1)	(0.45, 1.15)	(0.63, 1.02)	(0.93, 1.07)	(0.22, 13.71)	(0.80, 42.19)	(1.06, 17.04)	(0.06, 1.13)
SSES	Slow increasing	0.61	0.89	1.03	1.41	7.37	2.42	0.76
	hunger (2)	(0.42, 0.90)	(0.73, 1.08)	(0.97, 1.09)	(0.28, 6.95)	(1.36, 39.92)	(0.82, 7.13)	(0.21, 2.66)
	Decreasing	0.63	0.81	1.05	1.37	1.26	0.77	1.20
	hunger (3)	(0.41, 0.96)	(0.65, 1.01)	(0.98, 1.12)	(0.25, 7.38)	(0.15, 10.64)	(0.24, 2.48)	(0.31, 4.65)
	Low stable hunger (4)	ref	ref	ref	ref	ref	ref	ref
	Fast increasing	1.04	0.79	0.98	1.60	4.86	3.87	0.25
	hunger (1)	(0.70, 1.54)	(0.62, 1.00)	(0.91, 1.04)	(0.21, 12.51)	(0.68, 34.78)	(0.99, 15.2)	(0.06, 1.10)
SSS	Slow increasing	0.79	0.88	1.01	1.34	7.17	2.20	0.80
	hunger (2)	(0.57, 1.09)	(0.72, 1.06)	(0.96, 1.06)	(0.28, 6.33)	(1.36, 37.70)	(0.76, 6.35)	(0.23, 2.78)
	Decreasing	0.88	0.81	1.03	1.30	1.17	0.72	1.24
	hunger (3)	(0.62, 1.26)	(0.65, 1.00)	(0.97, 1.00)	(0.25, 6.72)	(0.14, 9.63)	(0.23, 2.26)	(0.33, 4.71)
	Low stable hunger (4)	ref	ref	ref	ref	ref	ref	ref

‘White’ is the reference group for race, and ‘female’ is the reference group for sex. ‘Asian’ was combined with the ‘multiple races’ group since these groups were not adequately distributed across the four hunger trajectories. SSS Interval is the visit at which the data for the current analyses was collected. SSS: Subjective social status; SSES: Subjective social status. Bolding indicates statistically significant association (95% CI does not include 1)

4. Discussion

In this study we tested whether lower subjective status (SSS and/or SSES) was associated with reduced sensations of satiation and satiety among children and adolescents. Overall, we observed support for both of our hypotheses. Hypothesis 1 was partially supported, in that lower SSS, but not SSES, was associated with reduced satiation immediately following shake consumption, although the effect was modest. Hypothesis 2 was supported, in that both lower SSS and SSES were associated with reduced satiety, as greater magnitude of hunger was experienced over 90 minutes following shake consumption. The relationship between subjective status and these outcomes were observed independently of household OSES. Notably, additional analyses examining trajectories of hunger ratings provide further support for Hypothesis 2, by demonstrating that participants reporting lower SSES were more likely to fall into the slow increasing hunger and decreasing hunger groups compared to the low stable hunger group. One potential reason for this might be because the slow increasing and decreasing hunger groups reported the highest level of hunger initially after consuming the shake (Figure 2). Thus, the relationship between low SSES and exhibiting these two trajectories of hunger may be due to a relationship between low SSES and higher ratings of hunger at the first measurement after the shake. This could explain why there was no relationship between SSES and placement in the fast-increasing hunger group relative to the low stable hunger reference group, as one similarity between both groups was low ratings of hunger shortly after shake consumption.

We observed more consistent relationships of SSS with satiation and satiety compared to SSES’s relationship with these outcomes in that SSS was associated with both outcomes, whereas SSES was only associated with satiety. This may be due to SSS being a more meaningful representation of subjective status among children and adolescents than SSES. These findings are consistent with prior work that indicates that SSS may be more associated with BMI than SSES among youths (Goodman et al., 2001). Similarly, a recent study that examined the interactive effects of distress from teasing and subjective status found that among children reporting distress from teasing, lower SSS (but not SSES) was associated with higher BMI, adiposity, and EAH due to negative affect (Cheon et al., 2024).

We did not observe any relationships between subjective status and changes in craving following consumption of the breakfast shake. The craving questions referred to a desire to consume “one or more specific foods,” but did not depict or describe any foods as a reference. Constructs related to subjective status, such as personal relative deprivation, have been associated with increased desire to eat foods that are presented as stimuli (Sim, Lim, Forde, et al., 2018). One possibility is that low subjective status may heighten craving or desire to eat in a reactive manner, such as when one encounters or is reminded of a palatable food.

Our findings of the relationship between lower subjective status and reduced satiation and satiety have important implications. Some obesogenic eating behaviors could be expressions of reduced sensations of satiation or satiety. These findings suggest that dysregulation of satiation or satiety may be a possible explanation for broader relationships between low subjective status and eating behaviors that may lead to the excess energy intake observed in prior studies. This work supports the concept that perceived deprivation of social and socioeconomic resources may coactivate or contribute to sensations of hunger and desire for energy intake (Briers & Laporte, 2013; Cheon & Hong, 2017). These observations may also help to explain why exposures to cues of environmental harshness, scarcity, or food insecurity may also increase energy intake and desire to consume energy-dense foods (Dhurandhar, 2016; Laran & Salerno, 2013; Swaffield & Roberts, 2015). Low subjective status may be a component of diverse social determinants of health, such as low income, poor prosects of social mobility, discrimination, stigma, and social exclusion. Many of these social determinants have been associated with risk of overeating among children, adolescents, and young adults (Kininmonth et al., 2020; Pink et al., 2024; Raney et al., 2023; Sim, Lim, Leow, et al., 2018; Vartanian & Porter, 2016), Our findings suggest that disruptions to satiation or satiety may play a role. These findings also have notable clinical implications. Prevention and treatment of eating behaviors that risk excess energy intake and obesity among youths, such as loss of control eating or EAH, could signal a need to screen for experiences that may threaten subjective status and social relationships like teasing, bullying, or social isolation. Further research is required on whether subjective status can be reliably changed through interventions. Recent research has suggested that subjective status may mediate the relationship between adolescent to adulthood socioeconomic mobility and dietary habits and metabolic health in adulthood (Bittner et al., 2024). Increasing one’s subjective status through societal and structural changes that promote upward socioeconomic mobility among youths could also support adoption of healthier eating behaviors in the transition to adulthood, in addition to reducing health disparities in general.

There are some limitations to this study. The sample size was relatively small and participants were recruited from a relatively affluent region of the United States. Thus, the findings may be less generalizable to other regions or internationally. Although energy content of the shakes was standardized to participants’ energy needs, we may have observed different effects if the energy was consumed from solid food rather than a beverage. Our sample consisted of a wide range of ages, from 8 years to 20 years old. Trajectories of SSES may vary across child age, with typically higher SSES being reported during childhood, which declines with age throughout adolescence (Amir et al., 2019; Mandalaywala et al., 2020). Although we controlled for age, we did not have sufficient sample size to conduct reliable subgroup analyses. Another limitation was that objective SES was only assessed among one caregiver (usually mothers) for most of our sample. Although people with similar socioeconomic status are more likely to marry one another and have children (Putnam, 2016), our measure of objective SES may not have been representative of overall household socioeconomic circumstances. Finally, the relationships between subjective status and satiation/satiety were examined cross-sectionally, and this study was a secondary data analysis. Future research that experimentally manipulates subjective status could potentially demonstrate causal effects of low subjective status on satiation and satiety.

Our study also has notable strengths. It is the first to directly examine subjective status’s relationships with satiation and satiety, doing so using standardized lab-based measures. These relationships were also examined in a sample that included children, which is lacking in research on subjective status and food-related responses. In addition to measuring both satiation and satiety as independent outcomes in the study, we analyzed hunger ratings over the 90-minute post-shake period with two different approaches: AUC and classification of distinct trajectories with LPA. Both approaches revealed converging findings, with subjective status (either SSS or SSES) associated with patterns of satiety generated by these approaches.

In conclusion, our findings suggest that low subjective status is associated with reduced levels of satiation and satiety among children and adolescents, independent of household OSES, which showed weaker and inconsistent relationships with these outcomes. These findings contribute to a growing body of literature on the relationship subjective status shares with appetite regulation and eating behaviors, while providing novel insights into the role that satiation and satiety may have in these relationships. An important avenue for future research would be to search for factors protective against reduced satiation and satiety among youths reporting lower subjective status.

Data availability

Data described in this manuscript will be made available upon request pending approval of the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development.