Change and Stability of Emotional Health of Rural Pennsylvania Youth During High School

Abstract

Purpose

Few prior studies have examined change in emotional health of high school students in a rural context. Considering the multifaceted nature of emotional health, this research aims to identify the patterns and explore change and stability of the emotional health of rural Pennsylvania youth. It also investigates the influence of family, peers, school, and the community environment on rural adolescents’ emotional health.

Methods

Using panel data from the Rural Youth Education Project (RYE), we employed latent transition analysis (LTA) to examine changes in patterns of rural students’ self-reported emotional health from 9^th grade to 11^th grade (N=1,294).

Findings

Four distinct emotional health sub-groups for rural adolescents were identified. Over half of the youth in the sample felt emotionally well, or positive, in both 9^th and 11^th grades. Roughly 60% of rural youth remained in the same emotional health category from 9^th to 11^th grade, but a substantial minority experienced change in emotional health. One-fifth reported lower emotional health status in 11^th grade, and one-fifth indicated more positive emotions in 11^th than in 9^th grade. We found strong evidence of family, school, community, and peer influences on the emotional health of rural youth in 9^th grade.

Conclusions

The results suggest that while a large share of rural youth exhibit positive emotional health and remain positive across their high school years, a substantial share of rural youth experience changing emotional health. The study underscores the important role that family, peers, school, and the community environment play for rural youth’s emotional health over time.

Adolescence is a time of development and change in many spheres of life. The emotional health of youth has increasingly drawn scholarly attention, since it is related with many critical outcomes in other life domains. Literature in psychiatry and psychology has documented that negative emotions, such as anxiety, depression or distress, are related to adverse health outcomes, poor academic performance, and behavior problems, including substance use and risky sexual behaviors.^1–3 Studies on emotional well-being find that positive emotions, such as happiness, interest in life, and competence, are strong predictors of many desirable mental and physical outcomes,^4,5 as well as positive thinking and initiative in work and school environments.⁶

Though there has been increased attention to the emotional health status of youth, few studies explore these indicators in rural adolescents. Youth in rural areas can have different family, peer, and community contexts than youth in suburban or urban areas, and different factors and experiences may influence their emotional health throughout their adolescent years. On one hand, certain characteristics of rural areas, such as overlapping social networks within family and community, are often seen as protective factors to promote positive youth development.^7,8 But fewer and less diverse job opportunities in rural areas create disadvantages for rural youth. As youth in rural areas make the transition to adulthood, fewer opportunities for jobs and higher education in rural areas make it more likely that rural youth will need to move to other areas to achieve their educational and career goals. This impending change can leave youth with feelings of uncertainty, anxiety, and/or excitement.⁹ Moreover, issues related to geographic isolation and poverty, such as transportation difficulties and limited access to physical and mental health services, may lead to lack of timely intervention when youth experience emotional problems and have access to few or no programs promoting emotional well-being.^9,10 Understanding the patterns and changes in emotional health of rural adolescents during high school can assist counselors, teachers, parents, and health practitioners to identify potential problems and to help youth to seek treatment. This knowledge also would assist practitioners to develop programs to help youth and their families to understand the issues rural youth face and to help them in resolving likely conflicts and concerns that affect rural adolescents’ emotional well-being.

In addition to a lack of focus on rural contexts, another issue in the emotional health literature is the conceptualization of emotional health status. Traditional research in psychopathology often explicitly or implicitly considers positive emotional well-being (such as happiness, joyfulness, and hopefulness) and negative emotions (such as anger, distress, and anxiety) to be opposite poles of a single emotional continuum. The absence of negative emotions suggests the presence of emotional well-being.¹¹ However, increased evidence suggests that the presence of emotional well-being does not mean the absence of illness, depression, distress, or anxiety.^12–15 Instead, emotional health should be regarded as a multifaceted construct with various dimensions, including a balance of positive and negative affect,¹⁶ pleasantness and activation,¹⁷ and tense and energetic arousal.¹⁸ To capture a comprehensive picture of a youth’s emotional health status, one needs to incorporate both positive and negative measurement of emotions.

Moreover, change in youths’ emotional health profiles over time does not follow a unified pathway.¹⁹ Youth experience physical and psychological changes as well as changing family, community, and social contexts, which may further affect mental and emotional health status over time. Therefore, it is possible that the prevalence of mental or emotional health statuses of youth would remain stable at the population level, while changes occur at the individual level.²⁰ Examining change in an individual’s mental or emotional health status over time requires panel data and longitudinal analysis.

On the empirical side, many factors related to different emotional health profiles have been identified. Overall, emotional well-being is closely related with social class and social relations. Economic hardship is consistently found to be related with poor emotional health profiles.²¹ Relative social status, and subjective measures of socioeconomic status (SES), have been associated with adolescent mental health, even when controlling for other measures of SES.²² Parent-child relationships can also affect adolescent emotional health, with conflict leading to more depressive symptoms and lower self-esteem.²³ Besides SES, there is consistent evidence of gender differences in emotional health. Males and females differ in their emotional health profiles, with females more likely to suffer depression and anxiety, and males to experience anti-social disorders.^24,25

Additionally, contexts of the school and community can influence adolescents’ emotional health. Community context, such as local social capital and socioeconomic status of residents, has been associated with mental and emotional health in children and adolescents.^26,27 The importance of schools as sites of adolescent development has been well documented.²⁸ In addition to school and family, peers also heavily influence adolescents. For example, smoking and drinking are common risk behaviors associated with adolescent depression symptoms. Individuals not suffering from depression increase their risk of developing symptoms of depression if they engage in risky behaviors such as smoking and drinking.²⁹ The risk of engaging in smoking and drinking is particularly likely to be triggered by peer conformity during the transition to adulthood.^30,31

Research in current youth emotional health literature that attempts to identify different emotional profiles of sub-groups of rural youth is limited. Few studies jointly consider both positive and negative emotions and change in these emotions over time. Such studies are particularly lacking in the rural context. Tracking a sample of rural Pennsylvania youth from 9^th to 11^th grades, this study extends prior literature by addressing the following questions: (1) How do different indicators of emotion combine to reflect rural youths’ emotional health status profiles? (2) How do youths’ emotional health statuses change over time? And, (3) how do family, school, community, and peer-related factors affect the emotional health of individual youth?

Data and Analytical Approach

Study Sites and Data

This paper uses data from the Rural Youth Education project (RYE), a study conducted to better understand the educational, occupational, and residential aspirations of youth growing up in rural Pennsylvania. Rural Pennsylvania is an appropriate site to study rural youth because the state has a large rural population, the rural areas of the state tend to differ somewhat in their economic base—from reliance on farming, to forestry, to mining and energy extraction, to being bedroom communities for the state’s large cities.^32,33 In some ways, the state’s rural population reflects variation found in rural areas across the United States. Thus, the results may be relevant in helping educators, counselors, community leaders, and parents across rural areas of the US to understand issues facing rural adolescents as they make decisions about their futures. While urban and suburban youth face the same basic issues as rural youth related to making education and career decisions, their experiences with whether they are able to stay in their home community and achieve their goals and whether their families expect them to stay in the area in which they grew up may be quite different from the expectations facing rural youth.

The RYE survey asks students about their future plans, current feelings about school and community activities, friends, and family dynamics. The longitudinal study followed 2 cohorts of students beginning in 2005, when the students were in 7^th grade (the younger cohort) or 11^th grade (the older cohort), to 2011, when students were 1 year or 5 years past high school graduation. Participants were surveyed biennially. Using purposive sampling, 10 rural school districts were selected to represent rural Pennsylvania school districts. We used data from the younger cohort, when participants were in 9^th grade (2007) and 11^th grade (2009), N=1,294. A detailed description of the RYE data can be found in McLaughlin and associates.³⁴

Measurement of Emotional Health

The RYE surveys ask rural youth 7 questions regarding their feelings and attitudes in the past month. Of the 7 responses, 3 statements measure positive emotions and 4 measure negative emotions. The positively worded items include “Happy,” “Calm,” and “Confident of what I was doing.” The 4 negatively worded responses include “Very nervous or anxious,” “Overwhelmed,” “Downhearted and blue,” and “So down in the dumps that nothing can cheer me up.” Rural students’ responses to these statements range from 1= “None of the time,” 2= “Some of the time,” 3= “Most of the time,” and 4= “All of the time.”

We collapsed the responses into 2 categories for clarity of interpretation. The first category includes the responses that are either “None of the time” or “Some of the time,” capturing those who had no or limited reports of experiencing these emotions. The second category includes responses that are either “Most of the time” or “All of the time” to represent those who experienced an emotion often.

Family, School, Community and Peer Influences

We examined family, school, community, and peer influences on rural youths’ emotional health. The measures are presented in Table 1. Family environment is captured by 3 covariates. Rural youth reported their experience of negative family events by answering the survey question “Over the past year, have you ever experienced these events?” A list of items, such as parents had divorced, parents fought a lot, a parent lost a job, or a close relative had died, were given as yes or no options. Subjective family means is a variable included to capture youths’ perceptions of their family’s income levels compared to others. The variable is coded 1 if youth agree or strongly agree with the statement “My family can afford to buy the things that other families can buy,” and coded zero otherwise. Argue with parents is coded as 1 if youth report “yes” to the survey question “Have you had a serious argument with your mother/father about your behavior in the past month?” and coded zero otherwise. Two covariates are included to capture youths’ feeling towards community and school. Like community is coded as 1 if a youth reports that he or she likes the community where he or she is currently living “a lot,” and coded zero otherwise. Similarly, Like school was coded 1 if youth reported liking school a lot, and coded zero otherwise. We also include 2 additional covariates to measure rural youths’ exposure to negative peer influences. The RYE survey asks youth, “Of your closest friends, how many have ever smoked?” and “Of your closest friends, how many get drunk at least once a month?” The responses to these 2 questions ranged from no friends engage in these behaviors to up to 4 friends. We created 2 binary variables: Friends get drunk is coded 1 if youth indicate that 3 or 4 of their closest friends get drunk at least once a month, and Friends smoke is coded 1 if youth report that 3 or 4 of their closest friends have ever smoked, and coded zero otherwise. The descriptive statistics are presented in Table 1.

Table 1.

Descriptive Statistics for Variables Used in Latent Transition Analysis, N=1294

	9th grade	11th grade
Survey items of emotional health
Percentage of Responses for “Most” or “All of the time”
“Very nervous or anxious”	33.2	34.5
“Overwhelmed”	30.5	46.9
“Downhearted and blue”	29.7	32.1
“So down in the dumps that nothing can cheer me up”	23.1	26.6
“Happy”	80.1	76.7
“Calm”	63.7	65.3
“Confident of what I was doing”	72.6	72.0
Covariates (percentage)
Female	50.5	50.5
Subjective family means	75.7	73.6
Serious argument with parent(s)	46.4	44.4
Like community “a lot”	27.0	24.0
Like school “a lot”	63.9	58.3
Friends drink (3 or 4 closest friends get drunk at least once a month)	10.4	20.9
Friends smoke (3 or 4 closest friends smoke)	8.9	17.0

Analysis Strategy

To identify different mental health patterns of rural youth, we applied latent transition analysis (LTA), an extension of latent class analysis (LCA). LTA is a statistical model that explains the correlations among observed survey responses by making assumptions about hidden, “latent,” statuses. Participants who have similar behaviors are clustered into different subgroups based on their responses to survey questions. LTA also includes a time dimension, enabling researchers to model individual’s transitions from one latent class to another over time.³⁵ The method has been widely applied in social and behavioral science research, including studies of substance use,³⁵ college drinking,³⁶ aggressive behavior,³⁷ and smoking cessation.³⁸ A detailed discussion of the parameter estimation process can be found in Collins and Lanza.³⁹

The analysis was carried out in 3 steps. First, we conducted separate LCAs for participants in 9^th and 11^th grade to explore the optimal number of latent classes within each grade. To do this, we fit different measurement models with different numbers of latent classes for each time point to explore the latent class structure within each time point. We obtained fit indexes for models with 2 to 7 classes at each grade. We then repeated the same procedure for an LTA using the entire sample. The optimal number of latent statuses was chosen based on the joint considerations of model fit statistics, interpretability and parsimony, as well as consistencies between cross-sectional LCA and full sample LTA results. After identifying the optimal number of latent statuses, we interpreted 3 sets of parameters (item-response probabilities, latent status prevalences, and transition probabilities) in order to present the complete picture of transitions in rural adolescents’ emotional health statuses from 9^th grade to 11^th grade. The last step of our analysis involved incorporating family-, school-, and community-related covariates to predict latent status membership when youth were in 9^th grade. The latent statuses were predicted using the maximum likelihood method. LTAs were conducted using both Mplus 8 (Mplus, Los Angeles, California)⁴⁰ and SAS Proc LTA (SAS Institute Inc., Cary, North Carolina).⁴¹ Results across platforms were quite similar. We chose to present the Mplus results due to the ease of obtaining confidence intervals and significance tests for covariates. The final LTA estimates were adjusted for sampling clusters at the school district level.

Results

Identification of Common Patterns of Emotional Health Status

In order to determine the optimal number of latent statuses, we compared model fit statistics for a series of cross-sectional LCA and full sample LTA models, ranging from 2 latent statuses to 7 latent statuses. The model fit statistics are presented in Table 2, which includes AIC (Akaike Information Criterion), BIC (Bayesian Information Criterion), and Entropy. Based on the fit information of the LCAs and LTAs, we narrowed our model choice to either 4-latent statuses or 5-latent statuses, since both models yield relatively small AIC and BIC. A further inspection of the 5-status model reveals that only 3% of youth in wave 2 and 5% of youth in wave 3 were categorized as the newly added fifth latent status. In addition, 3 item-response probabilities reached their boundaries (ie, the probabilities are either 1 or 0) in the 5-status model, which is regarded as an indication of the extraction of too many latent statuses.⁴² Moreover, boundary estimates could result in the problem of convergence to a local likelihood maximum.⁴³ These findings make the 5-latent status model less appealing for both statistical and substantive interpretation purposes. Therefore, we adopted the 4-latent status model as our final model.

Table 2.

Summary of Information for Selecting Numbers of Rural Youth’s Emotional Health Statuses

LCA for 9th grade
# of classes	AIC	BIC	Entropy
2	7486.02	7560.99	0.69
3	7339.32	7454.27	0.74
4	7280.51	7435.44	0.74
5	7260.09	7455.00	0.73
6	7261.16	7496.05	0.75
7	7266.73	7541.60	0.77
LCA for 11th grade
2	6845.09	6919.02	0.77
3	6643.08	6756.44	0.78
4	6572.51	6725.29	0.79
5	6540.52	6732.73	0.80
6	6536.04	6767.69	0.76
7	6539.26	6810.33	0.75
LTA for full sample
2	14246.74	14406.66	0.64
3	13895.37	14153.29	0.66
4	13747.23	14113.48	0.68
5	13681.80	14166.70	0.71
6	13667.83	14281.69	0.72
7	13664.33	14417.48	0.76

Note: AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion

LTA Results

Panel A in Table 3 presents the item-response probabilities of the 4-latent status model in 9^th grade. The response patterns are shown in Figure 1. Four distinct latent statuses emerged to express the self-reported emotional health of rural youth in this study: positive, moderate, negative, and non-expressive. These labels express the general affect of the responses associated with each latent status, rather than a psychological diagnosis.

Table 3.

Item-Response Probabilities and Latent Status Prevalences for the Selected LTA Model

Latent Status
	Positive	Moderate	Non-expressive	Negative
Panel A: Item-Response Probabilities
Response probabilities “Most” or “All of the time”
“Very nervous or anxious”	0.03	0.59	0.15	0.64
“Downhearted and blue”	0.02	0.28	0.11	0.96
“So down in the dumps that nothing can cheer me up”	0.00	0.17	0.03	0.64
“Overwhelmed”	0.20	0.71	0.23	0.69
“Happy”	0.97	0.93	0.34	0.12
“Calm”	0.81	0.48	0.26	0.16
“Confident of what I was doing”	0.89	0.75	0.26	0.24
Panel B: Latent Status Prevalences
Proportion of status at
9th grade	0.57	0.15	0.20	0.08
11th grade	0.56	0.11	0.22	0.11

Figure 1.

Graphic Illustration of Item-response Probabilities of Emotional Health Latent Statuses

Youth who are positive in 9^th grade have the highest probabilities to often experience feeling “Happy” (prob=0.97), “Calm” (prob=0.81), and “Confident of what I was doing” (prob=0.89); they have the lowest probabilities of feeling “Very nervous or anxious” (prob=0.03), “Downhearted and blue” (prob=0.018), “So down in the dumps that nothing can cheer me up” (prob=0.003), and “Overwhelmed” (prob=0.20). Those who are negative in 9^th grade have high probabilities of reporting often experiencing negative feelings such as “Very nervous or anxious” (prob=0.64), “Downhearted and blue” (prob=0.96), “So down in the dumps that nothing can cheer me up” (prob=0.64), and “Overwhelmed” (prob=0.69). They have lower probabilities of experiencing positive emotions such as “Happy” (prob=0.12), “Calm” (prob= 0.16), and “Confident of what I was doing” (prob=0.24). Moderate rural youth have high probabilities to often experience positive feelings of “Happy” (prob=0.93), “Calm” (prob=0.48), and “Confident of what I was doing” (prob=0.75). However, these youth also have relatively high probabilities of reporting feeling “Very nervous or anxious” (prob=0.59) and “Overwhelmed” (prob=0.71). Non-expressive rural youth have low probabilities (prob< 0.35) of endorsing any of the feelings.ⁱ

Latent Status Prevalences and Transition Probabilities From 9^th to 11^th Grade

Panel B in Table 3 shows that the most prevalent status in 9^th and 11^th grades is positive, which includes half of the total sample in both grades. This was followed, among 9^th graders, by moderate and non-expressive, both of which included roughly 21% of the sampled youth. Moderate declined to 14.8% of youth by 11^th grade, while non-expressive increased to 23.1% of youth in 11^th grade. The negative emotional status was the least prevalent when youth were in the 9^th grade, and it increased slightly to 11.7% of 11^th grade youth. This relative stability of overall patterns across the 2 grades may mask individual respondent’s transitions across emotional health categories from 9^th to 11^th grade.

The transition probabilities of individuals from 9^th to 11^th grades are shown in Table 4. The row indicates rural youths’ latent emotional status in 9th grade, and the column identifies the status in 11th grade. Youth in the positive status were the most stable. Youth who were moderate, non-expressive, or negative in the 9^th grade have relatively modest probabilities of being in the same emotional status when they reach 11^th grade, indicating that substantial shifting occurred. While it is difficult to know what youth in the non-expressive category are feeling, these 9^th grade youth had the highest chance (prob=0.35) to move to the positive category by 11^th grade. Youth who were in the negative and moderate statuses in 9^th grade had probabilities of 0.15 and 0.16, respectively, to transition into the positive status by 11^th grade. Those who were negative in 9^th grade had a 0.20 probability of being moderate by 11^th grade and a 0.22 probability of becoming non-expressive.

Table 4.

Latent Status Transition Probabilities of the Selected LTA Model

Transition from 9th grade (rows) to 11th grade (columns)
		11th Grade Emotional Health
		Positive	Moderate	Non-Expressive	Negative
9th Grade Emotional Health	Positive	0.77	0.03	0.15	0.05
	Moderate	0.19	0.42	0.23	0.16
	Non-expressive	0.37	0.05	0.45	0.12
	Negative	0.17	0.25	0.24	0.35

Note: Numbers in bold are probabilities of being in the same emotional status in 9th and 11th grades.

Predicting Latent Status Membership

The final step of our analysis examines how family, peer, school, and community-related factors are associated with rural adolescents’ latent status membership in 9^th grade. To achieve this goal, we applied multinomial logistic regression and chose youth from the negative status as our reference group. Each odds ratio was interpreted as the effect of the covariates on the odds of membership in a particular emotional status profile relative to membership in negative status. Table 5 summarizes the findings of the model predicting latent emotional status in 9^th grade.

Table 5.

Multinomial Logit Model Predicting Latent Status Membership in 9^th Grade (odds ratios reported)

	Latent Status
	Positive vs. Negative	Moderate vs. Negative	Non-expressive vs. Negative
Covariates
Female	0.31*	0.36+	0.28***
	[0.01,1.02]	[0.08,1.58]	[0.12, 0.64]
Subjective family means	1.91+	1.16	1.01
	[0.70,5.22]	[0.38,3.55]	[0.37,2.72]
Argued with both parents	0.30***	0.59	0.42+
	[0.14,0.63]	[0.25,1.40]	[0.13,1.40]
Like community “a lot”	8.13***	5.64**	3.86***
	[1.89, 35.02]	[0.92,34.50]	[1.02, 14.58]
Like school “a lot”	3.44***	3.72**	1.15
	[1.69,6.98]	[0.90, 15.37]	[0.41,3.24]
Friends smoke	0.45	0.89**	0.76
	[0.2,1.01]	[0.21,3.84]	[0.29,1.97]
Friends drink	0.76***	0.87	0.28**
	[0.37, 1.59]	[0.29,2.60]	[0.08,0.95]

Notes: The coefficients are presented in odds ratios, 95% confidence intervals are presented in brackets. Negative status is the reference group; membership in all other latent statuses is compared with this status.

⁺P < .1,

^*P < .05,

^**P < .01,

^***P < .001.

As Table 5 shows, rural girls are more likely to be in the negative status than in the other 3 statuses. Youth who report having desirable family income (subjective family means) have marginally significantly higher odds of being in the positive status relative to negative status. In terms of family relationships, compared to youth who never argued with parents, youth who often argue with parents have significantly lower odds of being in the positive emotional status relative to the odds of being in the negative status. We also found that across all of the covariates, youths’ feelings toward community and school have the largest explanatory power in predicting youths’ emotional status membership. In particular, youth who report that they like community a lot have higher odds of being in positive, moderate, or non-expressive status, relative to being in the negative status (OR=8.13, OR=5.64, OR=3.86, respectively). In other words, youth who report liking community are less likely to be in the negative emotional status. The same patterns also appear in the covariate like school a lot. Liking school is associated with higher odds of being in the positive or moderate emotional statuses relative to being in the negative status (OR=3.44, OR=3.72). Rural youth who report having close friends who smoke have lower odds of being in moderate relative to negative status. The same pattern also appears for the youth who report having friends who drink. Youth with friends who drink have higher odds of being in negative emotional status, relative to the other 3 statuses.

Discussion

Rural areas often are portrayed as rich in family networks and local community support for youth,⁴⁴ although this image does not hold for all youth in rural areas, or for all rural areas. In addition, rural areas have documented shortages in health care providers, especially those providing mental health services.^9,45,46 Because the context for rural youth may be quite different from that for urban and suburban youth, it is important to separately examine the emotional health and change in emotional health of rural youth.

Our analysis provides important details regarding how rural youth move across different emotional health statuses over time, and whether youth who have different family, school, and peer influences differ qualitatively in their emotional health status. Several findings from the LTA are important to underscore.

To begin, instead of measuring emotional health from a purely negative perspective, (distress, depression, and anxiety) or a purely positive perspective (in terms of well-being), our study includes both positively and negatively worded survey items, which enabled us to detect unique patterns of emotional health status that would otherwise be obscured when positive and negative aspects of mental health are examined separately. In revising the theoretical debate on single continuum versus dual continuum of emotional health construct, the evidence in this study aligns with the latter. For example, the LTA result suggests that there exists a sub-group of rural youth who reported experiencing predominantly positive (or negative) emotions. However, there is another group of youth who endorse happy and confident emotions, but also report relatively moderate probabilities of being overwhelmed and nervous (moderate status), suggesting the presence of positive emotions do not necessarily indicate the absence of negative emotions. Overall, our study reveals nuanced patterns of youths’ emotional status among sub-groups of rural youth. Instead of regarding rural youth as a homogenous group, our study highlights the importance of acknowledging heterogeneities within rural youth populations.

Second, there is a substantive percentage of rural youth who are non-expressive in emotions. There are many reasons youth could become less expressive in their emotions. Some studies interpret youths’ non-expressiveness of emotions as being emotionally numb or emotional inhibition, which are adversely related with psychological functioning,^47–49 while others see non-expressiveness as an outcome of youths’ emotional regulation, indicating their strong self-control and maturity.⁵⁰ While it is difficult to differentiate the nature of non-expressive youth given the available information in our sample, our study nevertheless suggests that it is important to consider the level of expressiveness as an independent factor when examining rural youths’ emotional health status. Non-expressive youth suggest a distinctive sub-group of youth that are not identified or recognized in the dual-continuum model.

Third, from the rural service providers’ perspective, examining the prevalence of different emotional health statuses and the extent to which they change during high school is critical for determining the likely caseload and need for service providers, as well as the share of the student body affected.⁵¹ For example, moderate youth and negative youth combined are roughly one-quarter of students in both 9^th and 11^th grades, and for both groups roughly 4 out of 10 remain in that emotional status from 9^th to 11^th grade. They also may be least “visible” in terms of needing or benefitting from mental or emotional health programs, but they may benefit from relatively modest efforts to improve emotional health. In addition, those who exhibit negative emotions in 9^th grade, although a small share of students, may benefit the most from efforts to provide support and guidance. Almost half of these youth stay in the negative status in 9^th and 11^th grades. Youth most at risk of poor outcomes later in high school due to poorer emotional health in 9^th grade may need the most intervention and assistance, with special attention paid to identifying these individuals early. Programs to alert school personnel, students, parents, and school volunteers to characteristics and behaviors associated with different emotional health statuses may increase the likelihood that youth with problems are identified and assisted.^51,52

Fourth, our results also shed light on the contextual factors that affect rural youths’ emotional health status. In addition to identifying patterns and transitions of emotional health statuses, we explored family, school, community, and peer influences on rural adolescents’ emotional health. In particular, we found that negative parent-child relationships and family economic hardship could contribute to rural youths’ negative emotional status. Moreover, other factors in a rural context serve as protective factors for rural youths’ development. For example, favorable communities and school environments are related with youths’ greater chances of feeling positive. However, negative peer influences serve as risk factors for reporting negative emotional health. Our study finds that all else being equal, peer substance use is associated with higher odds of being in a negative emotional health status, relative to non-expressive or moderate status. This finding is consistent with the existing literature on the associations between the depressed symptoms and youth substance use.^2,3 The findings on gender differences in emotional health status also are consistent with existing literature.⁵³ In particular, rural girls are more likely to be in a negative emotional status relative to other emotional statuses. Previous studies have found that among girls, a negative emotional status may be more likely to lead to behavioral problems such as eating disorders and self-harm.⁵⁴ Understanding and identifying states of emotional health and changes in emotional health among youth can contribute to timely and effective interventions. This may be especially helpful for families and schools in rural areas, where there are fewer mental health professionals and youth may face difficult decisions as they plan for the future.

Limitations and Conclusion

As with all studies, this research has limitations. Our study focuses on emotional health among rural Pennsylvania youth. Therefore, generalizations to other youth populations should be made with caution. In addition, although our study incorporates measurements of positive and negative experiences, the RYE data did not include a full assessment of all possible desirable or undesirable emotions that rural youth experience. It is possible, for example, that youth who are non-expressive in all 8 measurements in our study may report having salient emotional patterns in other domains. Future research may expand the current research by collecting data using a more comprehensive assessment of youths’ emotional experiences (such as the 12-item measures of the Scale of Positive and Negative Experience).⁵⁵ Furthermore, our study examined change and stability of youths’ emotional health status during high school. Additional research could investigate stability—improvements or declines in emotional health over a much longer time frame.

In conclusion, our study supports the theoretical and empirical understanding of emotional health of youth. In revisiting our 3 research objectives, we identified 4 major sub-groups of rural youth in Pennsylvania who exhibit distinctive emotional health statuses: positive, moderate, negative, and non-expressive. A large share of these youth remains stable in their emotional health status from 9^th to 11^th grade, while substantial movement among different mental health statuses also occurs. Supportive school and community environments serve to promote positive emotional health, while negative parental-children relationships and peer smoking and drinking are closely related with negative emotional status.