CASE STUDY On Presidential Research Report / Article to review is attached. Also 1 additional source is included.

Trump, Obama, Bush: Impacts of Presidential Elections on College Student Mental Health Brett M. Merrill a, Heidi Vogelera, Jessica Kirchhoefera, Shannon Tassa, Davey Erekson a, Mark Beechera, Vaughn Worthen, Klint Hobbsa, R. D. Boardmana, Jennie Binghama, Russell J. Baileya, Jonathan C. Coxa, Dever M. Carneyb, J. Ryan Kilcullenb, and Derek Griner a

aDepartment of Statistics , Brigham Young University, Provo, Utah, USA; bDepartment of Psychology, Pennsylvania State University, Pennsylvania, USA

ABSTRACT Research, media sources, and polls have identified negative effects associated with presidential elections. The aim of this research was to investigate associations between US presiden- tial election results and mental health outcomes in university students. This investigation consisted of two independent stu- dies. Study 1 analyzed data collected between the years 2000 and 2016 from students who utilized counseling services (N = 32,506) at a large, private, conservative institution in the western United States. Study 2 analyzed data collected between the years 2010 and 2016 from over 100 university counseling centers across the United States. Upon analyzing the quantita- tive and qualitative data, the results did not support the findings that presidential elections negatively impact the mental health outcomes of students who receive university counseling ser- vices. Furthermore, there was no detectable increase in student distress regardless of election year, age, ethnicity, gender, reli- gion, relationship status, sexual orientation, geographic region, citizenship, and first-generation student status.

KEYWORDS Presidential election; mental health; anxiety; Trump; treatment outcomes

Introduction

Over the past several United States (US) presidential election cycles, partisan- ship, polarization, and antipathy have increased among the US electorate (Pew Research Center, 2014). Impacts on US residents in these elections are very difficult to assess, however, given a paucity of empirical research. The few extant studies addressing psychological outcomes related to presidential elections suggest varied reactions: feelings of elation, optimism, anger, hopelessness, and fear, depending on the outcome of the elections (McCowan, 2015; Valentino, Brader, Groenendyk, Gregorowicz, & Hutchings, 2011). Some post- election studies have focused on positive effects, including feelings of optimism, perceptions of diminishing racism, and increases in efficacy (Franco & Smith- Bynum, 2016; McCowan, 2015; Merolla, Sellers, & Fowler, 2013). Other studies

CONTACT Brett M. Merrill, PhD. brettmm@byu.edu 1500 WSC Brigham Young University Provo, UT, USA 84602

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY https://doi.org/10.1080/87568225.2021.1888216

© 2021 Taylor & Francis

report no significant post-election differences concerning perspectives, lived experiences, aspirations to become president in children, and racial attitudes (Jones, 2014; Patterson, Pahlke, & Bigler, 2013; Schmidt & Axt, 2016). Yet some research, media sources, and polls have identified negative effects associated with presidential elections, such as increases in stress, anxiety, fear of incidents of hate and bias, physiological symptoms of stress, and triggers for past trauma (APA, 2019; DeJonckheere, Fisher, & Chang, 2018; Sifferling, 2016; SPLC, 2016). In fact, following the 2016 election, Newsweek magazine introduced the term “Trump Anxiety Disorder” to capture what they saw as a strong negative reaction to the election of Donald Trump (https://www.newsweek.com/thera pists-report-rise-anxiety-trump-was-elected-1046687).

In regards to the 2016 presidential election, specific psychological impacts on college-age individuals were identified. For example, the American Psychological Association conducted a “Stress in America Survey” in conjunc- tion with the 2016 election. They found that millennials (ages 18– 37) experi- enced increased stress that exceeded other age groups following the most recent election. The New York Times also shared that post-election contention and campus divisiveness were associated with feelings of fear and lack of safety for both liberal and conservative students alike (Hartocollis, 2016).

Researchers have examined the differential impacts on students from var- ious minority backgrounds related to social policy, discrimination, and finan- cial barriers (e.g., Grable & Joo, 2006; Mortenson, 2000; Pieterse, Carter, Evans, & Walter, 2010). For a variety of reasons, both international students and students from racial/ethnic minority backgrounds tend to underutilize mental health services available on college campuses (e.g., Clement et al., 2015; Davidson, Yakushka, & Sanford-Martens, 2004; Pendse & Inman, 2017). Generally speaking, international students attending university in the U.S. receive limited support while they face adjustments in learning new social and educational cultures (Olivas & Li, 2006).

Some studies have shown that discrete events can impact college stu- dent distress. For example, Lambert, Lambert, and Lambert (2014) sug- gested most students can handle typical stressful events like balancing course loads, employment, family, and social life, but some sources of stress beyond the control of students can have powerful effects. Events like terrorist attacks, school shootings, and natural disasters can be vicariously traumatizing to students, even if they were not directly affected or in proximity when the event occurred. These studies suggest that discrete events can impact student distress; however, there is sparse information linked to how a catalyzing political event (such as a presidential election) may spur a rapid change in psychological functioning in different student groups.

Within college student populations, specific student groups may have more personal connections to certain political issues, leading them to experience

2 B. M. MERRILL ET AL.

greater impact from political or social policies associated with different pre- sidential administrations. Examples include DACA students, students affected by executive order 13,769 (travel ban), or students affected by racism (Villavicencio, 2018). However, there is a gap in the literature regarding the specific impacts of presidential elections on college student psychological functioning, especially for minority and international students who might be adversely impacted.

Purpose of the study

The aim of this study was to investigate the possible association between US presidential election results and mental health outcomes in university stu- dents. Understanding how potential stressors impact student groups – and consequently how and whether counseling services are sought – would inform university counseling centers how to best meet the clinical needs of their students during presidential elections periods.

Research questions

This investigation consisted of two independent studies. Study 1 utilized data from a single university, with the following research questions:

(1) Do individual psychological outcomes change before and after a presidential election, and do they change when accounting for various demographic variables?

(2) Are crisis services utilized more frequently around the time of presi- dential elections, and do number of crisis appointments differ by citi- zenship and ethnicity?

(3) Are there qualitative indications of distress that are observable when examining case notes near presidential elections?

Study 2 utilized data from over 100 university counseling centers across the United States, with the following research questions:

(1) Do individual psychological outcomes change before and after a presidential election, and do they change when accounting for various demographic variables, including geographic region of the counseling center?

Study 1 methods

In order to better understand the impact of recent presidential elections on the mental health of college students generally and more specifically on college

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 3

students from various minority populations, archival data from a university counseling center were utilized to look at how distress scores changed before and after different presidential elections. These data included session-by- session psychotherapy outcome data as well as appointment attendance data. Upon intake, clients involved in this study completed an informed consent form that granted permission to use their de-identified information as part of future research. Approval for this informed consent was provided by this university’s Institutional Review Board (IRB).

Participants

The data for this study were collected between the years 2000 and 2016 from a large, private, conservative, religiously affiliated institution in the western United States. Participants were university students who utilized counseling center services (N = 32,506). The majority of clients were between the ages of 18 and 25, and identified as European American. For this study, participants were grouped based on different demographic variables that were collected by measurement at intake.

Measurement instruments

Demographic questionnaire (DQ) This measure gathers general demographic information such as gender, age, ethnicity, and citizenship. It also collects data on experiences and attendance in previous counseling. Political affiliation was not part of this questionnaire.

Outcome questionnaire-45 (OQ-45) The Outcome Questionnaire-45 (OQ-45; Lambert, Lunnen, Umphress, Hansen, & Burlingame, 1997) was utilized to measure mental health distress among students. The questionnaire is a 45-item survey that utilizes a five-point Likert scale. Items are scored on a scale from “Never” to “Almost Always” and participants are instructed to endorse how often they experience specific feel- ings, thoughts, or physiological symptoms within the past week. The OQ-45 has excellent internal reliability (alpha at .93), test–retest reliability (.84), and con- current validity with other instruments (Kadera, Lambert, & Andrews, 1996; Umphress, et. al, 1997). Subscale scores for the OQ-45 are also available, but they mostly describe the theoretical structure of the measure. Therefore, the total score was utilized in calculating differences instead of subscale scores. The OQ-45 was administered at intake and before every counseling session.

Clinician case notes Qualitative indications of distressed students were assessed using an examina- tion of case notes. Case notes from individual, couples, and group

4 B. M. MERRILL ET AL.

psychotherapy were scoured using search terms related to presidential elec- tions (Trump, Obama, election, Republican, Democrat, conservative, liberal, voted). Notes from 3 days pre- and post-election during the years 2008– 2016 were included in the search.

Procedure

The original sample of 32,506 was analyzed, and only cases that met the following criteria were retained: no missing OQ subscales and/or no more than 5 missing OQ responses; at least one OQ administration within the 14 days prior to an election date; and at least one OQ administration within the 14 days following an election. A 14-day window pre- and post-election was chosen, because most clients at the center were seen on a weekly, or bi-weekly basis. This span allowed for a large portion of clients to have at least two therapy sessions within this time frame; once before the election, and once after. After these cases were selected, if more than one OQ administration was identified in either window, the case with the OQ administration closest to the election was retained. This analysis resulted in a sub-sample of 6,616 cases with paired pre- and post-election OQ administrations. A summary of the demographics for this sub-sample is provided in Table 1.

Analysis

Data were analyzed utilizing both R and SPSS 26. Cases were selected for models based on proximity of the OQ administration to the election year. In the first model, cases from 2004 (election year) were compared to cases from 2003. The next three models followed suit, with cases from 2008 (election year) being compared to 2007, cases from 2012 (election year) being compared to cases from 2011, and cases from 2016 (election year) being compared to cases

Table 1. Demographics of paired sub-sample (n = 6,616). Category n % Category n %

Gender Ethnicity Female 3,687 55.73 White 5,323 80.46 Male 2,450 37.03 Hispanic 335 5.06 Unreported 479 7.24 Asian 194 2.93

Age Pacific Islander/Hawaiian 79 1.19 <18 27 0.41 Black 46 0.70 18 541 8.18 American Indian 96 1.45 19 649 9.81 Other/Missing 543 8.21 20 695 10.50 Citizenship 21 836 12.64 USA 5,856 88.51 22 904 13.66 Canada 43 0.65 23 840 12.70 Mexico 50 0.76 24 696 10.52 Other/Missing 667 10.08 25 467 7.06 26 288 4.35

>26 673 10.17

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 5

from 2015. Post-election years were not used for comparison purposes in order to keep the focus on the event of presidential elections, instead of impact of presidencies over times. A summary of the demographics for each of these models is presented in Table 2.

Study 1 results

Question 1

In order to address the first research question, a difference score was calculated for each case by subtracting the pre-election OQ score from the post-election OQ score. The mean difference scores were then com- pared between pre-presidential years and presidential years in each model, as well as in the aggregate, utilizing Independent Samples t-tests. In all models except the 2012 model, the pre-presidential mean difference score was more than the presidential mean difference score, though not signifi- cantly so. In the 2012 model, the presidential mean difference score was more than the pre-presidential mean difference score. The results are shown in Table 3.

Repeated Measures ANOVA was also utilized to examine the main effects of the difference between OQ scores both before and after and election (Time), whether it was a pre-presidential or presidential year (Presidential), and the interaction between the two. In all models, Time was a significant predictor of

Table 2. Demographics of model samples (N = 3,721). 2004 Model 2008 Model 2012 Model 2016 Model

2003 2004 2007 2008 2011 2012 2015 2016

Total (n) 395 386 356 392 510 564 568 550 Gender F 230 223 193 215 274 316 337 338

M 132 133 136 140 198 215 202 177 Other/Missing 33 30 27 37 38 33 29 35

Ethnicity White 328 309 294 321 404 456 452 433 Hispanic 13 14 13 17 18 28 37 38

Asian 6 9 8 5 18 18 24 22 Pacific Island/Hawaiian 3 6 3 5 8 11 11 6

Black 0 2 1 1 4 6 7 4 American Indian 6 5 8 4 13 9 5 10

Other/Missing 39 41 29 39 45 36 32 37 Citizenship USA 352 351 324 351 467 523 473 426

Canada 5 2 3 2 2 4 4 5 Other/Missing 38 33 29 39 41 37 91 112

AgeGroup 18– 22 215 215 196 220 269 312 327 333 23– 25 127 117 99 112 161 171 170 158

26+ 53 54 61 60 80 81 71 59 OQdiff Mean −2.81 −1.40 −2.31 −1.72 −0.97 −1.61 −3.03 −1.58

SD 12.92 12.17 15.18 14.49 12.95 13.04 13.31 13.52 Days_Between Mean 11.02 10.82 10.22 11.86 11.20 11.29 12.63 13.53

SD 4.15 4.55 4.12 4.90 4.05 4.30 4.96 4.79

AgeGroup = Age at time of first OQ Administration; OQdiff = OQ score post-election minus OQ score pre-election; Days_Between = number of days between OQ administrations

6 B. M. MERRILL ET AL.

OQ score, meaning OQ scores decreased as Time increased. However, Presidential was not a significant predictor; neither was the interaction between Time and Presidential. These results are presented in Table 4.

To account for the impact of various demographic variables, the four models were examined utilizing an Ordinary Least Squares (OLS) Regression Model. The dependent variable for all models was the calculated OQ difference score. A negative difference score equated to a lower OQ score after the election; while a positive difference score equated to a higher OQ score after the election, suggesting more distress.

In addition to OQ difference scores, each model contained six independent variables: number of days between OQ scores (Days_Between); age based on

Table 3. Mean difference in OQ scores: Independent samples t-test. Election Type n x̄ sd F Sig t Sig Effect Size

2004 Model Pre-Presidential 395 −2.81 12.92 1.65 0.20 −1.57 0.12 0.11 Presidential 386 −1.40 12.17

2008 Model Pre-Presidential 356 −2.31 15.18 0.03 0.86 −0.55 0.58 0.04 Presidential 392 −1.72 14.49

2012 Model Pre-Presidential 510 −0.97 12.95 0.33 0.57 0.80 0.42 0.05 Presidential 564 −1.61 13.04

2016 Model Pre-Presidential 568 −3.03 13.31 0.01 0.91 −1.81 0.07 0.11 Presidential 550 −1.58 13.52

Aggregate Pre-Presidential 1829 −2.27 13.53 0.66 0.42 1.57 0.12 0.05 Presidential 1892 −1.58 13.32

F = Levene’s Test for Equality of Variances

Table 4. Repeated measures ANOVA. Within-Subjects F Sig Between-Subjects F Sig

2004 Model Time 21.79 0.00 Time x Presidential 2.46 0.17 Presidential 2.39 0.12

2008 Model Time 13.83 0.00 Time x Presidential 0.30 0.58 Presidential 0.55 0.46

2012 Model Time 10.57 0.00 Time x Presidential 0.64 0.42 Presidential 0.04 0.83

2016 Model Time 33.08 0.00 Time x Presidential 3.26 0.07 Presidential 1.66 0.20

Aggregate Time 76.64 0.00 Time x Presidential 2.45 0.12 Presidential 3.28 0.07

Time = Difference between OQ scores before and after an election; Presidential = Type of election; Time x

Presidential = Interaction between OQ scores and election type

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 7

time of first OQ score (AgeGroup); Ethnicity; Gender; Citizenship; and whether the OQ scores were obtained in a presidential election year or not (Presidential). An examination of the results showed that number of days between OQ scores was a significant predictor in the 2004 model as well as in the aggregate model, with difference scores increasing as number of days between OQ scores increased. Presidential was a significant predictor only in the 2016 model (p = .04, effect size = 0.07), with difference scores decreasing significantly in the non-presidential election year. The findings of significance were confirmed utilizing stepwise regression techniques in order to isolate the significant predictor variables from all other predictors.

Question 2

To determine the impact of presidential elections on crisis services, a seven- day pre- and post-election window was utilized (see Figure 1). The 2012 and 2016 election year models were the only models with enough crisis visits to provide meaningful comparisons. A chi-square comparison of the number of crisis visits resulted in no statistical difference between election and non- election years. In terms of the impact of citizenship, the number of crisis visits for citizenship other than USA was not high enough to utilize chi- square comparison. The same holds true for ethnicity, with White being the only ethnicity with enough visits for chi-square comparison. In examining

Figure 1. Number of “crisis” appointments within 7 days of an election day.

8 B. M. MERRILL ET AL.

the chi-square comparison for aggregate citizenship data, no significant differences were found.

Question 3

The final question in Study 1 related to qualitative indications of distress related to presidential elections. A total of 13,699 case notes were reviewed. After removing duplicate notes from group psychotherapy, and notes that did not include any of the predefined search terms, only 23 case notes with identifiable election content remained: two case notes in 2008; three case notes in 2012; and 18 notes in 2016. Two independent raters coded references to election content in the case notes. An inter-rater reliability (IRR) score of .83 was calculated for the included search terms, suggesting adequate agree- ment between the two raters. An IRR of .98 was calculated for the valence (Positive, Negative, or Unclear) of the election term within the case note, suggesting very high agreement. The majority of political content within the notes was labeled as “negative.” One rater labeled 17 of the 23 case notes as negative; five as unclear; and one as positive. The other rater determined 16 of the 23 case notes as negative; six as unclear; and one as positive.

Study 2 methods

Study 2 sought to identify whether there might be more significant differences in student distress in association with presidential elections when looking at a nationally representative sample. Archival data from the Center for Collegiate Mental Health (CCMH), a collaborative effort of 619 college and university counseling centers, was utilized for this study. Approval for this study was provided by the IRB as well as the CCMH.

Participants

The data for this study were collected between the years 2010 and 2016 from college and university counseling centers who participate in the CCMH collaborative. Participants were university students who utilized counseling center services in a participating counseling center. As with the previous study, the raw data were paired based on students having at least one administration of the Counseling Center Assessment of Psychological Symptoms (CCAPS) within the 14 days prior to an election date (presidential or not); and at least one CCAPS administration with the 14 days following an election. A 14 day window pre- and post-election was chosen, because most clients at counseling centers were seen on a weekly, or bi-weekly basis. This span allowed for a large portion of clients to have at least two therapy sessions within this time frame; once before the election, and once after. Once these cases were selected, if

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 9

more than one CCAPS administration was identified in either window, the case with the CCAPS administration closest to the election was retained. This analysis resulted in a sub-sample of 20,327 cases with paired pre- and post- election CCAPS administrations.

Measurement instruments

Standardized data set (SDS) The SDS is a standardized set of demographic and other data materials that is utilized by counseling centers that participate in the CCMH.

Counseling center assessment of psychological symptoms (CCAPS) The CCAPS-34 was first released in September 2009, and was then updated in 2012. This instrument is a standardized measure of treatment outcome used by counseling centers that participate in the CCMH. It is comprised of 34- items that load on seven distinct subscales related to psychological symptoms and distress in college students. The scales are: Depression; Anxiety; Social Anxiety; Academics; Eating; Hostility; and Alcohol. It also incorporates an overall Distress Index. The instrument takes approximately 2–3 minutes to complete, and can be used at any point in treatment.

Analysis

As with the previous study, data were analyzed utilizing both R and SPSS 26. Cases were selected for models based on the year data was collected. However, in this study, there were only two models: cases from 2012 being compared to cases from 2011; and cases from 2016 being compared to cases from 2015.

Study 2 results

In order to address the research question, a difference score was calculated for each CCAPS Index Score for each case by subtracting the pre-election CCAPS score from the post-election CCAPS scores. The mean scores were then compared between pre-presidential years and presidential years in each model, as well as in the aggregate, utilizing independent-samples t-tests. In the 2012 model, the only significant difference (effect size = 0.12) was asso- ciated with the Eating index, where scores in the presidential year increased rather than decreased. In the 2016 model, the Depression, Anxiety, Academic, Hostility, and Distress indices were all significantly different (effect sizes = 0.067, 0.045, 0.037, 0.073, 0.071), again with the presidential year showing less improvement than non-presidential year. This pattern was mirrored in the aggregate comparison as well (see Table 5).

10 B. M. MERRILL ET AL.

Repeated Measures ANOVA was utilized to examine the main effects of CCAPS scores both before and after an election (Time), whether it was a pre- presidential or presidential election (Presidential), and the interaction between the two. The results of these analyses are shown in Table 6.

The models were also examined utilizing an Ordinary Least Squares (OLS) Regression Model. The dependent variable for all models was the CCAPS Index difference scores. A negative difference score equated to a lower CCAPS index score after the election; while a positive difference score equated to a higher CCAPS index score after the election, suggesting more distress. In addition to CCAPS difference scores, each model contained six independent variables: number of days between CCAPS scores (Days_Between); age based on time of first CCAPS score (AgeGroup); Ethnicity; Gender; USA Citizenship

Table 5. Aggregate independent samples t-test (N = 13,789). CCAPS Index

Pre-Presidential (n = 6,256) Mean CCAPSdiff (sd)

Presidential (n = 7,533) Mean CCAPSdiff (sd) t value Sig Effect Size

Depression −0.163 (0.631) −0.122 (0.621) 3.753 0.000 0.065 Anxiety −0.140 (0.579) −0.113 (0.583) 2.728 0.006 0.046 Social Anxiety −0.085 (0.499) −0.073 (0.480) 1.463 0.144 0.025 Academic −0.099 (0.672) −0.072 (0.664) 2.415 0.016 0.040 Eating −0.054 (0.596) −0.034 (0.578) 1.979 0.048 0.034 Hostility −0.090 (0.473) −0.063 (0.448) 3.354 0.001 0.059 Alcohol −0.061(0.453) −0.053 (0.432) 1.060 0.289 0.018 Distress Index −0.137 (0.448) −0.107 (0.440) 3.997 0.000 0.068

Table 6. Repeated measures ANOVA.

CCAPS Index, Election Type, and Interaction

2012 Model 2016 Model Aggregate Model

F Sig F Sig F Sig

Depression 130.243 0.000 584.383 0.000 709.012 0.000 Depression x Presidential 0.485 0.486 13.683 0.000 14.083 0.000 Presidential 2.449 0.118 1.683 0.195 0.450 0.502 Anxiety 63.482 0.000 585.138 0.000 646.350 0.000 Anxiety x Presidential 1.299 0.255 6.177 0.013 7.439 0.006 Presidential 3.125 0.077 1.666 0.197 3.524 0.061 Social Anxiety 55.225 0.000 300.873 0.000 355.507 0.000 Social Anxiety x Presidential 0.638 0.424 1.567 0.211 2.154 0.142 Presidential 6.711 0.010 2.989 0.084 6.488 0.011 Academic 48.661 0.000 177.291 0.000 223.820 0.000 Academic x Presidential 1.743 0.187 4.113 0.043 5.833 0.016 Presidential 9.612 0.002 0.409 0.523 2.770 0.096 Eating 4.664 0.031 72.941 0.000 76.575 0.000 Eating x Presidential 5.873 0.015 1.316 0.251 3.939 0.047 Presidential 3.105 0.078 1.279 0.258 0.126 0.722 Hostility 70.481 0.000 311.970 0.000 378.684 0.000 Hostility x Presidential 1.317 0.251 16.236 0.000 11.356 0.001 Presidential 0.204 0.651 3.167 0.075 2.543 0.111 Alcohol 39.819 0.000 188.221 0.000 227.801 0.000 Alcohol x Presidential 1.336 0.248 0.441 0.506 1.134 0.287 Presidential 4.041 0.045 0.812 0.367 2.923 0.087 Distress Index 150.005 0.000 885.022 0.000 1034.825 0.000 Distress Index x Presidential 1.098 0.295 14.787 0.000 15.978 0.000 Presidential 3.824 0.051 0.035 0.852 0.248 0.619

CCPAS Index x Presidential = Interaction Term between CCAPS Index and Election Type; CCAPS Index and Interaction Term = Within Subjects; Presidential = Between Subjects

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 11

or other (Citizenship), and whether the CCAPS scores were obtained in a presidential election year or not (Presidential). The results of each model are included in Tables 7 and 8.

Finally, the models were examined utilizing factorial ANOVA. The depen- dent variable for all models was the difference in CCAPS Distress Index scores, and the independent variables were region of the USA and presidential election year. The interaction between the two variables was also examined.

Table 7. 2012 model: Regression based on individual subscales (n = 1,661). CCAPS Index

Significant Predictor Variables (in the presence of all variables)

Unstandardized β t Sig

Effect Size

Depression Days_Between −0.006 −2.246 0.025 0.063 R2 = 0.014 Citizenship −0.093 −2.292 0.022 0.029

Presidential −0.180 −3.233 0.001 0.079 Anxiety Days_Between −0.006 −2.356 0.019 0.065 R2 = 0.006 Social

Anxiety N/A

R2 = 0.005 Academics Gender 0.064 2.079 0.038 0.009 R2 = 0.007 Eating N/A R2 = 0.005 Hostility N/A R2 = 0.003 Alcohol N/A R2 = 0.004 Distress Presidential −0.089 −2.110 0.035 0.026 R2 = 0.008

Days_Between = Number of days between CCAPS administrations; Gender: Male = 0; Citizenship: USA = 0

Table 8. 2016 mode: Regression based on individual subscales (n = 12,128).

CCAPS Index Significant Predictor Variables

(in the presence of all variables) Unstandardized β t Sig Effect Size

Depression AgeGroup 0.023 2.918 0.004 0.030 R2 = 0.004 Gender 0.027 3.518 0.000 0.049

Presidential −0.039 −3.407 0.001 0.034 Anxiety Days_Between −0.002 −2.508 0.012 0.022 R2 = 0.004 AgeGroup 0.019 2.551 0.011 0.027

Gender 0.016 2.252 0.024 0.045 Presidential −0.021 −2.028 0.043 0.023

Social Anxiety Gender 0.013 2.127 0.033 0.013 R2 = 0.001 Citizenship −0.030 −2.182 0.029 0.011 Academics Days_Between −0.004 −3.428 0.001 0.031 R2 = 0.003 Gender 0.024 2.935 0.003 0.036 Eating N/A R2 = 0.000 Hostility AgeGroup 0.015 2.562 0.010 0.025 R2 = 0.003 Gender 0.015 2.625 0.009 0.033

Presidential −0.031 −3.710 0.000 0.037 Alcohol Days_Between 0.001 2.072 0.038 0.018 R2 = 0.001 Citizenship 0.040 3.205 0.001 0.031 Distress Days_Between −0.002 −2.955 0.003 0.027 R2 = 0.006 AgeGroup 0.019 3.403 0.001 0.031

Gender 0.021 3.901 0.000 0.035 Presidential −0.027 −3.317 0.001 0.030

Days_Between = Number of days between CCAPS administrations; AgeGroup = Age at first CCAPS administration; Gender: Male = 0; Presidential = type of election; Presidential: Presidential Election = 0; Citizenship: USA = 0

12 B. M. MERRILL ET AL.

In both the 2016 model and the aggregate model, election year status was the only significant effect, with R2 values of 0.002 and 0.001, respectively. In the 2012 model, none of the effects was significant, and the R2 value was 0.003.

Discussion

The 2016 election results attention suggesting Trump’s election elicited a great deal of distress, fear, contention, and divisiveness (Hartocollis, 2016). Given the amount of media attention focused on potential student distress related to the most recent presidential election, it was presumed that some subgroups of students experienced elevated distress related to the election. Study 1, con- ducted in a single site, large, private, conservative university located in the western US, found no difference in distress pre- or post-election for any groups of students who were engaged in counseling services over multiple election year cycles. More specifically, no significant differences in OQ-45 scores and no change in crisis utilization (30 days before or after the election, 15 days before and after, and 7 days before and after) for students were found. Moreover, there were no significant differences in OQ-45 scores during the two weeks before, or after, the election, when accounting for various demo- graphic variables (gender, age, ethnicity, or citizenship).

In an effort to examine both objective and subjective data regarding the election, therapy case notes were reviewed for terms associated with past elections. Review of clinician notes offered extremely sparse evidence for elevated concerns about the election (17 out of 13,699 notes). Thus, even though the dataset for Study 1 was very large (N = 32,506), with ample power to detect meaningful differences, results yielded no significant quanti- tative or qualitative evidence for elevated distress related to the presidential election.

Given these counterintuitive findings in light of such strong media atten- tion, it is possible that the lack of significant findings were the result of a constricted sample (i.e., private, conservative, predominantly White, reli- giously sponsored institution in the West). Thus, Study 2 was initiated, which examined a nationwide collegiate dataset from over 600 college and university counseling centers across the US. Various demographic variables were utilized in our analyses that might have been differentially impacted by previous election cycles: ethnicity; gender; religion; relationship status; sexual orienta- tion; international student status; and whether students in the sample were first-generation students. Data by geographic region of the US (Midwest, Northeast, South, and West) were also analyzed in an effort to examine areas that are typically conveyed as having high or low support for different presidential candidates.

After accounting for these myriad variables, there was still no significant evidence that presidential elections negatively impacted students’ mental

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 13

health, regardless of demographics. Additionally, results suggested no signifi- cant differences by geographic region in terms of CCAPS Distress Index, indicating that there was no difference in overall distress level measured among students of differing geographic regions. In sum, with a large nation- wide sample that included multiple demographic variables, there was no significant evidence that students receiving counseling services in a collegiate counseling center were negatively, or positively, impacted by recent presiden- tial election cycles.

Unremarkable aggregate data does not diminish significant individual experiences. A review of therapy case notes in Study 1 suggested a few students were negatively impacted by Trump’s election into office, but the data suggest these experiences seem to be more the exception than the rule. It is possible that the media has overstated the impact of presidential elections on college students by identifying a few vocal outliers and incorrectly broadcasting their experiences as commonplace. That being said, anecdotally many of the authors of this paper reported either having their clients (especially minority students) indicate distress or the authors personally experienced distress immediately following the most recent election results.

Limitations

There are several factors that limit the findings of this study. All participants in the sample for Study 1 were from the same large, private, religiously oriented, predominantly White university. Because nearly all (98%) students attending this university belong to the same faith, it is likely that many of them had similar worldviews shaped by their common membership in this faith. Participants in the sample for Study 1 also included only students receiving therapy at this particular university’s counseling center. Extrapolating the experiences of students attending therapy to those who were not attending therapy is inappropriate. Since the university in which Study 1 took place was predominantly White, it was also likely to adequately capture the experiences of students from racially/ethnically diverse backgrounds who are less likely to seek out mental health services (Clement et al., 2015; Kearney, Draper, & Barón, 2005; Smith & Trimble, 2016). The two advantages of the Study 1 dataset were the access to session by session OQ-45 scores and the ability to examine counselor case notes. In sum, there were significant limitations associated with external validity related to Study 1. To help ameliorate this limitation, data were gathered from the national CCMH dataset to get better representation of students across the country.

All participants in Study 2 were also attending therapy through a campus counseling center. Similar to Study 1, students from racially/ethnically diverse backgrounds were statistically less likely to seek out mental health services. Again, extrapolating the experiences of those attending therapy across college

14 B. M. MERRILL ET AL.

campuses to those who are not is inappropriate. It is plausible that those seeking therapy across college campuses are already more distressed than the general campus body, and the data did not accurately reflect the experiences of those with less distress. Students attending therapy may have significant mental health concerns that took priority over concerns about elections.

Both of the studies relied on archival data and were unable to add variables of interest. For example, the measures administered in both studies were designed to be sensitive to clinically significant change in mental health concerns in a clinical population, but were not designed to measure political distress in a general population. In other words, none of the measures had questions specifically related to political concerns. It is also possible that the measures our participants took were not sensitive to changes in distress that did not rise to a clinical level. Therefore, some election effects may have been missed due to the focus of the instruments or their sensitivity to change. Furthermore, the demographic information questionnaires that were given to students attending therapy had no questions related to political affiliation, which prevented us from knowing how this factor may have influenced the results.

In an effort to gather information about clients’ distress related to the elections that may have been discussed in therapy, case notes of students attending therapy in Study 1 were examined. Case notes were inaccessible for Study 2. The number of case notes that mentioned distress due to the presidential election were sparse. One possible reason for this finding is that the examined notes were written from the lens of the clinician. Hence, the clinician may not have deemed election stress clinically relevant and may not have accurately depicted students’ experiences in regard to the election in their case notes. One other possibility is that students were reluctant to share their distress related to the election, because they were unsure about the political affiliation of their therapist. This may be particularly true for minority stu- dents, who would likely be hesitant to share their viewpoints with conserva- tive, White faculty members in positions of power.

Directions for further research

Many areas exist for further exploration of the intersection between political activity and mental health, and the current study should not be interpreted as a definitive statement on mental health as it is impacted by presidential elections. To more fully understand the relationship between elections and mental health, it would be beneficial to include multiple representative sam- ples in varying life circumstances. For example, recruitment from community mental health centers or Veteran Affairs would likely provide more diversity in terms of socioeconomic status, education levels, ethnicity, etc. Some prelimin- ary evidence suggests dramatic societal events, like presidential elections, can

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 15

differentially effect the health of distinct populations (e.g., Trump supporters may experience increased psychological well-being, pride, and hope, while Anti-Trumpers or marginalized groups may experience increased fear of hostility or discrimination; Williams & Medlock, 2017). Recruitment from more nationally representative samples should allow researchers to specifically examine mental health effects associated with presidential elections.

Research examining the intersection between politics and mental health would also benefit from assessment of political activation and political alle- giance of populations studied as moderators for mental health distress. Allegiance to a particular party or level of political activism are both variables that may have differential impacts on clinical distress or mental health. Pairing measures that evaluate aspects of an individual’s political experience with measures that assess various domains of mental health represents a logical next step in study design.

A final area for future research could examine the impact of political climate or presidential elections on mental health over time. A longitudinal design could elucidate changes in mental health as they occur over the course of different presidencies or as the result of institution of national policies. It is also expected that a longitudinal design may be more sensitive to reactions and distress that would not be present in the immediate aftereffects of a presidential election, but may exhibit a delayed response (Bryant & Harvey, 2002).

Conclusion

The empirical data did not support findings from media reports and anecdotal evidence that presidential elections negatively impact the mental health out- comes of students who receive university counseling services. Furthermore, there was no detectable increase in distress regardless of election year, age, ethnicity, gender, religion, relationship status, sexual orientation, geographic region, citizenship, and first-generation student status. Further inquiry into this potential phenomenon is encouraged.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Brett M. Merrill http://orcid.org/0000-0002-1053-2946 Davey Erekson http://orcid.org/0000-0001-6214-485X Derek Griner http://orcid.org/0000-0002-0378-6403

16 B. M. MERRILL ET AL.

References

American Psychological Association. (2019). Stress in America: Americans feeling stressed about presidential election, health care and mass shootings. Stress in America™ Survey. https://www. apa.org/news/press/releases/stress/2019/stress-america-2019.pdf

Bryant, R. A., & Harvey, A. G. (2002). Delayed-onset posttraumatic stress disorder: A prospective evaluation. Australian and New Zealand Journal of Psychiatry, 36(2), 205–209. doi:10.1046/j.1440-1614.2002.01009.x

Clement, S., Schauman, O., Graham, T., Maggioni, F., Evans-Lacko, S., Bezborodovs, N., . . . Thornicroft, G. (2015). What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies. Psychological Medicine, 45(1), 11–27. doi:10.1017/S0033291714000129

Davidson, M. M., Yakushka, O. F., & Sanford-Martens, T. C. (2004). Racial and ethnic minority clients’ utilization of a university counseling center: An archival study. Journal of Multicultural Counseling and Development, 32, 259.

DeJonckheere, M., Fisher, A., & Chang, T. (2018). How has the presidential election affected young Americans? Child and Adolescent Psychiatry and Mental Health, 12(1), 8. doi:10.1186/ s13034-018-0214-7

Franco, M., & Smith-Bynum, M. (2016). The relationship between racial identity and perceived significance of the election of President Barack Obama among African American mothers. Identity: An International Journal of Theory and Research, 16(3), 182–189. doi:10.1080/ 15283488.2016.1190728

Grable, J. E., & Joo, S. (2006). Student racial differences in credit card debt and financial behaviors and stress. College Student Journal, 40(2), 400–408.

Hartocollis, A. (2016, December 8). On campus Trump fans say they need ‘safe spaces. The New York Times. Retrieved from https://www.nytimes.com/2016/12/08/us/politics/political- divide-on-campuses-hardens-after-trumps-victory.html

Jones, D. (2014). Counseling low-income African American men post-2008 presidential election: A phenomenological study of lived experiences [Unpublished doctoral disseration]. Capella University.

Kadera, S. W., Lambert, M. J., & Andrews, A. A. (1996). How much therapy is really enough: A session-by-session analysis of the psychotherapy dose-effect relationship. Journal of Psychotherapy Research and Practice, 5, 132–151.

Kearney, L., Draper, M., & Barón, A. (2005). Counseling utilization by ethnic minority students. Austin, TX: The Counseling & Mental Health Center.

Lambert, M. J., Lunnen, K., Umphress, V., Hansen, N., & Burlingame, G. M. (1997). Administration and scoring manual for the Outcome Questionnaire (OQ-45.1). Salt Lake City, UT: IHC Center for Behavioral Healthcare Efficacy.

Lambert, S. F., Lambert, J. C., & Lambert, S. J., III. (2014). Distressed college students following traumatic events. VISTAS Online, 1–14. Retrieved from https://www.counseling.org/docs/ default-source/vistas/article_21.pdf?sfvrsn=8

McCowan, J. R. (2015). Black, White, and Obama: A phenomenological examination of the impact of Barack Obama’s 2008 Presidential Election on the lives of biracial Americans [Unpublished doctoral dissertation]. The Chicago School of Professional Psychology.

Merolla, J. L., Sellers, A. H., & Fowler, D. J. (2013). Descriptive representation, political efficacy, and African Americans in the 2008 presidential election. Political Psychology, 34(6), 863–875. doi:10.1111/j.1467-9221.2012.00934.x

Mortenson, T. G. (2000). Poverty, race, and the failure of public policy: The crisis of access in higher education. Academe, 86(6), 38–43.

JOURNAL OF COLLEGE STUDENT PSYCHOTHERAPY 17

Olivas, M., & Li, C. S. (2006). Understanding stressors of international students in higher education: What college counselors and personnel need to know. Journal of Instructional Psychology, 33(3), 217–222.

Patterson, M. M., Pahlke, E., & Bigler, R. S. (2013). Witnesses to history: Children’s views of race and the 2008 United States presidential election. Analyses Of Social Issues And Public Policy (ASAP), 13(1), 186–210. doi:10.1111/j.1530-2415.2012.01303.x

Pendse, A., & Inman, A. G. (2017). International student-focused counseling research: A 34-year content analysis. Counselling Psychology Quarterly, 30(1), 20–47. doi:10.1080/ 09515070.2015.1128395

Pew Research Center. (2014, June 12). Political polarization in the American public. Retrieved from http://www.people-press.org/2014/06/12/political-polarization-in-the-american- public/

Pieterse, A. L., Carter, R. T., Evans, S. A., & Walter, R. A. (2010). An exploratory examination of the associations among racial and ethnic discrimination, racial climate, and trauma-related symptoms in a college student population. Journal of Counseling Psychology, 57(3), 255–263. doi:10.1037/a0020040

Schmidt, K., & Axt, J. R. (2016). Implicit and explicit attitudes toward African Americans and Barack Obama did not substantively change during Obama’s presidency. Social Cognition, 34(6), 559–588. doi:10.1521/soco.2016.34.6.559

Sifferling, A. (2016, November 1). What women are telling their therapists about election stress Time. Retrieved from http://time.com/4544738/what-women-are-telling-their-therapists- about-election-stress/

Smith, T. B., & Trimble, J. E. (2016). Foundations of multicultural psychology: Research to inform effective practice. Washington, DC: American Psychological Association.

Southern Poverty Law Center. (2016). The Trump effect: The impact of the 2016 presidential election on our nation’s schools. Retrieved from https://www.splcenter.org/20161128/trump- effect-impact-2016-presidential-election-our-nations-schools

Umphress, V. J., Lambert, M. J., Smart, D. W., Barlow, S. H., & Clouse, G. (1997) Concurrent and construct validity of the Outcome Questionnaire. Journal of Psychoeducational Assessment, 15, 40–55. doi: 10.1177/073428299701500104

Valentino, N. A., Brader, T., Groenendyk, E. W., Gregorowicz, K., & Hutchings, V. L. (2011). Election night’s alright for fighting: The role of emotions in political participation. The Journal of Politics, 73(1), 156–170. doi:10.1017/s0022381610000939

Villavicencio, K. C. (2018). Migrants in the biosphere: Migrants, water, and the politics of pain [Unpublished doctoral dissertation]. Yale University.

Williams, D. R., & Medlock, M. M. (2017). Health effects of dramatic societal events – Ramification of the recent presidential election. New England Journal of Medicine, 376 (23), 2295–2299. doi:10.1056/NEJMms1702111

18 B. M. MERRILL ET AL.

Copyright of Journal of College Student Psychotherapy is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.