Psychological interventions to improve sleep in college students: A meta‐analysis of randomized controlled trials

Sleep disturbances and insomnia are common in college students, and reduce their quality of life and academic performance. The aim of this meta‐analysis was to evaluate the efficacy of psychological interventions aimed at improving sleep in college students. A meta‐analysis was conducted with 10 randomized controlled trials with passive control conditions (N = 2,408). The overall mean effect size (Hedges’ g) of all sleep‐related outcomes within each trial was moderate to large (g = 0.61; 95% confidence interval: 0.41−0.81; numbers‐needed‐to‐treat = 3). Effect sizes for global measures of sleep disturbances were g = 0.79; 95% confidence interval: 0.52−1.06; and for sleep‐onset latency g = 0.65; 95% confidence interval: 0.36−0.94. The follow‐up analyses revealed an effect size of g = 0.56; 95% confidence interval: 0.45−0.66 for the combined sleep‐related outcomes based on three studies. No significant covariates were identified. These results should be interpreted cautiously due to an overall substantial risk of bias, and in particular with regard to blinding of participants and personnel. Nevertheless, they provide evidence that psychological interventions for improving sleep are efficacious among college students. Further research should explore long‐term effects and potential moderators of treatment efficacy in college students.

disturbances include cases in which symptoms of insomnia do not reach clinical significance, as well as cases in which there is not enough information present to diagnose insomnia.
College students in particular are at risk for experiencing impaired sleep, in large part due to their high levels of perceived stress; likewise, an epidemiological study found that 60% of college students surveyed reported experiencing sleep disturbances (Lund, Reider, Whiting, & Prichard, 2010). The prevalence of clinical insomnia in college students reaches nearly 10% (Schlarb, Kulessa, & Gulewitsch, 2012;Taylor et al., 2011). As emerging adults trying to earn academic degrees, young college students face numerous new challenges, such as establishing identity and future life goals, reaching momentous life decisions, living independently, handling finances, managing academic demands, and modifying existing and adopting new social roles (Arnett, 2000;Brougham, Zail, Mendoza, & Miller, 2009;Pierceall & Keim, 2007). In contrast to other young adults, college students face a less prestructured everyday life due to self-organized schedules. In this respect, evidence shows that college students often shift to an irregular sleep−wake cycle by short sleep length on weekdays and later wake-up times on weekends (Buboltz, Brown, & Soper, 2001;Machado, Varelle, & Andrade, 1998) due to, for example, long learning hours, parties and noise pollution in student residences.
Given this transition phase in life, there are several reasons to assume that the efficacy of psychological interventions to improve sleep might differ between college students and the general population. Young adults (ages 19-29 years) show especially high strain in their everyday lives due to sleep disturbances and the worst sleep hygiene (SH) behaviour (Rosenberg et al., 2011). More severe sleep disturbances have been shown to be a predictor for larger sleep improvements within the scope of psychological intervention (Espie, Inglis, & Harvey, 2001;van Houdenhove, Buyse, Gabriëls, & van den Bergh, 2011;Murwaski, Wade, Plotnikoff, Lubans, & Duncan, 2018).
Furthermore, psychological interventions aimed at changing beliefs and behaviour patterns might be more effective for college students than for other populations. Because cognitive flexibility is correlated with intelligence (Colzato, van Wouwe, Lavender, & Hommel, 2006)which in turn is fostered by education (Ritchie & Tucker-Drob, 2018) -and decreases with increasing age (Peltz, Gratton, & Fabiani, 2011), college students may have above average cognitive flexibility. This increased cognitive flexibility may lead to improved response to psychological intervention. Finally, research has found increasing evidence that early intervention predicts better treatment outcomes for a variety of psychological disorders (McGorry, Purcell, Goldstone, & Amminger, 2011). Because sleep disturbances and insomnia often begin during the college years, this is an essential time for intervention (Lund et al., 2010).
People suffering from insomnia show a two-to threefold (odds ratio [OR] = 1.98−2.98) increased risk for depression, a significantly greater risk for anxiety disorders (OR = 1.63−2.64), and a significantly greater risk for substance abuse disorders, including alcohol abuse (Baglioni et al., 2011;Pigeon et al., 2017;Sivertsen et al., 2014).
Cognitive behaviour therapy for insomnia (CBT-I), including stimulus control, sleep restriction, progressive muscle relaxation, SH and cognitive therapy, meets the American Psychological Association's criteria for empirically validated psychological interventions for insomnia (Morgenthaler et al., 2006;Morin, Bootzin, et al., 2006). CBT-I is recommended for the general population as first-line treatment for insomnia (Qaseem, Kansagara, Forciea, Cooke, & Denberg, 2016;Riemann et al., 2017), as it has been shown to be effective in improving sleep quality (SQ) and quantity in adult samples (van Straten et al., 2018).
Despite this, little is known about the efficacy of psychological interventions among college students, as there are no meta-analyses available referencing this specific population. Thus, investigating treatment response in college students is of particular importance due to their increased vulnerability to sleep disturbances and the negative health outcomes they experience as a result. Given that emerging adulthood -including college years -represents a time of major life transition and the highest risk of developing mental disorders, reaching emerging adults through interventions is of paramount importance Kessler, Berglund, et al., • Self-help interventions are as effective as F2F interventions.
• The effects persist after the end of treatment.
2005; Kessler, Berglund, et al., 2005). College students experiencing a variety of stressors show an increased onset of mental health problems (Pedrelli, Nyer, Yeung, Zulauf, & Wilens, 2015;Thurber & Walton, 2012), with over 20% of all college students suffering from a mental disorder . This high prevalence depicts the elevated vulnerability for mental disorders at a time of major life transitions, that also influences further life of college students considerably, as this critical time period is substantial for basic life events as, for example, attaining educational levels/degrees Bruffaerts et al., 2018). The number of affected college students far exceeds the resources of most treatment options at universities, resulting in substantial treatment gaps of mental health issues among college students Beiter et al., 2015).
Further, college students are a key group in society in terms of human capital (Abel & Deitz, 2012), driving future societal economic growth and innovation. Higher education institutions provide an opportunity to become a key setting for the prevention and early treatment of sleep disturbances and insomnia, with a high reachability of over 50% of emerging adults being enrolled in higher education (Aud et al., 2011;Reavley & Jorm, 2010). Beyond that, because sleep disturbances are a common complaint that lack the stigma associated with other mental health issues, psychological interventions aimed at improving sleep may provide an acceptable low-threshold opportunity for a first step in a care pathway that have the potential to positively affect other co-morbid mental and physical health issues (Freeman et al., 2017;Friedrich & Schlarb, 2017;Wu, Appleman, Salazar, & Ong, 2015).
To the best of our knowledge, there are no meta-analyses available so far, focusing on psychological sleep treatments in college students. To date, two descriptive reviews have synthesized findings on psychological interventions to overcome sleep disturbances in college students. One of these reviews (Dietrich, Francis-Jimenez, Knibbs, Umali, & Truglio-Londrigan, 2016) focuses on education programmes to improve sleep. SH education in varying formats was identified as the main intervention form in the articles included in the review. The interventions on SH education differed in their format with regard to the integration of additional content, delivery strategies, length and duration of the programme. Studies were included that meet the primary outcome measures of SH knowledge, SH behaviour or SQ. However, because such programmes are a poorly explored subsection of psychological sleep interventions (at least among college students), this review includes only three randomized controlled trials (RCTs) and one quasi-experimental study. Significant effects of sleep education on SH knowledge and behaviour, respectively, were found in one of two studies, and effects on SQ were found in one of four studies. Thus, this review only provides preliminary evidence, and does not allow to draw conclusions on the overall effectiveness of sleep education interventions in college students.
Furthermore, the review focuses on sleep education only, without targeting and analysing all types of psychological sleep interventions. Friedrich and Schlarb (2017) examined 27 evaluation studies on sleep treatments that included psychological components. The treatments were categorized in SH interventions, CBT-I, relaxation and other treatments (e.g. gestalt therapy or imagery rehearsal therapy). They found small to moderate effect sizes for SH interventions (d = 0.32-0.61), large effects for CBT-I (d = 1.06-1.77), and moderate effects for other psychological interventions (d = 0.45-0.61). Fifteen out of the 27 studies were RCTs, seven studies used a controlled design without randomization, and five trials had no control condition.
Because it was the first review on the topic of sleep disturbances in college students, rather broad eligibility criteria were chosen that focused more on sensitivity than specificity (Friedrich & Schlarb, 2017).
Generalizability and interpretability of these two reviews are limited due to the small number of included studies, and the inclusion of studies that were not RCTs. The authors could, therefore, not calculate pooled effect sizes using meta-analytic techniques. Because several RCTs on this issue have recently been published, we performed a meta-analysis to evaluate the efficacy of psychological interventions aimed at improving sleep among college students compared with control conditions. Thus, the aim of the current meta-analysis is to deliver quantitative information on the effectiveness of sleep interventions in college students including RCTs only, focusing on psychological treatment with reported outcomes that are related to sleep.

| MATERIAL S AND ME THODS
This study was carried out as part of the WHO World Mental Health

| Study selection
Publications were identified by searching three major electronic databases, starting from database inception on 28 September 2017 up to 20 March 2018: the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and PsycINFO. As part of the WMH-ICS, the search string was compiled on a superordinate level to provide a foundation for investigating a variety of psychological interventions in college students. Thus, the search string did not contain terms that would restrict the search to the disorders or delivery modes targeted in this analysis; it accepted a high number of references for screening in order to minimize the risk of missing relevant studies. There were no restrictions on publication date or status. The search was based on a string combining terms text words, index, and free terms indicative of RCTs that evaluated psychological interventions in tertiary education settings (see Appendix A). From there, references in identified studies and previous systematic reviews of overlapping topics were checked for earlier publications. The WHO International Clinical Trials Registry Platform was also searched for unpublished trials as a way to identify grey literature. Authors of study protocols without published results were contacted to determine the eligibility of unpublished data for the meta-analysis.
We screened the titles and abstracts of all articles for overall fit for this analysis. Full texts of selected articles were then retrieved and independently assessed for eligibility. Both steps were performed independently by two researchers (KSa, TB). Discussion between researchers was initiated in case of assessor disagreement; two senior researchers (DEE, HB) were consulted when disagreement could not be resolved. Consistent with Cuijpers, van Straten, Warmerdam, and Andersson (2008), psychological interventions were defined as: interventions in which verbal communication between a therapist and a client was the core element; or in which a systematic psychological method was written down in book format or on a website (bibliotherapy), while the client worked through it more or less independently.

| Inclusion criteria
Psychological interventions consisting of educational elements, cognitive methods, behavioural methods or other CBT-I-related techniques delivered face-to-face (F2F), written, by phone, tape or in computerized form were included. We elected to include all manner of psychological interventions currently practiced in order to provide a representative and overarching depiction of practiced psychological interventions. A subgroup analysis that focused on CBT-I exclusively was also conducted.
Thus, non-psychological interventions like medication, manipulation of light and sound, physical exercise or surgery were excluded.
Studies exploring non-specific and hybrid interventions were also excluded, as they did not focus primarily on sleep. Furthermore, studies containing any participants who were not enrolled in higher education institutions were excluded, even if the majority of the sample was. Finally, case studies, cross-sectional studies, non-randomized trials, non-controlled trials, and trials with control groups containing interventions or other active manipulation such as placebo control conditions were also excluded. Our main research question was to analyse non-specific treatment effects of sleep interventions in college students as there was, to date, no evidence whether psychological interventions to treat sleep disturbances were also effective in the vulnerable group of college students (Ohayon & Reynolds, 2009). Thus, active control conditions were excluded as we did not intend to analyse specific effects of treatment in superiority or non-inferiority trials.

| Data extraction and classification
The following data were extracted for each article, if reported or applicable: (a) bibliographical data (first author, year of publication); (b) sample characteristics (sample size, gender and age distribution, dropout rate, university course[s] of the participants, study subject, compensation for participation, country); (c) participation criteria (inclusion and exclusion criteria)l; (d) intervention characteristics (mode of delivery, frequency and number of contacts, duration of intervention, therapeutic content, type of control group); (e) outcome measures (time points of assessments, outcomes of interest). In addition to sleeprelated outcomes, we were also interested in measures of academic performance due to their reported relationship to sleep disturbances.
If relevant information to examine the eligibility of a study was not available, authors were contacted a maximum of two times to attain or clarify information. If the authors did not respond, and the information given in the publication was insufficient to perform a meta-analysis, the article was excluded. Data extraction was conducted independently by two researchers (KSa, TB).

| Quality assessment
Study validity was evaluated by two researchers (KSa, TB). The risk of bias assessment was carried out according to the updated method guidelines for systematic reviews with RCTs of the Cochrane Collaboration (Furlan, Pennick, Bombardier, & van Tulder, 2009).
We examined bias due to: (a) selection (sequence generation, alloca- (2009), we reported: (e) researcher allegiance, as it has been shown to be associated to outcome (Munder, Brütsch, Leonhart, Gerger, & Barth, 2013); and (f) whether the sample was a convenience sample chosen due to reachability. The risk of researcher allegiance was considered high if at least one of the authors developed the intervention and: (a) delivered it himself; or (b) did not deliver it himself but trained or supervised those that did (Gaffan, Tsaousis, & Kemp-Wheeler, 1995;Leykin & DeRubeis, 2009). The studies' risk of bias was considered low if: (a) at least six of Furlan et al.'s (2009) 12 suggested risk of bias criteria were rated low; and (b) the risk of researcher allegiance was rated low.

| Outcome measures
The included studies reported a number of different outcome measures and used different measurement methods (primarily questionnaires or daily sleep diaries, but also actigraphy and elec-  outcomes reported in the studies and their frequency can be seen in Table 1.
The majority of studies reported more than one outcome assessing sleep, and a primary outcome was not defined. Because of this, we used combined effect sizes for each study. Thus, effect sizes contain effects of the psychological intervention on all obtained sleep-related facets.
Effect sizes were calculated by standardizing all sleep-related outcomes within a study and combining them to one effect size. When calculating a combined effect size within a single study, the intercorrelations between included outcomes need to be considered. We assumed a conservative intercorrelation of r = 1. To account for the potential distortion of effect size created by this assumption, we added a sensitivity analysis with an intercorrelation estimation of r = 0 (Becker, 2000).

| Power calculation
An a priori power analysis (Borenstein, Hedges, Higgins, & Rothstein, 2010) indicated that at least 15 studies with at least 20 participants per condition, respectively, six studies with at least 50 participants per condition, would be needed to detect an effect of g = 0.3 (with moderate heterogeneity) with a statistical power of 0.8 and a significance level of α = .05.

| Effect size calculations
An effect size for each sleep-related outcome in each study was calculated. These effect sizes compared post-treatment values of intervention groups and control groups. As many included studies had small sample sizes, bias was corrected for by calculating Hedges' g as an effect size metric (Hedges & Olkin, 2014). The effect sizes were calculated using reported means and standard deviations or t-values (Lipsey & Wilson, 2010). Effect sizes below g = 0.32 were rated as small, between g = 0.33 and 0.55 as moderate, and above 0.55 as large (Lipsey & Wilson, 1993). To facilitate clinical interpretation of standardized mean difference (Hedges' g), we calculated numbers-needed-to-treat (NNT) to generate one additional clinically significant change using the formulae of Kraemer and Kupfer (2006). We applied the random effects model for pooled effect sizes (Borenstein et al., 2010

| Subgroup analyses
Subgroup analyses were based on the mixed effects model; a fixed effects model was used to test differences across subgroups, while a random effects model was used within subgroups (Borenstein et al., 2010).

| Specific sleep constructs
In addition to the calculation of the overall effect size, we analysed the effect of the interventions on specific sleep constructs. Indicator with eight items , and the sleep diary item "sleep quality"). To reduce the risk of over-representing single aspects of sleep disturbances, sleep diary items focusing on a single aspect (e.g. "number of awakenings") were excluded. The construct "fatigue and daytime functionality" includes all questionnaires assessing fatigue and daytime functionality in connection to SQ, sleep continuity or sleep duration. The construct "pre-sleep behaviour and experiences" includes all questionnaires assessing SH or attitudes towards sleep.
The construct "SOL" includes all measures assessing the duration of time spent trying to fall asleep to actually falling asleep; this includes measures derived from the sleep diary, actigraphy and EEG (Table 1).

| Follow-up analysis
The effect sizes for available follow-up measurements were calculated by comparing the treatment conditions with the control conditions. Follow-up measurements include all measurements declared by the authors of the studies as such and timed after the end of intervention and post-test measurement.

| Selection of studies
We screened a total of 12,206 studies for titles and abstracts (search string n = 11,936; secondary search strategies n = 270). After this screening, 23 studies of the main search string and 12 studies of the additional secondary search strategies remained to be assessed on the full text level. The study selection process is described in a PRISMA flowchart (Figure 1). Finally, 10 studies were included in the meta-analysis. The inter-rater reliability concerning study selection was very good (κ = 0.83).

| Missing data
With regard to missing data, 14 authors were asked via email to provide the full text. Five of the missing studies were conference abstracts and thus not able to be included. Four studies could be identified as unpublished master theses with no access given. Three authors did not provide full text articles needed, and thus respective studies had to be excluded. Two of the studies received were excluded because they: (a) examined a treatment focusing on headache with sleep disturbances as a co-morbidity; and (b) investigated the effectiveness of booster sessions after the initial treatment only.

| Study characteristics
Detailed study characteristics can be found in Table 2.
The 10 studies included 2,408 college students in total (treatment conditions: n = 1,002; control conditions n = 1,406). In the seven studies reporting gender, the majority of participants were female (70.1%).
Five studies included only psychology students, while the other five included students from a variety of disciplines. In five of the studies, participants did not receive any incentive for study participation.
Conversely, of the remaining five studies, two offered course credit, two offered monetary compensation, and the final study offered both. Four studies compared CBT-I with control groups, and one of these studies added music therapy to CBT-I. Freeman et al. (2018) used the unguided web-based CBT-I intervention Sleepio, which is based upon a validated manual (Espie et al., 2001(Espie et al., , 2007(Espie et al., , 2008  Five studies used standalone behavioural interventions: relaxation (k = 3), stimulus control (k = 1) and exposition (k = 1). The three relaxation interventions contained either group sessions (Borkovec & Weerts, 1976;Steinmark & Borkovec, 1974)  that trained participants in PMR and instructed them to complete PMR twice daily, including once prior to sleep. The stimulus control intervention (Zwart & Lisman, 1979) required participants to follow stimulus control rules, such as getting up after 10 min of not being able to fall or return to sleep. In the exposition intervention (Carrera & Elenewski, 1980), participants were asked to visualize lying in bed and having difficulties falling asleep. Then, they were encouraged to attend to bodily sensations and interpret them as a strange disease. Finally, they were instructed to imagine their own death.
One study used a sleep education programme (Barber & Cucalon, 2017) in which participants were informed about general SH information and technology boundary management via PowerPoint presentation.
Five of the psychological interventions were self-help programmes, of which four were delivered electronically and two were preceded by F2F training sessions. The other psychological interventions were delivered F2F as group therapy (k = 3) and individual therapy (k = 2). The group size ranged from two to seven. Considering all 10 studies, the number of contacts ranged from one to 11, and averaged 4.7 contacts (including e-mail contacts). The post-test measurements were conducted on average 4.5 weeks after the baseline measurements (range 1-10 weeks). This duration also reflects the intervention duration.
To measure sleep outcomes, sleep diaries (k = 6), objective measures such as actigraphy and overnight EEG (k = 3), and questionnaires (k = 6) were used. No study evaluated effects on academic functioning. Inter-rater agreement concerning study characteristics was very good (κ = 0.87).

| Quality assessments
The complete assessment for each study is presented in Table 3.
Overall risk of bias was considerable. Of the 10 studies, five showed a high risk of bias. On average 7.5 of 14 assessments per study were rated unclear or high. Six studies did not report information on sequence generation and allocation concealment. Two studies did not report on demographic and baseline differences between conditions. One study had a high risk of bias in allocation concealment. All studies revealed an unclear or high risk of bias in at least two of the three blinding situations. Five of the studies revealed low risk of bias in the domain of incomplete outcome data.
Concerning other threats to validity, most criteria showed on average low risk of bias. Only selective reporting showed on average a higher risk of bias, with nine studies being rated unclear and only one study being rated low. Three studies had an unclear or high risk of compliance differences between the treatment conditions, and three showed a high risk of researcher allegiance. Inter-rater reliability regarding risk of bias assessments was good (κ = 0.76).

TA B L E 2 Selected characteristics of included RCTs (n = 10)
Author (

| Overall effect of psychological intervention
The overall effect size of combined outcomes of the 10 studies was moderate to large (g = 0.61; 95% confidence interval

| Follow-up
Only three studies reported follow-up results. The studies gathered follow-up measures 1 month, 3 months or 12 months, respectively, after the end of psychological intervention. The combined follow-up effect size was g = 0.56 (95% CI: 0.45−0.66) with low heterogeneity of I 2 = 0 (95% CI: 0−95).

| Subgroup analyses
Subgroup analyses yielded no significant differences between subgroups, indicating that mode of delivery, risk of bias and type of treatment did not affect the effect size (see Appendix B).

| Specific sleep constructs
For a summary of the effect sizes of specific sleep constructs, see Table 4.

| Global measures of sleep disturbances
Six to high heterogeneity (I 2 = 65; 95% CI: 16−85; see Appendix C). We carried out the same subgroup analyses as in the sensitivity analyses for the overall effect size.

| Publication bias
Both Egger's tests and visual inspection of funnel plots did not show any significant publication bias for all but one analysis. Only the visual inspection of the funnel plot for the effect size distribution of SOL suggested some publication bias. According to Duval and Tweedie's trim and fill procedure, three studies were imputed, resulting in a smaller but still significant effect size of g = 0.46.

| DISCUSSION
This meta-analysis suggests that psychological intervention may reduce sleep disturbances in college students, as determined by global measures of sleep disturbances and SOL. Overall risk of bias was substantial, but no indication of larger effects through distortion could be found relating to the risk of bias of a study. Indication for publication bias was only found for the construct SOL, in which adjusting for potential unpublished studies resulted in change. Thus, the present analysis implies psychological intervention to be an appropriate way to help the large number of college students with sleep disturbances.
Our results are in line with previous research among college students. In their review, Friedrich and Schlarb (2017)  for ISI; all adults: g = 0.42 for SOL, g = 0.55 for SE, g = 0.73 for ISI).
Assuming this difference in effect sizes is due to an actual increased benefit for college students compared with older adults (rather than representing methodological or other biases), it might be explained by two factors. First, young adults experience especially high strain from sleep disturbances, and show the worst SH compared with older adults (Rosenberg et al., 2011). Thus, college students (who are, in general, young adults) may have on average more room for improvement regarding their sleep. This assumption is similar to results found in research on depression, showing that higher pre-intervention symptom severity is associated with higher efficacy of psychological intervention (Driessen, Cuijpers, Hollon, & Dekker, 2010). In accordance with these findings, the study by van Straten et al. (2018)  Second, as outlined above, college students are likely to show above average cognitive flexibility (Colzato et al., 2006), resulting in a higher capacity to change their sleep behaviour and thus a potentially heightened benefit of psychological intervention. Future research is needed to confirm such assumptions.
The present meta-analysis did not find efficacy differences between different types of psychological interventions. This is inconsistent with the suggestion of Friedrich and Schlarb (2017) that CBT-I is the most effective psychological intervention, and SH interventions the least effective, a specification supported by research in the general population (Morin, Culbert, & Schwartz, 1994;Morin et al., 1999;Morin, LeBlanc, et al., 2006;Qaseem et al., 2016). It is noteworthy, though, that the latest and most comprehensive meta-analysis on psychological interventions for sleep disturbances and insomnia in the general population also failed to find this specification. The authors suggested that this may be due to variability within each type of psychological intervention, or a potential correlation of psychological intervention type, study quality and variation in effect size with time of publication; this would result in smaller effects for the types of psychological interventions investigated more recently with higher study quality (van Straten et al., 2018). This explanation might also apply to the results of the present meta-analysis. It is also noteworthy that the small number of trials in our meta-analysis might have reduced power to detect true differences in efficacy between different types of psychological intervention. However, effects of specific psychological intervention techniques could also be overestimated, and the reduction of sleep disturbances might be evoked by generic mechanisms of change common to all types of psychological intervention like therapeutic alliance, empathy, goal consensus and collaboration.
Although a previous study found F2F psychological intervention to be more effective in improving sleep than self-help psychological intervention in the general population (Lancee, van  The current meta-analysis has some limitations. First, the number of included studies was small. However, it is not only the number of studies that makes a meta-analysis reliable, but also the number of participants per study (Borenstein et al., 2010). The 10 studies that were included in this meta-analysis had an average sample size of more than 100 participants per condition, which is sufficient to reliably detect effect sizes of 0.3 and above, even given a small number of trials with high heterogeneity (Borenstein et al., 2010).  (Breslau, Roth, Rosenthal, & Andreski, 1996;Daley et al., 2009;Kessler, Berglund, et al., 2005;Kessler, Chiu, et al., 2005;Sivertsen et al., 2014). Above that, such interventions may improve the cognitive performance of college students, who represent a large portion of a society's human capital (Gomes et al., 2011 This in turn reduces the economic cost associated with these disorders and conditions. Above that, such interventions may improve the cognitive performance of college students, who represent a large portion of a given society's human capital. The task of providing psychological treatment to college students includes not only establishing sufficient provision of psychological interventions, it also includes raising awareness of intervention opportunities and providing psychological interventions that are suited to the needs of college students in particular.