Mental Health as a Predictor of Student Retention

1
Whats New in the Study of Mental Health and
Student Retention?
John Achter, Ph.D. AUCCCD Fall 2008
2
Collaborators

• Co-Investigators
• Kristina Gorbatenko-Roth, Ph.D. Dept. of
  Psychology
• Jenna Maas, M.A. Applied Psychology masters
  program
• Assistants
• Bob Spencer, Jenna Simon, Abby Laib, EmmaLee
  Ericksen, Katie Hosley, and Sara Grzelak
  Applied Psychology program
• Lindsey Grush and Mark Mittag Mental Health
  Counseling program

3
Study 1 - Introduction

• The relationship between mental health and
  student retention has been understudied, despite
  a strong focus on uncovering predictors of
  retention in higher education contexts
• Existing research among counseling center samples
  generally shows a retention advantage for
  students receiving counseling for personal
  problems compared to non-counseled students
• Limited research exists examining the effect of
  mental health issues on retention among a general
  student population

4
Study 1 - Methods

• Questions
• Does mental health status predict retention rates
  of college students?
• For those with mental health needs, does seeking
  mental health care predict retention?
• Sample
• 2201 undergraduate students at midsized
  Midwestern public university
• 57 female 93 Caucasian equal class
  representation
• Representative in terms of ACT, HS Rank, College
  GPA

5
Study 1 - Methods

• Procedures
• Step 1 Assessed undergraduate mental health
  needs and treatment seeking via electronic health
  survey to all students (39 response rate)
  Spring 2006
• MH Need Status As a college student, have you
  ever needed healthcare services for mental health
  concerns (e.g. stress, depression, relationship
  issues)?
• MH Care Seeking Status Did you ever actually
  seek care anywhere for mental health concerns?
• Step 2 Assessed retention status in Spring 2007
  (1 year post-survey)

6
Study 1 Results

• Descriptives
• Experienced a Mental Health Need During College
• 16.8 (N369) of total sample (N2201)
• Sought Mental Health Care Anywhere
• 62.3 (N230) of people with need sought care
• 37.1 (N137) of people with need did NOT seek
  care
• 0.5 (N2) no response to item query

7
Study 1 Results
Retention Comparisons
8
Study 1 -Results
Significant Bivariate Relationships (Test of
Independence (?2)
CF Cramers V Black Negative Relationship
(positive on condition less likely to be
retained) Red Freshmen Seniors less likely to
be retained then Sophomores and Juniors
9
Study 1 - Results

• Binary Logistic Regression Full Sample (n1402)
• Block 1 Academic demographic predictors
• Gender, Minority status, Low income status, 1st
  Generation status, Year in College, ACT, High
  School Rank, Cumulative GPA
• Nagelkerle R2 .199
• Block 2 Mental health status during college
• Nagelkerle R2 .199
• Mental Health need did not account for additional
  variance
• Significant predictors (p .02) and direction of
  relationship
• Class year (mixed), Cum. GPA (), high school
  ile rank (), 1st gen. status (-)

10
Study 1 - Results

• Binary Logistic Regression Mental Health Sample
  (n219)
• Block 1 Academic demographic predictors
• Nagelkerle R2 .272
• Block 2 Mental health care-seeking during
  college
• Nagelkerle R2 .321
• Seeking MH treatment accounts for 5 more
  retention variance and increases ID of
  non-retention by 10.8 (full model18.9 reduced
  model8.1)
• Significant predictors (p .02) and direction of
  relationship
• Class year (mixed), Care seeking (-), 1st gen.
  status (-), low inc. status ()

11
Study 1 - Conclusions

• Mental health (MH) need related to lower
  retention in bi-variate but not multivariate
  analyses (possible co-variation effect)
• For those with a MH need, seeking care was
  predictive of retention, above and beyond
  traditional predictors
• 2nd strongest predictor after year in college
• Unexpectedly, seeking care was related to lower
  retention.
• Current study is 1st known to look at a
  population (vs. treatment) sample and to control
  for other variables related to retention
• Hypothesis seeking MH care is a proxy for MH
  severity
• More research is needed to tease out the complex
  relationship between mental health needs,
  treatment, college persistence.

12
Study 1 - Limitations

• Self-report data
• No measure of type or severity of mental health
  issue
• No data on whether treatment was actually
  received, nor on type or amount of treatment
• No information available regarding reasons
  students were not retained

13
Future Research

• Seek replication of findings among other general
  university populations
• Examine several types of mental health related
  data, not just self-report data
• Examine whether type or amount of treatment
  impacts retention
• Examine whether severity of mental health need
  relates to retention

14
Study 2 - Methods

• Questions
• Among counseling center clients, does severity of
  distress predict retention?
• Among counseling center clients, does treatment
  length predict retention?
• Sample
• 757 undergraduate students at midsized Midwestern
  public university, seeking counseling from
  2003-2007
• No career counseling no mandated alcohol clients
• 69 female 94 Caucasian Class year FR-30
  SO-22 JR-22 SR-26
• Representative in terms of ACT, HS Rank, College
  GPA

15
Study 2 - Methods

• Measures
• Retention Students were considered retained if
  they earned credits or graduated within one year
  after a treatment episode
• Episodes defined as specific periods of
  treatment need for each subject
• 865 treatment episodes 80 had 1 16 had 2
• of sessions ranged from 1-51 Mean5.26
  Median3
• Severity Subjects maximum score during episode
  on the Outcome Questionnaire 45.2 (OQ-45.2)
• Mean75 SD25

16
Study 2 Results

• Retention Rates by Class Status

17
Study 2 - Results

• Retention Significant bi-variate relationships
• Year in College ?2 (3) 58.513, p.00
• Treatment Sessions r .109, p.001
• Severity of Mental Health Need r -.08 (p.02)
• Note Severity of need and of sessions also
  correlated r .153, p.001. This suggests a
  possible interaction and the need to look closer
  at relationship between these variables and how
  they impact retention.

18
Study 2 - Results

• Binary Logistic Regression
• Block 1 Academic demographic predictors
• Age, Gender, Low income status, 1st Generation
  status, Year in College, ACT, High School Rank
• Nagelkerle R2 .142
• Block 2 Severity and treatment length
• Nagelkerle R2 .183
• Severity and Treatment Length accounted for 4
  more variance and increased ID of non-retention
  by 11.8 (full model13.6 reduced model1.8)
• Significant predictors (p .04) and direction of
  relationship
• Freshman status (-), OQ score (-), of sessions
  ()

19
Study 2 - Conclusions

• Results were in expected directions
• As severity of distress increased, the likelihood
  of being retained decreased
• As treatment length increased, the likelihood of
  being retained increased
• Treatment length and severity of distress
  predicted retention above and beyond academic and
  demographic factors
• However. . .
• Retention rates among counseling center clients
  is lower than in the general student population

20
Study 2 - Limitations

• Self-report data
• No information available regarding reasons
  students were not retained
• Did not look at reasons for seeking counseling

21
Future Research

• Seek replication of findings among other
  university counseling center populations
  include effect sizes
• Examine several types of mental health related
  data
• Look at interactions between severity and
  treatment length variables
• Look at changes in distress over time
• Control for other variables known to impact
  retention

22
Implications

• I still cant explain study 1 results!
• Could avoidance be functional?
• Are treatment-seeking students more severe? Did
  they really seek help?
• It may not be that counseling centers can boast
  higher retention rates than the general student
  populationat least not for those presenting with
  mental health needs
• Nonetheless, studies like Wilson et. al (1997),
  and the present study, show that for this at-risk
  population, students who need treatment are more
  likely to be retained if they get treatment than
  if they dont
• The more severe the distress, the more
  intervention may be needed. Are our centers
  equipped for this?

23
Implications

• Do we have sufficient resources to reach out and
  identify the most at-risk students, and to
  provide adequate services once they seek help?
• Your thoughts?