The University of North Carolina at Chapel Hill School of Social Work SOWO 916 Structural Equation Modeling Spring Semester, 2015 INSTRUCTOR Natasha K. Bowen, Ph.D. Room 245E, Tate Turner Kuralt Building CB #3550, School of Social Work, Chapel Hill, NC 27599-3550 Phone: (919) 843-2434 Email: nbowen@email.unc.edu CLASS MEETING TIMES & OFFICE HOURS Class meets on Wednesdays 9:00-11:50 am in TTK Room 101 Office hours are Wednesdays 12:00 – 1:00 or by appointment This course was originally developed by Dr. Shenyang Guo. COURSE DESCRIPTION Structural equation modeling (SEM) is a general statistical method that can be employed to test theoretically derived models. It is “a class of methodologies that seeks to represent hypotheses about the means, variances, and covariances of observed data in terms of a smaller number of ‘structural’ parameters defined by a hypothesized underlying model” (Kaplan, 2000). In this course, students will learn fundamental concepts and skills to conduct SEM, and how to apply these techniques to social work research. COURSE OBJECTIVES At the completion of the course, students will be able to: Understand the fundamental hypothesis of SEM and its relationship to the specification, identification, and estimation of a structural equation model; Run path analysis and test mediating hypotheses using SEM; Conduct confirmatory factor analysis to evaluate measurement validity; Conduct structural equations with latent variables and apply the method to test/confirm a theoretically derived model; Understand statistical indices measuring goodness-of-fit of a model; Conduct multiple group comparisons with SEM to test moderating effects; Perform power analysis with SEM and know how to determine minimum sample size needed; Understand basic concepts and skills to deal with interactions and quadratics in latent variables, and categorical variables; Understand the linkage between SEM and hierarchical linear models, and conduct multilevel analysis and latent growth curve analysis with SEM; Understand strategies for dealing with missing data. PRE-REQUIREMENT Students are assumed to be familiar with descriptive and inferential statistics. A solid understanding of multiple regression analysis is a key. They should have statistical and statistical software background at least equivalent to that provided by SOWO919 (applied regression analysis and generalized linear models), SOCI209, PSYC282, EDUC284 (linear regression), or SOCI211 (categorical data analysis). SOFTWARE PACKAGES Students may choose to use SAS, SPSS, or Stata for data management and matrix calculations. Class examples in these areas will be in Stata. The course will employ Mplus for running SEM. Mplus is available through the university’s Virtual Computer Lab. Students are expected to use the Lab’s help resources to learn how to establish connections with the Lab. TEXTBOOKS Required: Bowen, N. K., & Guo, S. (2011). Structural Equation Modeling. New York, NY: Oxford University Press. Recommended for students who will be conducting SEM’s in their futures: Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: John Wiley & Sons. Byrne, B. M. (2012). Structural equation modeling with Mplus. New York: Routledge. Hoyle, R. H. (Ed.). (2012). Handbook of structural equation modeling. New York: Guilford Press. Kline, R. (2011). Principles and Practice of Structural Equation Modeling, 3rd Ed., New York: NY: Guilford Press. ASSIGNMENTS 2 peer evaluations 5 IRATS 5 TRATS Team Application Activities 7 Homeworks Final Exam TOTAL GRADE PERCENTAGE 10% 10% 15% 30% 14% 21% 100% 2 GRADING SYSTEM The standard School of Social Work interpretation of grades and numerical scores will be used. H = 94-100 P = 80-93 L = 70-79 F = 69 and below POLICY ON CLASS ATTENDANCE Class attendance is critical for team performance and individual learning. You are expected to attend all scheduled sessions. Each class session will cover a great deal of material, and you will fall behind in the course when you miss even one class. In addition, your team will be disadvantaged by your absences. It is your responsibility to inform the instructor in advance about missing a class session. Starting after the second absence, your course grade will be reduced by 10% for each session missed. This reduction will occur after the calculation of your final grade and will be in addition to the effects absences may have on your peer evaluations. If you miss a class, you are responsible for obtaining all notes, announcements, handouts, and assignments from the missed class from your classmates. If you miss three classes, you will be expected to meet with the instructor and Associate Dean of Student Services to discuss if further participation in the course is possible. A grade of incomplete will only be given under extenuating circumstances and in accordance with University policy. POLICY ON INCOMPLETE AND LATE ASSIGNMENTS Assignments are to be turned in to the instructor at the beginning of class on the due date noted in the course outline if the class meets on a due date. If assignments are sent by email, do not assume they have been received until confirmation is received. Extensions may be granted by the professor given advance notice of at least 24 hours and extenuating circumstances. Unannounced late assignments will automatically be reduced 10%. An additional 10% grade reduction on the assignment will occur for each day late (including weekend days). A grade of incomplete will only be given under extenuating circumstances and in accordance with University policy. POLICY ON ACADEMIC DISHONESTY In-class TRATs and application activities are designed to create a collaborative learning environment. However, IRATs, homeworks, and the final exam are to be completed totally independently. Once an assignment has been handed out, it is assumed that students will not discuss any aspect of the assignment together—not even general principles or topics related to the assignment. Such exploring or sharing of knowledge that is even peripherally related to class assignments gives an unfair advantage to students who have peers from the same department or school enrolled in the course. Violations constitute violations of the honor code and will be treated as such. 3 Please include the honor code statement along with your signature on all assignments: “I have neither given nor received unauthorized aid on this assignment.” This statement indicates to the instructor that you did not consult with students or other human sources on any aspect of the assignment. It also represents your pledge that you have not made use of information or materials related to graded assignments from last year’s course. For assignments that are open book and take home (such as homeworks), consulting non-human sources (text and online sources) is encouraged. For example, you may read existing SEMNET postings, articles and Powerpoints about topics that are posted online, and book chapters, but you may not post a question related to an assignment to a discussion board. All questions about assignments and these policies should be referred to the instructor. Please refer to the APA Style Guide (6th edition), the SSW Manual, and the SSW Writing Guide for information on attribution of quotes, plagiarism and appropriate use of assistance in preparing assignments. If reason exists to believe that academic dishonesty has occurred, a referral will be made to the Office of the Student Attorney General for investigation and further action as required. POLICY ON ACCOMMODATIONS FOR STUDENTS WITH DISABILITIES Students with disabilities that affect their participation in the course and who wish to have special accommodations should contact the University’s Disabilities Services and provide documentation of their disability. Disabilities Services will notify the instructor that the student has a documented disability and may require accommodations. Students should discuss the specific accommodations they require (e.g. changes in instructional format, examination format) directly with the instructor immediately after the first class session and before the second class session. Retroactive accommodations will not be implemented. The instructor is happy to make any necessary accommodations to ensure optimal learning for every student. POLICIES ON THE USE OF ELECTRONIC DEVICES IN THE CLASSROOM The use of electronic devices for non-class related activities (e.g. checking messages, playing games) is prohibited. If any such uses occur during class time, laptops will not be allowed for any students for the remainder of the course. 4 DATES, TOPICS, AND DESCRIPTIONS OF ASSESSMENTS AND HOMEWORK ASSIGNMENTS 3 DATE ASSIGNMENT HW 1 assigned: OLS 1/7 regression, Mplus & virtual computer lab 1/14 RAT 1 HW 1 due HW 2 assigned: Direct/Ind effects 1/21 HW 2 due 4 1/28 5 2/4 6 2/11 7 2/18 8 2/25 9 3/4 1 2 DESCRIPTION Introduction to SEM and Team-based Learning Path analysis I: Specification, direct & indirect effects Path analysis II: Identification, fit, parameter estimates; Intro to matrices Matrices in SEM RAT 2 HW 3 assigned: CFA on your own HW 3 due HW 4 assigned: Justifiable modifications HW 4 due Confirmatory factor analysis I: Theory, specification, identification General structural models: Theory, specification, identification, alternative models Options for special data: Non-normal, ordinal, clustered data and data with missing values 11 3/25 RAT 3 HW 5 assigned: General SEM on your own HW 5 due HW 6 assigned: Special analyses on your own Task: identify 2 articles in your area of interest, bring to class next week HW 6 due Bring two SEM articles HW 7 assigned: Article critique RAT 4 12 4/1 HW 7 due 13 4/8 14 4/15 15 4/22 RAT 5 Longitudinal Models I: Repeated Measures Longitudinal Models II: Growth Curves Advanced Readings in SEM 10 3/18 CFA II: Alternative models, modifications, correlated errors CFA III: CFA in Scale development Critiquing SEM studies Power analysis Multiple Group Analysis Final exam Reminder: Homework assignments are due at the beginning of class and must include the student honor pledge (I have neither given nor received unauthorized assistance on this assignment) to be accepted. 5 OUTLINE OF COURSE TOPICS AND READINGS 1/7/15: Introduction to SEM and Team-based Learning (class 1) 1. Intro to TBL 2. Intro to SEM HW1 assigned Also by next week, become versatile at gaining access to Mplus 7 at the UNC virtual lab, and finding files on your computer while using the virtual lab. Mplus is now at the new virtuallab.unc.edu address. Intra- and inter-team information sharing about using the virtual lab is encouraged. The virtual lab website also has helpful information. 1/14/15: Path Analysis I: Specification, Direct and Indirect Effects (class 2) 1. Specification—Path diagrams, equations, Mplus syntax 2. Direct and indirect effects 3. Mplus: running path models 4. Examining path coefficients RAT1 HW1 due Required readings for today: Bowen & Guo, chapter 1, pp. 1-15; part of chapter 2, pp.17-34; pp. 202-203 Kline, R. (2005). Principles and practice of structural equation modeling, 2nd ed. New York, NY: Guilford Press. pp. 93-105 and 123-131, 162. Optional readings to strengthen understanding (read now or any time during semester): Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74-94. Homework 2 assigned 1/21/15: Path Analysis II: Identification, Fit, Parameter Estimates, Intro to Matrices (class 3) 1. Identification 2. Mplus: running path models 4. A first look at the evaluation of fit 5. An intro to the many matrices of SEM HW2 due Required readings: (recommended order—alphabetical) Bowen & Guo, pp. 49 – 51 in chapter 2; pp. 141-149 in chapter 6 Kline, R. (2005). Principles and practice of structural equation modeling, 2nd ed. New York, NY: Guilford Press. pp. 105-120, 131-150. 6 1/28/15: Matrices in SEM (class 4) 1. Input matrix generation from raw data 2. Input, implied, and residual matrices 3. Specification of path models with matrices 4. Matrix operations with analysis matrices Required readings (recommended order, B & G, Bollen): Bollen (1989), Appendix A: Matrix Algebra Review (also see posted guide to the Appendix) Bowen & Guo Chaper 2, pp. 34-49 Additional resources are posted: Matrix operations review by Ender; Multivariate matrices; and matrix operations in Stata by Ender (Ender’s pdf file on matrix algebra summarizes important matrix operations) Let me know if you prefer information on matrix operations in SPSS or SAS. 2/4/15: Confirmatory Factor Analysis I: Theory, Specification, Identification (class 5) 1. Latent variables 2. Classical Test Theory 3. Specification—path diagram, equations, matrices, Mplus 4. Identification 5. More on model fit RAT 2 Required readings (recommendation: read B&G, Byrne, then Bollen): Bollen, K. A. (2000). Modeling strategies: In search of the Holy Grail. Structural Equation Modeling 7, 74-81. Bowen & Guo, Chapter 4, pp. 73-108; and pp. 141-144 on χ2 in Chapter 6. Byrne, B. M. (2005). Factor analytic models: Viewing the structure of an assessment instrument from three perspectives. Journal of Personality Assessment, 85(1), 17-32. An optional (user-friendly) reading: Iacobucci, D. (2009). Everything you always wanted to know about SEM but were afraid to ask. Journal of Consumer Psychology, 19, 673-680. HW 3 assigned 2/11/15 Confirmatory Factor Analysis II: Alternative models, modifications, correlated errors (class 6) 1. Alternative models 2. Second-order CFA models 7 3. Correlated measurement errors 4. Improving and comparing models--Intro HW 3 due Required readings: Bentler P. M. & Mooijaart, A. (1989). Choice of structural model via parsimony: A rationale based on precision. Psychological Bulletin, 106, 315-317. Bowen & Guo, Chapter 6, pp. 135-166 (some parts are review) Gerbing, D. W. & Anderson, J. C. (1984). On the meaning of within-factor correlated measurement errors. Journal of Consumer Research, II, 572-580. Werner, C. & Schermelleh-Engel. (2010). Deciding between competing models:Chi-square difference tests. (Posted at website.) Retrieved in 2013 at: http://user.uni-frankfurt.de/~cswerner/sem/chisquare_diff_en.pdf HW 4 assigned 2/18/15: Confirmatory Factor Analysis III: CFA in Scale Development (class 7) 1. Theory 2. Validity 3. Reliability 4. Method effects and Multi-method multi-trait models HW 4 due Required readings: Bowen & Guo, review pp. 73-81 in chapter 4 Eid, M., Nussbeck, F. W., Geiser, C., Gollwitzer, M., Cole, D. A., & Lischetzke, T. (2008). Structural equation modeling of multitrait-multimethod data: Different models for different types of methods. Psychological Methods, 13, 230-253. Gregory, Jr., V. L. (2012). Gregory Research Beliefs Scale: Discriminant construct, concurrent criterion, and known-groups validity. Journal of Evidence-Based Social Work, 9, 465-480. Pohl, S., & Steyer, R. (2012). Modeling traits and method effects as latent variables. In S. Salzborn, Davidov, E., & Reinecke, J. (eds.). Methods, theories, and empirical applications in the social sciences. Wiesbaden, Germany: Springer. Midterm peer evaluations Midterm course feedback 2/25/15: General SEM: Theory, Specification, Identification, Estimation, Model Improvement (class 8) 1. Theory 2. Specification-- path diagram, matrices, equations 3. Identification 4. Estimation 8 5. Improving models RAT 3 Required readings: Anderson, J. C. and D. W. Gerbing (1988). Structural equation modeling in practice A review and recommended two-step approach. Psychological Bulletin 103, 411-423. Bowen & Guo, chapter 5, pp. 109-134 and review , pp. 157-166 in chapter 6 Rothbard, N. P. (2001). Enriching or depleting? The dynamics of engagement in work and family roles. Administrative Science Quarterly, 46, 655-684. Optional user friendly article: Iacobucci, D. (2010). Structural equations modeling: Fit indices, sample size, and advanced topics. Journal of Consumer Psychology, 20(1), 90-98. Homework 5 assigned 3/4/15: Options for Special Data: Non-normal, Ordinal, Clustered Data, and Data with Missing Values (class 9) HW 5 due Required Readings: Allison, P. D. (2003). Missing data techniques for structural equation modeling. Journal of Abnormal Psychology, 112, 545-557. Bowen & Guo, Chapter 3, pp. 52-72 Flora, D. B., & Curran, P. J. (2004). An empirical evaluation of alternative methods of estimation for confirmatory factor analysis with ordinal data. Psychological Methods, 9(4), 466-491. Wegmann, K. M., Thompson, A. M., & Bowen, N. K. (2010). A confirmatory factor analysis of home environment and home social behavior data from the ESSP for Families. Social Work Research, 35, 65-128. Optional article and good reference on missing value analysis: Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Structural Equation Modeling, 8, 430-457. HW 6 assigned Also for 3/18, find two SEM articles in your area of interest. Bring them to me for review as possible articles for your article critique (HW 6) 3/11/15: HAVE A GOOD SPRING BREAK! 9 3/18/15: Critiquing SEM studies, Power Analysis (class 10) 1. Critiquing SEM studies 2. Power analysis for SEM studies HW 6 due Bring two SEM articles to class for my review Required readings: Boomsma, A. (2000). Reporting analyses of covariance structures. Structural Equation Modeling: A Multidisciplinary Journal, 7(3), 461-483. Bowen & Guo, review Chapter 3, pp. 52-72 Bowen & Guo Chapter 7, sections on power analysis and the problem of underidentification, pp. 167-179 Lee, T, Cai, L., MacCallum, R. C. (2012). Power analysis for tests of structural equation models. In R. H. Hoyle (Ed.). Handbook of structural equation modeling. (pp. 181-194). New York: Guilford Press. MacCallum, R.C., Browne, M.W., & Sugawara, H.M. (1996). Power analysis and determination of sample size for covariance structure modeling, Psychological Methods 1, 130-149. HW 7 assigned 3/25/15: Multiple Group Modeling (class 11) 1. Path analysis 2. CFA 3. General structural models 4. Ordinal data RAT 4 Required readings: Bowen & Guo, review Chapter 6, 149-157 and Chapter 7, pp. 180-186 Campbell, H. L., Barry, C. L., Joe, J. N., and Finney, S. J. (2008). Configural, metric, and scalar invariance of the modified achievement goal questionnaire across African American and White university students. Educational and Psychological Measurement, 68, 988-1007. Dimitrov, D. M. (2010). Testing for factorial invariance in the context of construct validation. Measurement and Evaluation in Counseling and Development, 43, 121-149. Sass, D. A. (2011). Testing measurement invariance and comparing latent factor means within a confirmatory factor analysis framework. Journal of Psychoeducational Assessment, 29, 347-363. Simons, R. L., Johnson, C., Beaman, J., Conger, R. D., and Whitbeck, L. B. (1996). Parents and peer group as mediators of the effect of community 10 structure on adolescent problem behavior. American Journal of Community Psychology, 24, 145-171. Recommended: Cheung, G. W. & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255. 4/1/15: Longitudinal Analyses in SEM I—Repeated Measures Models (class 12) 1. Autoregressive, cross-lagged models 2. Specification, correlated errors 3. Identification HW 7 due Required readings: Cole, D. A. & Maxwell, S. E. (2003). Testing meditational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal of Abnormal Psychology, 112, 558-577. Find and read a recent SEM mediation study in your area of interest Be prepared to discuss its strengths and weaknesses 4/8/15: Longitudinal Analyses in SEM II--Latent Growth Models (class 13) RAT 5 Required readings: Curran, P. J., & Willoughby, M. T. (2003). Implications of latent trajectory models for the study of developmental psychopathology. Development and Psychopathology, 15. 581-612. Find and read a recent SEM growth model study in your area of interest Be prepared to discuss its strengths and weaknesses 4/15/15: Advanced Readings (class 14) Ferron, J. M. and M. R. Hess (2007). Estimation in SEM: A concrete example. Journal of Educational and Behavioral Statistics, 32: 110-120 Raykov, T., Marcoulides, G. A., & Li, C. (2012). Measurement invariance for latent constructs in multiple populations: A critical view and refocus. Educational and Psychological Measurement, 72, 954-972. Tomarken, A. J., & Waller, N. G. (2003). Potential problems with “well fitting” models. Journal of Abnormal Psychology, 112(4), 578-598. Others TBA 4/22/15: Final Exam, in-class (class 15) 11