Academics Navigation
Academics
You are here
Courses
Campus | Bryn Mawr |
Semester | Fall 2020 |
Registration ID | PSYCB314001 |
Course Title | Multivariate Statistics |
Credit | 1.00 |
Department | Health Studies |
Instructor | Schulz,Marc |
Times and Days | Th 09:40am-12:30pm
|
Room Location | |
Additional Course Info | Class Number: 2255 This course is designed to improve your data science skills by introducing you to advanced statistical techniques that have become increasingly important in psychology and a variety of fields. The focus will be on understanding the advantages and limitations of regression approaches and multivariate analytic techniques that permit simultaneous prediction of multiple outcomes. Topics covered will include basic regression approaches, advanced regression strategies, structural equation modeling, factor analysis, measurement models, path modeling, modeling of longitudinal data sets, multilevel modeling approaches and growth curve modeling. Students will gain familiarity with these techniques by working with actual data sets. The last part of each class will be reserved for lab time to apply lessons from class to an assignment due the following week. Students are welcome to stay beyond the noon ending time to complete the assignment. Prerequisites: Required: PSYC Research Methods and Statistics 205 (BMC), Psych 200 (HC) Experimental Methods and Statistics, or BIOL B215 Experimental Design and Statistics. Students with good statistical preparation in math or other disciplines and some knowledge of core methods used in social science or health-related research should consult with the instructor to gain permission to take the class.; PRELIMINARY SYLLABUS Psychology 314 Advanced Data Science: Regression and Multivariate Statistics Fall 2020 Marc Schulz: 610-526-5039 (mschulz@brynmawr.edu) Meeting Time: Thursday 9:30-12:30 (in most weeks, lab sessions will be held during the last hour of the class) Office Hour: TBA TA: Kate Petrova (epetrova@brynmawr.edu) TA Drop-in hour: TBA Course Description: This course is designed to improve your data science skills by introducing you to advanced statistical techniques that have become increasingly important in psychology, health-related disciplines, education and a variety of other fields. The focus will be on understanding the advantages and limitations of multivariate analytic techniques that permit simultaneous prediction of multiple outcomes. We will begin with basic regression approaches, which are the foundation for the more complex approaches that follow. We will cover structural equation modeling, factor analysis, measurement models, path modeling, analysis of change and longitudinal data, multilevel modeling approaches, and growth curve modeling. Students will gain familiarity with these techniques by working, each week, with actual data sets and learning to implement analyses in specialize software. Emphasis will be placed on helping you critically evaluate applications of these techniques in the literature and the utility of applying these techniques to your own work. Methodological considerations will also be highlighted as we regularly use data from actual studies. Course Objectives and Learning Goals • Improve your data analytic knowledge and skills • Improve your ability to connect analytic strategies to research questions • Develop your understanding of the strengths and weaknesses of different methodological and statistical approaches • Develop your ability to interpret statistical analyses and discuss the implications of findings • Improve your ability to report/present statistical analyses effectively • Learn multivariate approaches to analyzing data Course Requirements: Readings for a given week should be done before class. Some of the material from the readings is challenging and may become clear only after class and completing relevant exercises. However, completing the reading prior to class will facilitate your learning of class material. This year, I am likely to pre-record some lecture material that you should watch prior to class (but after doing the reading for the week). This pre-posted material is intended to free up more time for the lab portion of the class during our scheduled 3-hour time block. There will be some “low-stakes” questions to complete after each pre-recorded lecture. Active engagement with the material covered in class is critical. The best way to learn complex statistics is to plan and carry out actual analyses with real data. Homework assignments will be given each week and are due prior to the next class. The end-of-class lab time is designed to help you get a strong start on analyses that are part of the assignment for each week. Depending on class size AND on software and room availability issues related to COVID, students will either be able to do the in-person lab each week OR we may split the class into two groups with each group alternating in-person labs with on-line, Zoom access to labs. This is a “hands-on” course that requires consistent attendance and active engagement. Timeliness of handing in homework, effort in completing the homework, completion of the low-stakes pre-recorded lecture assignments, engagement in class and lab, and the quality and thoroughness of the write-up of the homework will be the primary determinants of your grade. Attendance is expected. Class attendance, participation, and overall engagement will count for 15% of your grade. Technology “etiquette” for our class: 1. Technology brings many advantages but it also can be a significant source of distraction for the user and for those around the user. It is important that we all remain engaged in class and able to listen to each other. Please leave your cell phones and other electronic devices off or on silent status during class. I promise to do the same. If you have a compelling reason (e.g., family emergency) to keep your phone on, please speak to me ahead of time. If you have a strong preference for using a laptop or tablet to take notes, you may do so but under no circumstance is the device to be used for non-notetaking activities (e.g., web-browsing, Facebooking, etc.) during class as these activities not only distract you but they also distract the rest of us. 2. The best way to contact me outside of class or office hours is by email. I strive to respond to email within 24 hours. During the weekends, please expect a longer delay in response. Accommodations: Bryn Mawr College is committed to providing equal access to students with a documented disability. Students needing academic accommodations for a disability must first register with Access Services. Students can call 610-526-7516 to make an appointment with the Director of Access Services, Deb Alder, or email her at dalder@brynmawr.edu to begin this confidential process. Once registered, students should schedule an appointment with me as early in the semester as possible to share the verification form and make appropriate arrangements. More information can be obtained at the Access Services website. (http://www.brynmawr.edu/access-services/) Any student who has a disability-related need to record this class first must speak with the Director of Access Services and to me, the instructor. Class members need to be aware that this class may be recorded. On-time Completion of Assignments: The content of this course is challenging, and each week’s work builds on the previous week. I expect work to be completed on time. Late work will disrupt your progress and also raises challenging issues about fairness. If clearly justified, I will consider granting a rare extension if you make the request for the extension at least 24 hours prior to the deadline. If an emergency develops (e.g., sudden illness, family crisis) the 24-hour rule does not apply but please let me know as soon as it arises. Responsible Data Use: You will be working with actual data that have been de-identified to protect the anonymity of the participants. These data can only be used for analyses for the course; they cannot be downloaded and used for your own purposes outside of class. These procedures help protect the participants (and researchers) that generously contributed to this research. Texts REQUIRED TEXT: Kline, Rex B. (2016). Principles and Practice of Structural Equation Modeling (4th edition). NY: Guilford. • A copy will be on reserve in Canaday library if there is access to Reserves during the semester. OPTIONAL TEXT: Singer, J.D. & Willett, J.B. (2003). Applied longitudinal data analysis. NY: Oxford • Copy is on reserve in Canaday library. Other Readings Additional required readings will be available on Moodle. Useful On-Line Resources: • Free 30-day trial from SPSS of AMOS SEM software: • https://www.ibm.com/us-en/marketplace/structural-equation-modeling-sem • Free student version of the HLM 8 software we use in class: • HLM: http://ssicentral.com/index.php/products/hml/free-downloads-hlm • Other helpful websites for SEM, Multilevel Modeling or General Statistics Information: • David Kenny’s website (leader in dyadic analyses, meditational analyses and multivariate modeling generally): http://davidakenny.net/kenny.htm • For a good source of references on Structural Equation Modeling, Multilevel Modeling and other statistical approaches, visit Jason Newsom's webpage: http://web.pdx.edu/~newsomj/ • Discussion of multivariate stats within the MPLUS stats package; helpful for other stats packages as well: http://www.statmodel.com/cgi-bin/discus/discus.cgi • Ed Rigdon’s Compilation of Useful SEM WEB resources: • http://www2.gsu.edu/~mkteer/ • ListServ Discussions of Multivariate Techniques (registration may be required) • SEMNET ListServ Discussion: http://www2.gsu.edu/~mkteer/semnet.html • Multilevel modeling issues: http://www.jiscmail.ac.uk/lists/multilevel.html Mediation and Moderation: • For references on mediation and moderation (including useful macros and calculators), visit Kristopher Preacher’s website: http://www.quantpsy.org/medn.htm • Dave McKinnon’s website on mediation analysis: http://ripl.faculty.asu.edu/mediation/ • Preacher and Hays’ bootstrap and Sobel macros: http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html Outline of Course, Readings and Assignments PART I: Regression and Structural Equation Modeling 9/10, Week 1: Review of Correlation and Multiple Regression; Overview of Multivariate Techniques, Introduction to Covariance Matrices • SEM Applications • Multilevel Modeling Applications (Hierarchical Linear Modeling • Correlations & Introduction to Canadian National Public Health Survey (NPHS) data • Effect sizes • Bivariate Regression 101 • Standardized vs. Unstandardized Assignment: Correlation in SPSS, learning to report statistical associations in clear, simple language (will use Canadian PHS data) • Due Week 2 (Monday by 2 PM) in box in BYC mailroom 9/17; Week 2: Multiple Regression • Partial Regression Coefficients • Types of Multiple Regression • Effect Sizes o Standardized estimates o Variance Explained • Multiple Regression implemented in SPSS o Interpreting output o Reporting Regression Analyses • Correlations and Covariances Reading: Kline, Chapter 2, “Regression Fundamentals” (skip pp. 42-47) and Chapter 3, pp. 49-54 (standard errors, critical ratios and power and types of null hypotheses) Assignment: Multiple Regression Analyses in SPSS and Reporting Regression Results 9/24, Week 3: Introduction to AMOS and Structural Equation Modeling • Using SEM to do multiple regression analysis • SEM drawing conventions • Introduction to Covariance Structure Analysis • Using AMOS Reading: Kline, Chapter 6, pp. 117-122, skim 122-126, read 129-135 Assignment: Fitting Multiple Regression Models Using AMOS (due before Week 3) 10/1 Week 4: Multivariate Regression: Predicting Multiple Outcomes • Model Identification • Goodness of Fit Indices • Model Modification • Nested Models and Hypothesis Testing Reading: Kline : • Kline, Chapter 7, pp. 145-148 on Identification • Kline, Chapter 12, 262-266, 270-272 ; 277 (on SRMR), and 280-289 (Testing Hierarchical Models) Data example for class will be drawn from: Davis, M.H., & Franzoi, S.L. (1986). Adolescent loneliness, self-disclosure, and private self-consciousness: A longitudinal investigation. Journal of Personality and Social Psychology, 51, 595-598. Familiarize yourself with how to use AMOS by reading through the Tutorial (CHAPTER 2), Getting Started with AMOS GRAPHICS, in the AMOS (26) USER's Guide: • The User Guide is also available on the course Moodle Page or via link on the following web page: https://www-01.ibm.com/support/docview.wss?uid=ibm10878803 HELPFUL REFERENCES ON FIT INDICES: Special issue on fit indices in Personality and Individual Differences: Volume 42, Issue 5, pages 811-898 (May 2007) McDonald, R.P., & Ho, M.H.R., (2002). Principles and practice in reporting structural equation analyses. Psychological Methods, 7, 64-82. o Technical but also helpful as a guide for writing up SEM results McDonald, R.P. (2010). Structural models and the art of approximation. Perspectives in Psychological Science, 5(6), 675-686 Boomsa, A. (2000). Reporting analyses of covariance structures. Structural Equation Modeling, 7, 461-483. Hu, L.-t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1-55. Assignment: Fitting a Multivariate Regression Model BACKGROUND REFERENCE FOR ASSIGNMENT: Pajares, F. & Miller, M.D. (1995). Mathematics self-efficacy and mathematics performances: The need for specificity of assessment. Journal of Counseling Psychology, 42, 190-198. 10/8, Week 5: Path Analysis; Measurement Model 1: Using reliability estimates to improve model estimates • Additional AMOS features • Interpreting Path Coefficients • Standardized Solutions • Examining Residuals • Measurement theory and reliability estimates Reading: Kline: • pp. 88-94 on reliability and validity • pp. 266-270 (approximate fit indexes and recommended approach to fit evaluation), 273-2 Data example for class will be drawn from: Taylor, R.D., Casten, R., Flickinger, S.M., Roberts, D., & Fulmore, C.D. (1994). Explaining the school performance of African-American adolescents. Journal of Research on Adolescence, 4, 21-44. • Concentrate on pages 21-37, including figures 1-4, but ignore Tables 3-8. Hoyle, R. H., & Panter, A. T. (1995). Writing about structural equation models. In R. H. Hoyle (Ed.), Structural equation modeling (pp. 158-176). Thousand Oaks, CA: Sage. (Skim the section on fit indices) Helpful REFERENCES ON DIAGNOSING PROBLEMS IN SEM: Tomarken, A.J., & Walker, N.G. (2003). Potential problems with “well-fitting“ models. Journal of Abnormal Psychology, 112, 578-598. Chen, F., Bollen, K.A., Paxton, P., Curran, P., & Kirby, J. (2001). Improper solutions in structural equation models: Causes, consequences, and strategies. Sociological Methods & Research, 29, 468-508. Assignment: Conducting Path Analysis and Using Estimates of Reliability to Disattenuate the Findings BACKGROUND REFERENCE FOR ASSIGNMENT: Dumka, L.E. & Roosa, M.W., (1993). Factors mediating problem drinking and mothers’ personal adjustment. Journal of Family Psychology, 7, 333-343. 10/15 Week 6: Using Multiple Indicators to Account for Measurement Error • Latent Constructs and Manifest Indicators • Hybrid Models: Structural model and measurement model • Hypothesis Testing and Model Constraints Reading: Data examples for class will be drawn from: Holahan, C.J., Valentiner, D.P., & Moos, R.H. (1994). Parental support and psychological adjustment during the transition to young adulthood in a college sample. Journal of Family Psychology, 8(2), 215-223. Conger, R.D., Conger, K.J., Elder, G.H., Lorenz, F.O., Simons, R.L., & Whitbeck, L.B. (1993). Family economic stress and adjustment of early adolescent girls. Developmental Psychology, 29(2), 206-219. Assignment: Fitting a SEM with multiple indictors of a latent variable BACKGROUND REFERENCE FOR ASSIGNMENT: Berndt, T.J., & Miller, K.E. (1990). Expectancies, values and achievement in junior high school. Journal of Educational Psychology, 82, 319-326. 10/22, Week 7: Confirmatory Factor Analysis (CFA) • Goals of Factor Analysis • Exploratory vs. Confirmatory Factor Analysis • Orthogonal vs. Oblique Factor Structure • Identification guidelines for CFA • Factor Loadings • Higher Order Factor Analysis Reading: Kline: • Chapters 9 and 13 (pp. 300-321) • MacCallum, R.C., Roznowski, M., & Necowitz, L.B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111(3), 490-504. o Concentrate on the beginning (490-494) and end (501-503) of this article where the authors caution about modifying hypothesized models. Data example for class will be drawn from: SKIM: Benson, J. & Bandalos, D.L. (1992). Second-order confirmatory factor analysis of the reaction to tests scale with cross-validation. Multivariate Behavioral Research, 27, 459-487. General Reference (not required reading) MacCallum, R. C., & Austin, J. T. (2000). Applications of structural equation modeling in psychological research. Annual Review of Psychology, 51, 201-226. References on Exploratory Factor Analysis: Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4, 272-299. Preacher, K.J., & MacCallum, R.C. (2003). Repairing Tom Swift’s electric factor analysis machine. Understanding Statistics, 2, 13-43 Assignment: Conducting a First-Order Factor Analysis BACKGROUND REFERENCE FOR ASSIGNMENT: Lee, R.M. & Robbins, S.B. (1995). Measuring belongingness: The social connectedness and the social assurance scales. Journal of Counseling Psychology, 42, 232-241. 10/29, Week 8: October 23, Week 7: SEM and Longitudinal Data Analysis • Advantages of longitudinal designs • Stability of rank ordering over time vs. individual growth trajectories • Autoregressive Models • Handling missing data Reading: Kline: Section on missing data strategies: pp. 82-88 Data examples for class will be drawn from: Farrell, A.D. (1994). Structural equation modeling with longitudinal data: Strategies for examining group differences and reciprocal relationships. Journal of Consulting and Clinical Psychology, 62, 477-487 Conger, K.J. & Conger, R.D. (1994). Differential parenting and change in sibling differences in delinquency. Journal of Family Psychology, 8(3), 287-302. Assignment: Fitting a Structural Equation Model with Longitudinal Data BACKGROUND REFERENCE FOR ASSIGNMENT: Conger, K.J. & Conger, R.D. (1994). (listed above) 11/5, Week 9: Multi-group Analysis, Analysis of Mediation and Indirect Effects Reading: Kline, Chapter 16 (skim both data examples – both the continuous and the ordinal indicators) Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the study of mediators and moderators: Examples from the child-clinical and pediatric psychology literatures. Journal of Consulting and Clinical Psychology, 65(4), 599-610. Hayes, A.F. & Scharkow, M. (2013). The Relative Trustworthiness of Inferential Tests of the Indirect Effect in Statistical Mediation Analysis: Does Method Really Matter? Psychological Science, 1918-27. o this one is a bit technical, but giving it a read will help prep you for our discussion about mediation and estimating indirect effects Data example for class will be drawn from: Holahan, C.J. & Moos, R.H. (1991). Life stressors, personal and social resources, and depression: A 4-year structural model. Journal of Abnormal Psychology, 100, 31-38. Classic References on Mediation and Moderation (included with Moodle readings but not required reading): Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1278. MacKinnon, D.P., Lockwood, C.M., Hoffman, J. M., West, S.G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104. (Not in Moodle readings) References on sample size limitations and power in SEM: Maas, C. J. M., & Hox, J. J. (2005). Sufficient Sample Sizes for Multilevel Modeling. Methodology. European Journal of Research Methods for the Behavioral and Social Sciences, 1, 85-91. Maas, C. J. M., & Hox, J. J. (2004). Robustness issues in multilevel regression analysis. Statistica Neerlandica, 58, 127-137. 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. Assignment: Multi-group SEM analysis BACKGROUND REFERENCE FOR ASSIGNMENT: LIKELY TO CHANGE Harold, G.T., Fincham, F.D., Osborne, L.N., & Conger, R.D. (1997). Mom and Dad are at it again: Adolescent perceptions of marital conflict and adolescent psychological distress. Developmental Psychology, 33, 333-350. PART II: Multilevel Modeling: Repeated Measures and other Nested Data Designs 11/12, Week 10: Analyzing Change over Time, Introduction to Multilevel Models for Individual Change, Growth Curve Modeling I • Approaches to assessing change • Growth Curve Modeling and Multilevel Models • Application: Marital Satisfaction and the transition to parenthood Reading: Singer, J.D. & Willett, J.B. (2003). Applied longitudinal data analysis. NY: Oxford, pp. 3-40. Cowan, C.P. (1988). Working with men becoming fathers: The impact of a couples group intervention. In P. Bronstein & C.P. Cowan (Eds.), Fatherhood today (pp. 276-298). NY: Wiley. Assignment: Graphing individual growth trajectories of marital satisfaction 11/19, Week 11: Introduction to HLM; Growth Curve Modeling II • Nested data structures • HLM nomenclature • How to do HLM analyses Reading: Singer & Willett: pp. 45-74 (skim pp. 65-68); 92-104. Karney, B.R. & Bradbury, T.N. (1995). Assessing longitudinal change in marriage: An introduction to the analysis of growth curves. Journal of Marriage and the Family, 57, 1091-1108. Assignment: Create a MDM file; run growth curve analyses THANKSGIVING BREAK (and tra Enrollment Cap: 16; If the course exceeds the enrollment cap the following criteria will be used for the lottery: Major/Minor/Concentration; Senior; Junior; Sophomore; The course will be taught primarily synchronously and in-person. Lab time and weekly assignments will give you the opportunity to conduct analyses with real data asking addressing meaningful questions. In most weeks, lab sessions will be held during the last hour of the class (from 11:30-12:30). This year, I am likely to pre-record some lecture material that should be watched prior to class. This pre-posted material is intended to reduce our in-person exposure risks (although the class will be held in a large room with good circulation) AND free up more time for the lab portion of the class during our scheduled 3-hour time block. The lab time is designed to help you get a strong start on analyses that are part of the assignment for each week. Depending on class size AND on software and room availability issues related to COVID, students will either be able to do the in-person lab each week OR we may split the class into two groups with each group alternating in-person labs with on-line, Zoom access to labs. |
Miscellaneous Links |