2015 PSYCHOSTATISTICS Newsletter of the Quantitative Methods Section of CPA

Canadian Psychological Association Volume 1

Spring 2015 PSYCHOSTATISTICS Newsletter of the Quantitative Methods Section of CPA

Message from the Chair It is hard to believe that it has could draw an audience from ever issue of the QM In This Issue: already been three years since not only our section, but also Newsletter is evidence that the Quantitative Methods (QM) from a broad range of other some of the goals of the Message from the Chair ... 1 section of the Canadian CPA sections. In addition to section are becoming realities. Student Rep Message ... 2 Psychological Association the academic portion of CPA, We also have an awesome QM Laugh ... 2 Visualization Insights ... 3 (CPA) was formed. However, it many members of the section website, which you can view Consulting Corner ... 4 is even harder to believe that it also made it out to the QM at: 2015 Invited Speaker ... 4 took more than 70 years for social/dinner where it was http://www.cpa.ca/aboutcpa/c 2014 Convention ... 5 CPA to get a section related great to mingle with other pasections/quantitativemethod QM Grad Programs ... 6 specifically to QM. When you section members. s/. Section Elections ... 6 consider the vital role that QM QM Section Executive ... 7 plays in psychology, and the I am very excited about the In addition to these goals, we upcoming 2015 CPA would also like to find ways to incredible QM researchers from Canada, it really is hard to convention in Ottawa. Ottawa better promote QM imagine that the CPA existed has historically been a workshops, encourage for so long without a section popular CPA location, and we students to study QM in dedicated to statistics, are hoping that means a psychology graduate psychometric methods, etc. continuation of the large programs, promote audiences we have seen for discussions on better ways to QUANTITATIVE The 2014 CPA convention in our QM presentations. All teach QM, etc. We have METHODS Vancouver was another big CPA members will be started working to these goals SECTION OF success for the QM section, delighted to hear that Dr. via the website, which CPA building on the success of the James Ramsay from McGill contains links to many helpful very first convention for the University will be giving this resources. We hope that section in Quebec City in 2013. year's QM Invited Speaker many of you will become Website: In addition to the well-attended presentation. Jim will be active in the section and help workshops and symposia, the discussing psychometrics, us meet these goals and, http://www.cpa.ca/abo highlight of the convention was focusing on the interpretation more generally, grow the field utcpa/cpasections/qua the first ever QM invited of latent traits. Also keep your of QM for psychology in ntitativemethods/ speaker presentation by Dr. eyes open for the date and Canada. Bruno Zumbo from the time of the QM social. University of British Columbia. Dr. Zumbo entertained and Although the convention has Listserv: I hope to see you in Ottawa! enlightened a packed house on been the primary focus of the QM section over the past [email protected] the merits of Pratt Indices for Rob Cribbie couple years, there are many exploring the contribution of York University ways in which we are looking predictors in linear models. As [email protected] a new section, it was fantastic to grow the section. The fact to see that a QM focused talk that you are reading the first page 2 PSYCHOSTATISTICS

Message from the Student Rep Hello current and future quant coming methods, workshops, encourage you to attend all af icionados! or learning opportunities in of the sessions that you can. QM, then this section is for Also note that if you are “If you are a My name is Alyssa Counsell you too! We currently have a interested in learning more student with and I am the student great base of students from a about the area and how to interests in representative for the variety of backgrounds who get involved in the QM quantitative Quantitative Methods (QM) receive information via our section, you are welcome to section. That’s right, you methods, joining listserv, website and now this attend our section’s business heard me…quantitative newsletter. meeting. the QM CPA methods section! This year section will marks our third birthday since One of my favourite features I have so much more that I connect you with the official section creation at about the QM section is its would love to say about faculty and the Quebec City convention in small size. This means that CPA’s QM section, but I can’t students with 2013, and this article makes the convention is a great put it all into this small similar research history in our first section opportunity to meet like- section. Instead, I am happy interests” newsletter! minded individuals. And let’s to answer more specific be honest, everyone wants to questions about the section, Some of you may be thinking have at least one quant being a student in a QM that you can’t imagine anyone friend…At the Vancouver graduate program, etc. wanting to devote themselves convention last year, I met Feedback and suggestions to research in statistics, some great people at our QM from students about the QM psychometrics, and related social, and I know that others area are also welcome and topics, but believe it or not, felt the events were wonderful encouraged! there are many such opportunities to network and The QM researchers in psychology mingle. This year we will have I’m very much looking Section departments in Canada. If you similar events with details forward to reconnecting with currently has are a student with interests in provided closer to the dates of QM friends and colleagues at quantitative methods, joining the upcoming CPA 249 student the Ottawa convention. the QM CPA section will Everyone interested in QM is convention, but I am members ... connect you with faculty and welcome! especially excited to meet all students with similar research of the new QM members! interests or if you are a This year’s CPA convention See you in Ottawa this June! student who would simply like will provide many exciting QM to learn more about sessions: workshops, posters, Alyssa Counsell quantitative research or be ‘in symposia, and our annual York University the know’ about up-and- business meeting. We [email protected]

QM Laugh

Source: http://www.explainxkcd.com/

page 3 PSYCHOSTATISTICS

Visual Insights

QM Funny

Thank you to Matt Sigal, a QM doctoral student in the Department of Psychology at York University, for this issue's Visual Insight. Matt specializes in, among other things, visualizing psychological data. Send 'Visualization Insights' ideas to Nicole Aitken ([email protected]), Communications Director of the QM Section

PSYCHOSTATISTICS page 4

Consulting Corner Dear Stats Consultant, coefficients in your full model variances. Also be aware are only looking at the unique that the relationship between I have three subscale scores contribution of each a subscale and the outcome that I would like to include as predictor. When predictors can actually be stronger predictor variables in a share a large amount of when controlling for the regression model. All of the variance with each other, the remaining predictors. Lastly, constructs represented by shared variance is not taken if you are interested in the subscales are into account for predicting predicting the outcome from theoretically important. When your outcome variable. It is both the unique and shared I run the regression model typical for subscale scores to variability, then you could run with all of the subscales in be related to one another. separate regression analyses the regression only two of for each predictor. Generally

them are significant. It is important for you to this is not what researchers However, running separate determine which variability are interested in, and may regression models using one you are interested in: 1) the provide misleading subscale at a time produces shared variability; 2) the information about the significant results for all three unique variability; 3) unique strength of the relationship subscales. Why does this plus shared variability. If you between each predictor and happen and what is the are interested in the shared the outcome variable (e.g., correct model? variability then you could try the sum of the separate R2 summing the scores on the values does not represent Sincerely, three subscales and using the total variability explained Jane Tukey those as predictors in your by the three predictors). Send consulting regression model, or run a When running regression Dear Ms. Tukey, structural equation model models, including all corner suggestions with a latent variable (or theoretically relevant to Nicole Aitken This is a common problem in variables) representing the predictors is the best method regression. You likely have ([email protected]), shared variability among the for obtaining unbiased significant predictors in Communications predictors . If you are regression estimates and individual models but not in interested in the unique maximizing the amount of Director of the QM the multiple regression model variability, then you will want variance accounted for in because the regression Section to leave all predictors in the your outcome variable. coefficients in the full model regression. Be sure to check are partial regression though that collinearity is not Sincerely, coefficients, whereas in a an issue, since highly Your Friendly Neighbourhood simple regression they are correlated predictors can Stats Consultant not. This means that the lead to inflated error

QM Section Invited Speaker at CPA 2015 The QM Section's Invited McGill in 2007 after His research interests Speaker at this year's CPA approximately 40 years of include psychometrics, dynamic systems and Convention is Dr. James service. Currently, Jim also has affiliations with the functional data analysis, of Ramsay, Emeritus Professor Department of Chemical which he is the founder. He of Psychology from McGill Engineering at Queen's also developed much of the University. The topic of Dr. University and the statistical theory behind Ramsay's presentation is A Department of Mathematics multidimensional scaling. Jim Friendlier Psychometric and Statistics at the is the author of four Theory for Tests and University of Ottawa. He influential books and over Psychological Scales. received his PhD from 100 peer-reviewed articles in statistical and psychometric Princeton University in 1966 Dr. Ramsay retired from in Quantitative Psychology. journals.

PSYCHOSTATISTICS page 5

2014 CPA Convention: A Great Success for QM! Section Keynote An Overview of New Developments in Psychometrics with Implications for Validation Studies: Pratt Indices for Latent Variable Models, Dr. Bruno Zumbo, University of British Columbia

Preconvention Workshop The QM Section Introduction to Structural Equation Modeling, Rob Cribbie & Alyssa Counsell, York University Program for 2015 in Ottawa Workshops will be available Getting Started with R: An Introduction to Data Management and Analysis using Free Software, soon. Watch our S. Sampasivam, University of Ottawa; Nicole Aitken, University of Ottawa; Joo Ann Lee, York University; Susanna Cheung, University of Ottawa website for more information. Structural Equation Modeling: Made Easy in Free Software, Katherine Collins, University of Ottawa; Susanna Cheung, University of Ottawa; Alyssa Counsell, York University

Analyzing Dyadic Data with the Actor-Partner Interdependence Model (APIM): A Systematic Walkthrough of Recent Applications and Developments, Alexandre Garneau, University of Ottawa; Josée Fitzpatrick, University of Ottawa; Patrick Gaudreau, University of Ottawa; Marie-France Lafontaine, University of Ottawa.

Symposia Advanced Methods for Longitudinal Data Analysis, Cathy LaBrish, Victoria Ng, David Flora, York University

Posters Investigating the robustness of the nonparametric Levene test with 3 or more groups, David Nordstokke, University of Calgary; Mitchell Colp, University of Calgary

Power and type I error rates of three robust chi-square difference tests in studies of measurement 2015 Pre- invariance, Jordan Brace, University of British Columbia; Victoria Savalei, University of British Columbia Convention Workshop: Dr. Neyman and Pearson versus Fisher controversy is informative of current practice in inferential statistics, Michael T. Bradley, University of New Brunswick; Andrew Brand, King's University College; Luke Macneill, Dennis Jackson University of New Brunswick from the University of A general approach to combining parallel factor analysis and cluster analysis, Ji Yeh Choi, Heungsun Hwang, Mcgill University Windsor will be giving a pre- Comparing pre-post change across groups: Guidelines for choosing between difference scores and convention ANCOVA, Megan A. Jennings, Robert A. Cribbie, York University workshop on The status of models in construct validity research: Theoretical models or theoretical entities?, Maria factor analysis. Nikol, Kathleen L. Slaney, Simon Fraser University Sign up now so

Apples, file drawers and garbage: Revisiting validity issues in meta-analysis, Sarena Poets, Donald you are not Sharpe, University of Regina "latent"

Disparities in the use and interpretation of meta-analysis models: An empirical examination, Kathleen L. Slaney, Simon Fraser University; Ramsay Malange, University of Victoria

How to minimize the impact of reverse-worded items on psychological scales, Xijuan (Cathy) Zhang, Victoria Savalei, University of British Columbia

Demonstrations of the attenuation of effect sizes by fluctuating levels of measurement error, Shawna Zuda, University of British Columbia; Brian P. O'Connor, University of British Columbia

page 6 PSYCHOSTATISTICSpage 5

QM Graduate Programs in Canada Source: http://www.apa.org/divisions/div5/docprogs.html#capsydev

University of British Columbia MA/PhD, Quantitative Methods http://psych.ubc.ca/graduate/research-areas/quantitative-methods/ Contact: Jeremy Biesanz Email: [email protected]

Numerous resources University of Manitoba related to the study of MA/PhD Methodology quantitative methods for http://umanitoba.ca/faculties/arts/departments/psychology/graduate/programs/analysis.html psychology can be found Contact: Johnson Li on the APA Website. See: Email: [email protected] http://www.apa.org/ McGill University research/tools/quantitative/ PhD, Quantitative Psychology and Modeling There you will find, among http://www.psych.mcgill.ca/research.html#q other things, that relative Contact:Yoshio Takane to other areas of Email: [email protected] psychology there is a much greater chance of Simon Fraser University getting a job with a PhD in MA/PhD, Quantitative Methods Quantitative Methods. http://www.psyc.sfu.ca/index.php?topic=theory Contact: Rachel Fouladi Email: [email protected]

University of Western Ontario MSc/PhD, Personality and Measurement Program http://www.psychology.uwo.ca/research/personality_and_measurement/index.html Don Saklofske Email: [email protected]

York University MA/PhD, Quantitative Methods http://qm.info.yorku.ca/ Contact: David Flora Email: [email protected]

2015 -2016 QM Section Elections A list of available executive Elections for QM Section positions will occur at the QM Annual positions in the QM section can be found on pages 7 Meeting during the CPA Convention. If you interested in running for and 8 a position, or if you would like to nominate someone for a position, you can do so by emailing Rob Cribbie ([email protected]) or nominations will also be accepted during the Annual Meeting.

PSYCHOSTATISTICS page 7

Meet Your 2014-2015 QM Section Executive

Chair: Rob Cribbie Quantitative Methods Program Department of Psychology

York University Special points [email protected] of interest:

• QM Section of CPA was formed in 2013

• Don Sharpe was the first chair of Chair-Elect: the section • CPA 2015 is in Michael Bradley Ottawa; CPA 2016 is in Department of Psychology Victoria University of New Brunswick [email protected]

Past-Chair: Donald Sharpe Department of Psychology University of Regina [email protected]

page 8 PSYCHOSTATISTICS

Meet Your 2014-2015 QM Section Executive

Secretary/Treasurer: R. Nicholas Carleton Department of Psychology Do you want to get University of Regina involved with the [email protected] Quantitative Methods Section of CPA?

If so, email any of the members of executive ... we'd love to have you! Communications Director: If you are not Nicole Aitken already a member of our listserv, please Department of Psychology send an email to University of Ottawa Rob Cribbie so you [email protected] don't miss out on future newsletters,

convention news, training opportunities, etc. Student Representative:

Alyssa Counsell Quantitative Methods Program Department of Psychology York University [email protected]