Arizona Methods Workshops: About the Workshops and Instructors

MORNING SESSIONS- 9-12 each day

Data Management and Programming in Stata
Jennifer Earl

Although standard statistics courses prepare scholars to design and run statistical analyses, courses generally do not spend a great deal of time discussing data management and workflows, which are critical to making research replicable, efficient, and accurate. This is unfortunate because the best designed statistical analysis is easily undone by poor data management, whether through misconstructed variables, unreplicable workflows, and/or poorly commented or documented workflows and programs. This course addresses this oversight using Stata by harnessing Stata’s powerful tools for managing and preparing data before analyses. The course will cover fairly basic data manipulations as review but then progress to increasingly powerful programming tools, including extensive use of by statements, collapse, post, loops, and sub-routines in Stata. Students interested in applying these skills in other programs (e.g., R) will find that although the specific commands and formatting differs between packages, the logics are quite similar and thus skills are surprising portable between programs—it is like learning regional dialectics of the same language.

Jennifer Earl is a Professor of Sociology at the University of Arizona. Her research focuses on social movements, information technologies, and the sociology of law, with research emphases on Internet activism, social movement repression, and legal change. She is the recipient of a National Science Foundation CAREER Award for research from 2006-2011 on Web activism. She is also a member of the MacArthur Research Network on Youth and Participatory Politics. She has been programming in statistical packages for over 15 years and has taught graduate classes on programming in both Stata and SAS.

Qualitative Comparative Analysis
Claude Rubinson

Invented by Charles Ragin in the late-1980s, QCA offers an alternative to conventional statistical methods based on the analysis of set-theoretic relationships and is suitable for small-, medium-, and large-N studies. This workshop will describe the method, examine its strengths and weaknesses, discuss recent developments and extensions, suggest best practices for conducting and presenting QCA, and recommend resources for further study. Software for conducting QCA will be reviewed and demonstrated. This workshop is designed both for novices and those with previous QCA experience. Participants are asked to bring their laptop (Windows, Mac, or Linux/Unix). Those currently working on a QCA project are encouraged to also bring their dataset; time will be reserved for discussing these projects and working with the software.
 
Claude Rubinson is an associate professor of sociology at the University of Houston-Downtown where he teaches research methods, sociology of culture, and social inequality.  He received his Ph.D. from the University of Arizona in 2010, under the direction of Charles Ragin, with whom he has published a series of articles discussing comparative research and QCA.  Professor Rubinson's research program has three tracts: methodological research on developing formal methods of qualitative research, focusing on QCA; sociological research on the relationship between the global political-economy and aesthetic form; and health services research on the conditions that hinder/facilitate the success of interventions designed to improve patient outcomes.  In addition to leading didactic seminars on QCA and regularly consulting on a variety of QCA-related research projects, Professor Rubinson is a member of the management team of COMPASSS, the international, inter-university consortium of QCA methodologists and practitioners, and editor of the COMPASSS Working Paper series.  He is currently studying the contextual conditions that affect the success of an intervention program designed to prevent medication-associated acute kidney injury in pediatric patients.
 
Intro to R
Katerina Sinclair
 
R is a free, open-source statistical analysis program that has taken the world, and the job market, by storm. Participants will learn how to get data into R and conduct simple analyses, including descriptive analyses and data cleaning, t-tests, ANOVAs, correlations, and uni- and multi-variate regressions. We'll learn about cheater packages that turn R into a point-and-click program and we'll even glide over a few graphics packages where R's value really shines! Participants should be familiar with basic statistical concepts, have used some type of analysis program before (e.g., SAS, SPSS, Excel), and should bring laptops with R already loaded (instructions on how to do this will be sent before the workshop).

Dr. Katerina Sinclair has been an enthusiastic R useR for over a decade. After initially using it to do things no other program could do, she eventually came to use it to do pretty much everything. She earned a concurrent Masters of Applied Statistics and a PhD in Human Development and Family Studies, with a focus on research methodology and statistics, from the Pennsylvania State University. She has worked in academia and industry and currently works as the Principal Data Scientist for UnitedHealth Group in an R & D department. Her past work included a stint at the U of A wherein she tried to convince every social science department to implement “R-only” policies. She eats data for breakfast and retains a steadfast commitment to improving people's lives by writing better open-source algorithms.​​

AFTERNOON SESSIONS - 2-5 pm each day

Social Network Analysis
Joseph Galaskiewicz

The purpose of this workshop is to give an introduction to the methods used in the social and behavioral sciences to do network analysis.  We will also introduce you to UCINET, software that is a good beginner’s tool.  We will focus on data gathering, network measures, and the fundamental analytic strategies.  The course cannot be anything but an introduction to the field, so we will go over the material quickly, but provide you with references which can take you more in-depth. 

There is no assumption that you ever did network analysis before, so the first day will be very elementary.  The second day will explore measures and simple analytic techniques.  Finally, the third day will cover some advanced topics so as to motivate you to study more.  I will not assign homework, but I hope that you can do readings before or in-between classes.

Professor Galaskiewicz (Ph.D. Sociology, University of Chicago, 1976) contributes to theory and research on social networks, nonprofit organizations, and urban inequality.  He is now introducing the concept of spatial capital, which is based on the idea of spatial networks, into the sociological literature.  Spatial capital refers to the differential access of urban residents to goods and services across the urban landscape via the urban grid.  Currently he does research on the Phoenix metro area.  Professor Galaskiewicz received National Science Foundation funding almost continuously for the last 37 years and has served on various NSF panels.  He received the 2014 Award for Distinguished Achievement and Leadership in Nonprofit and Voluntary Action Research awarded by the Association for Research on Nonprofit Organizations and Voluntary Action.  He has published three research monographs and one edited volume and lists 56 articles published in refereed journals or edited volumes including ASR, AJS, ASQ, SF, ARS, and AMJ

Conducting Survey, Lab, and Field Experiments in the Social Sciences
S. Michael Gaddis

During the past decade, experiments in the social sciences have gained in popularity as the internet has made implementing experiments easier, cheaper, and faster.  However, although social scientists may have a conceptual knowledge of how experiments work, the actual experience of implementing an experiment for the first time is often frustrating and time consuming.  Researchers without prior experience often struggle with a number of issues such as navigating IRB, obtaining true random sampling and assignment, understanding blocking, and interpreting different types of treatment effects.  The initial learning curve may be steep but the rewards are plentiful as experiments produce original data often highly valued in top journals, lend themselves to causal analysis in ways that traditional survey data cannot, and become easier to implement as a researcher’s experience level increases. 

This course is designed for a variety of social scientists who are interested in social science experiments (survey, lab, and/or field).  Those with only textbook knowledge of experiments will leave with the confidence to conduct their own experiments.  Others with some experience designing and implementing experiments will leave with a deeper understanding of methods that can help save them time and effort with future experiments.  In this course, we will begin with identifying what types of questions to ask with experiments and work our way through advanced analysis techniques.  Hands-on activities will give students the opportunity to design and implement a survey experiment and explore the basics of a field experiment during the workshop.

S. Michael Gaddis is an Assistant Professor of Sociology at UCLA.  His work explores the causes and consequences of racial and educational inequality, often through experiments.  He has led the data collection efforts on nearly a dozen field and survey experiments.  He is the editor of and a contributor to the forthcoming book Audit Studies: Behind the Scenes with Theory, Method, and Nuance due out from Springer in late 2017.  His research has been published in the American Journal of Sociology, Social Forces, Social Science Research, and Educational Evaluation and Policy Analysis, among other outlets.

Data Science in R
Katerina Sinclar & Yotam Shmargad

OuR adventures continue as we focus on more advanced programming and packages in R.  Participants should be well-versed in statistical concepts, have some experience with R or with data science methods and terms, and should bring laptops with R already loaded (instructions on how to do this will be sent before the workshop).  This workshop has three distinct components.  The first day will focus on classification models and be taught by Katerina.  Topics include supervised and unsupervised machine learning techniques and decision trees.  The second day will cover the construction of ego networks from Retweets and will be taught by Yotam.  He will go through the details of obtaining authorization from Twitter as well as taking into account the limits that Twitter imposes on how much data can be extracted from their servers within a short window of time. Instead of focusing on who follows whom on the Twitter platform, this method instead conceptualizes edges as existing between users if they have recently Retweeted each other, so it can be used to analyze even relatively popular users.  The third day will focus on clustering models, including PCA and nearest neighbors, and will be taught by Katerina.

Dr. Katerina Sinclair has been an enthusiastic R useR for over a decade. After initially using it to do things no other program could do, she eventually came to use it to do pretty much everything. She earned a concurrent Masters of Applied Statistics and a PhD in Human Development and Family Studies, with a focus on research methodology and statistics, from the Pennsylvania State University. She has worked in academia and industry and currently works as the Principal Data Scientist for UnitedHealth Group in an R & D department. Her past work included a stint at the U of A wherein she tried to convince every social science department to implement “R-only” policies. She eats data for breakfast and retains a steadfast commitment to improving people's lives by writing better open-source algorithms.​

Yotam Shmargad is an Assistant Professor at the University of Arizona’s School of Information. He received his Ph.D. in Marketing from Northwestern University’s Kellogg School of Management in 2014. In his research, he analyzes the databases of social media companies and designs network experiments to study how social media augment users’ networks, and the impact this has on how products and information spread through these networks.