On this page I am happy to share some methodological resources that may be potentially useful to researchers interested in the respective techniques.
Mobile Experience Sampling Software (SurveySignal)
Experience sampling is quickly gaining popularity as a research method (see Mehl & Conner, 2012). However, many of the existing methods require a high technical proficiency. In search for an easy and user-friendly solution, we developed SurveySignal, a web-based tool that facilitates experience sampling and other smartphone research. SurveySignal marries together three basic features useful for experience sampling projects and other smartphone research:
(1) A simple, foolproof, and fast signup system that allows
people to register their own smartphone and provide informed
consent, (2) the use of cellphone text messaging as a signaling
device according to random, fixed, or mixed schedules, and (3)
the benefits of highly developed and easy-to-use survey software
such as Qualtrics or SurveyMonkey with which most researchers
are already familiar.
Here's how it works: The text messages contain individualized
links that take participants to mobile-compatible surveys as
created through Qualtrics, Surveymonkey, or any other survey
software (thus eliminating the need to program the surveys) and
transport further information such as the send and response time
of the signal. A beta version of SurveySignal has been piloted
successfully in a number of experience sampling projects
conducted at the University of Chicago and in collaboration with
researchers from Duke University, Northwestern University, Wake
Forest University, McGill University, Canada, and the University
of Basel, Switzerland. SurveySignal is available for anyone
interested in using it. Registering an account and exploring the
system is free, but we charge a small amount per signals used in
order to cover our development, operating, and support costs.
Click HERE for a short presentation given at the
2012 Society for Experimental Social Psychology symposium "The
future is now: How smart phones are transforming psychological
research" organized by Maarten Bos.
Slides on Implicit Measures in Attitude and Personality Research
HERE, you can find a presentation on "the
use of implicit measures in attitude and personality research" I
gave at the 2012 SPSP GSC and Training Committee Innovative
Methods pre-conference organized by Marina Milyavskaya and
Stacey Sinclair. The presentation is intended to provide a
hands-on introduction to implicit measures as research tools
while at the same time placing these measures into a theoretical
and conceptual context.
Workshop on Moderated Regression
HERE, you can find powerpoint slides from a workshop on moderated regression I gave at the 2008 EASP summer school in Cardiff. It provides a beginner-oriented overview of how to conduct and interpret moderated regression analyses. Specifically, the workshop covers prototypical cases such as continuous variable by nominal scale interactions and interactions among two continuous predictors. It also addresses questions such as how to compute regression syntax and interpret output from a statistical software program such as SPSS, how to illustrate interaction effects graphically (in EXCEL), how to conduct simple slope tests, how to code nominal variables, and how to calculate effect sizes for the interaction effects. Click HERE for a zip-file containing the accompanying SPSS example data, syntax, and EXCEL slides used to illustrate different types of interaction effects. For a quicker and more user-friendly way to plot interactions in SPSS see the macro below.
Interaction Macro Extension
This is an extension of the MODPROBE macro (Hayes, 2009) for probing single-df interactions in OLS and logistic regression models in SPSS. The extension (Hofmann, 2010) includes a variety of options for graphical inspection, uni- and multivariate outlier detection, and probing. Main features:
• Request unstandardized, half-standardized, or
fully-standardized interaction plot ready for export
• Plot predicted log-odds or probabilities for logistic regression analysis
• Test and plot at different values of the moderator
• Specify any range of the predictor variable for the plot
• Conduct univariate and/or multivariate outlier exclusion using a variety of statistics such as studentized residuals, Mahalanobis distance, leverage, dffits, or Cook’s Distance
• Check for normality
• Include additional filter variable
Example of SPSS Interaction Plot
Extension of Meta-Analysis Macros by David Wilson
The goal of this extension was to make the useful Anova and Regression macros by David Wilson (see Lipsey & Wilson, 2001) somewhat more flexible with regard to a number of data handling and analysis strategies (e.g., outlier exclusion etc.). The computational “core” of the macros by D. Wilson has been left untouched. However, I added some features (click HERE for a more detailed description) that should make the use of both macros easier and less time-consuming. The extension may be particularly helpful (a) in meta-analyses that involve a larger number of moderators or (b) when the researcher is interested in conducting sensitivity analyses to check how certain variations in the meta-analytic procedure (e.g., fixed vs. random effects, outlier inclusion vs. exclusion) affect the outcome of the meta-analysis. Furthermore, the results will be summarized in .sav and .xls output files. These can be conveniently fed into an Excel-spreadsheet designed to present the results in pre-arranged tables and graphs ready to be exported into other applications.
Download macro package HERE.