A Tutorial for Exploratory Research: An Eight-Step Approach

Abstract

Currently in psychological science considerable effort is directed towards confirmatory practices. Much less attention has been devoted to how to do exploratory research. In this article, we support researchers in expanding their methodological toolbox by adding one more technique of exploratory research. The majority of this article is a hands-on tutorial that explains how exploration can be done using state-of-the-art statistical methods, ultimately leading to an in-depth demonstration of machine learning techniques. The practical part of this tutorial explores one of our own datasets, the Human Penguin Project (IJzerman, Lindenberg et al., 2018). The reader can follow the tutorial by recreating our analyses in their own RStudio, apply our annotated code to her own data or other secondary data, and repeat our steps. We show how to get familiar with datasets the researcher wants to use for machine learning, inspect it in many useful ways, and make predictions using machine learning algorithms. We close with describing the limitations related to causal inference and clarifying that finding robust patterns does not equate generating a comprehensive theory. Our tutorial requires basic knowledge of statistics and programming language R (R Core Team, 2016), but we provide resources for absolute beginners.

Stress Regulation via Being in Nature and Social Support in Adults, a Meta-analysis

“…There are reasonable solutions to such "small N, large P" problems. We recommend to use machine-learning methods to 1) reduce potential confounds by adding predictors and 2) to better extract patterns from data by using random forests to identify the most important predictors of the projected outcome (cf., Wittmann et al, 2021; for a tutorial, see Szabelska et al, 2022). Once identified, researchers can generate relatively precise predictions to be tested in either longitudinal designs or experimental studies to bolster the strength of the causal inference.…”

“…For this, we have two recommendations. First, like for being in nature, it is possible to improve the quality of causal inferences by surveying participants on a host of additional (potentially relevant) variables (such as neuroticism, attachment security, the quality of one's social network, and so forth), and thereafter applying analytic techniques that are less prone to overfitting and to problems with collinearity (such as conditional random forests; Szabelska et al, 2022). Second, one could attempt to improve the quality of emotional social support from a support-giver by letting dyads participate in relationship-focused therapy (see, for instance, Johnson et al, 2013) .…”

A Manifesto for More Productive Psychological Games Research

“…In sum, it is crucial for the progress-oriented games researcher to think carefully about the use of theory, and the specificity of predictions. I encourage readers to read more carefully about the relationship between theory, hypothesis, and model (e.g., Scheel et al 2021b;Guest and Martin 2021), review the basics of causal modeling (Rohrer 2018), and consider conducting more explicitly descriptive or exploratory studies (for a tutorial, see e.g., Szabelska et al 2021; for varied examples, see e.g., Festic et al 410 2021; Karakus et al 2008;DeCamp 2017).…”

A meta-scientific approach to understanding stress regulation

“…(2013), surveyed groups walking in different environments on mental well-being, depression, and self-reported stress. In such a case, we recommend authors to use experimental methods to bolster the strength of causal inferences and/or to reduce potential confounds by adding predictors and using conditional random forests as an analysis method to identify the most important predictors of the projected outcome (cf., Wittmann et al, 2021;Szabelska et al, 2022).…”

“…Szabelska et al, 2022). Second, one could attempt to improve the quality of emotional social support from a support-giver by letting dyads participate in relationship-focused therapy (see, for instance,Johnson et al, 2013) and comparing this with an active and passive control condition.More generally, to better map how emotional social support (if at all) and being in nature affect stress (in experimental or cross-sectional studies), we think that a much better record is needed.…”