Centering predictor variables in cross-sectional multilevel models: a new look at an old issue.

By Craig K. Enders, Davood Tofighi. Psychological methods, 2007

4,105 citations recorded by the source corpus.

Subfields: Education, Radiological and Ultrasound Technology

View Paper: https://www.semanticscholar.org/paper/9b29601fd63805afffd2e826f4dec8e610db3024

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Education

In Radiological and Ultrasound Technology

Citations in the corpus

School Size, Social Capital, and Student Victimization (2011, methodology): “This centering is used because we have a substantive interest in the effects of Level 2 predictors on victimization, and it is necessary to adjust for between-school variance that arises because of the grouping of individuals with similar characteristics in schools (Enders and Tofighi 2007).”
The impact of the multi-channel retail mix on online store choice: : Does online experience matter? (2015, methodology): “Lastly, to disentangle whether the online buying experience effects are driven by betweenhousehold differences rather than within-household differences over time, we included a household mean of the online grocery buying experience in the model as an extra variable (based on Enders & Tofighi 2007).”
Environmental Income and Rural Livelihoods: A Global-Comparative Analysis (2014, background): “23 The rationale for doing this is that individual and aggregated indicators of site-level well-being may exert different and independent effects on environmental income and reliance outcomes (Enders & Tofighi, 2007).”
The Etiology of the Multilevel Paradigm in Management Research (2011, background): “…each higherlevel unit (group mean centering) or around the overall mean (grand-mean centering) changes the inherent interpretation of the RCM intercept and slope parameters, particularly in the case of modeling influences on lower-level slopes (see Enders & Tofighi, 2007; Hofmann & Gavin, 1998).”
Personal values and prosocial behaviour in strategic interactions: Distinguishing value‐expressive from value‐ambivalent behaviours (2013, background): “In sum, the results of our 1Group-mean centring means that we centred the scores of each of our values scales within samples—in linear mixed-model analyses, variables should be left uncentred only if they have a raw metric with a meaningful zero point (e.g. Enders & Tofighi, 2007).”
Within-person analyses of situational interest and boredom: Interactions between task-specific perceptions and achievement goals. (2014, background): “All predictors were group-mean centered, so ß0j represents participant j’s predicted interest when the participant has average scores for expectancy, utility, and difficulty (Enders & Tofighi, 2007).”
Contributions of Academic Language, Perspective Taking, and Complex Reasoning to Deep Reading Comprehension (2016, background): “This approach allows unbiased estimates of relations between variables at the individual level and produces the most accurate estimates of the slope variance (Enders & Tofighi, 2007; Raudenbush & Bryk, 2002).”
Belief in a Zero-Sum Game as a Social Axiom (2015, positive): “This type of centering has been recommended by Enders & Tofighi, 2007.”
Is This How I Will Be Treated? Reducing Uncertainty through Recruitment Interactions (2013, background): “Given our focus on within-person effects, we centered the time-varying predictors and mediator on each individual’s means (Enders & Tofighi, 2007) to remove any interindividual variance in the estimates of intraindividual effects.”
Classroom emotional climate, student engagement, and academic achievement (2012, background): “Either grand- or group-mean centering may be used to examine the differential influence of a variable at Levels 1 and 2 (Enders & Tofighi, 2007), but grand-mean centering is preferred when between-classroom effects are of theoretical interest (Preacher, Zyphur, & Zhang, 2010).”
Teachers’ emotions and emotional authenticity: do they matter to students’ emotional responses in the classroom? (2020, background): “In study 2, we investigated within-person relations by group-mean centering all variables (that is centering at the person mean, see also Enders & Tofighi, 2007) prior to analyses; thus, results denote relations within individuals (i.e. students) and reflect situational fluctuations in variables…”
CEO Succession Origin and Firm Performance: A Multilevel Study (2017, positive): “To analyze the data, we used a three-level hierarchical-linear-modeling (HLM) technique with all variables centered to the grand mean (Enders and Tofighi, 2007).”
The Joint Moderating Effects of Activated Negative Moods and Group Voice Climate on the Relationship between Power Distance Orientation and Employee Voice Behavior (2017, background): “For Hypotheses 2a and 2b, PDO and activated negative moods were group-mean centered, and group voice climate was grandmean centered (Enders & Tofighi, 2007).”