Figure - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Model Specifications in Dishop (2022)
Source publication
Dishop (2022) identifies the consensus emergence model (CEM) as a useful tool for future research on emergence but argues that autoregressive models with positive autoregressive effects are an important alternative data-generating mechanism that researchers need to rule out. Here, we acknowledge that alternative data-generating mechanisms are possi...
Contexts in source publication
Context 1
... proposed methodology to address these issues, the CEM, is based on an extended two-or three-level mixed-effects model (see Model C1 in Table 1) with observations nested in unit members and unit members nested in units. The model includes a time trend to account for changes in the between-group variance over time. ...Context 2
... fully understand the Dishop (2022) article, we review the model specifications in each part of the article. Table 1 provides an overview of the models described in the article, appendix, and supplementary material. ...Context 3
... the text, Dishop (2022) correctly describes and discusses the CEM and provides the formula for the residual variance within the model on page 2 and in Table 1, e ijt N(0, σ 2 exp[2δTIME t ]). In this formula, negative values indicate consensus emergence and positive values indicate dissensus. ...Context 4
... (2022) indicates that this code was used to test Models A1 and A2; however, the code in the appendix shows that the models underlying Simulations 1-3 differ from Models A1 and A2 as described in the article. Table 1 provides the details of Models A6 and A7 using the nomenclature from Dishop (2022) and details how these models differ from Models A1 and A2. ...Context 5
... the specification in Table 1, it is clear that A6 and A7 are combined models that use a different specification for different points in time. The rationale behind the use of these models and the strong scaling parameters was not clear to us especially because Dishop (2022) explicitly states that α was set to 0.8 on page 3. The autoregressive effects are not scaled similarly to the other parameters resulting in models that appear to produce very small autoregressive effects-a point which we verify later. ...Context 6
... Question 1: Does fitting the CEM to simulated data from the random walk models (Models A1 and A2) in Dishop (2022) yield negative δ values (direct replication of Table D1 and Figure D1)? ...Context 7
... replication of Figure D4). Table D1, and in Figure D1 To answer whether fitting a CEM to simulated data from Models A1 and A2 in Dishop (2022) yields negative δ values, each author independently wrote a simulation based on the model description in the article (Models A1 and A2). Both independent simulations provided similar results. ...Context 8
... importantly, applying CEMs to these data yielded a positive δ of 0.12 reflecting the observed pattern of dissensus in Figure 1. Our results are substantially different from the reported estimates for δ of −0.48 (Table D1) and −0.82 ( Figure D1A) both in terms of magnitude and sign. ...Context 9
... revealed a negative δ effect of −0.38 that is in line with the results in the article (Table D1 and Figure D1). We next graphed the data and the predicted trend from the CEM. ...Context 10
... simulations are based off the models described in the text, no evidence of systematically negative δ values were observed (see Replications 1 and 2). Furthermore, we believe that the model specifications used in the supplemental materials (Models A6 and A7 in Table 1) are not likely to include notable autoregressive effects because any autoregressive effects will be small compared to the process error in the first and the individual differences components in the second model. ...Citations
Despite the prevalence of research on the consequences of collective turnover (TO), we lack an understanding of how, when, and why changes in the external environment influence collective turnover. The present study extends context emergent turnover and threat-rigidity theories to consider temporal changes in rates of collective turnover brought on by an external disruption. We also conduct variance decomposition to evaluate the relative influence of internal and external factors on collective turnover and examine how changes in the external environment impact relative influences. Finally, we examine the role of collective engagement in explaining patterns of collective turnover over time. Our study is based on a large, geographically dispersed U.S. firm. Findings from a two-phase longitudinal model reveal that rates of collective turnover change over time in ways that are predictable from threat-rigidity theory. Variance decomposition analysis finds that internal store-level factors explain substantially more variance than external factors, but the balance changes in response to an external disruption. We also show that collective engagement can mitigate increases in collective turnover. Results inform theory regarding the relative importance of internal versus external factors in influencing collective turnover and provide a framework for predicting how contextual change in the external environment impacts collective turnover over time.
