Exploring perceived innovativeness in Central America and the Caribbean: cross-level interactions of perceived camaraderie, organizational camaraderie climate, and organizational gender diversity

Abstract

We propose a cross-level perspective on the relationship between individual-level perceived camaraderie and organizational-level camaraderie climate which interact to predict employee perceptions of innovativeness. Additionally, organizational gender diversity weakens the cross-level interaction. We tested our hypotheses by conducting a multi-level study with 39,574 employees working in 53 firms in Central America and the Caribbean. The positive link between perceived camaraderie and perceived innovativeness was stronger in firms with higher organizational camaraderie climate, and this interaction was moderated in firms with more gender diversity. We discuss the international human resource management implications of social context for the organizational gender diversity and innovation literature in an often-overlooked region.

Keywords:

• Camaraderie climate
• gender diversity
• innovation
• social context
• multi-level

Acknowledgement

We thank Len Tevino, Jill Ellingson, and three Academy of Management reviewers for their valuable comments on the manuscript. Financial support from the Asociación Mexicana de Cultura, A.C. is gratefully acknowledged by Claudia Gonzalez Brambila.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.

Notes

1 Due to the large sample size and the low percentage of imputed values, we feel confident that the imputation did not impact the robustness of our results. Further, post-hoc we confirmed that there were no statistically different findings between the imputed and the original data set (with listwise deletion) for the hypothesized relationships.

2 As the scales have not all been previously established and confirmed in academic literature, we conducted an EFA on the perceived camaraderie and perceived innovativeness scales. Using PCA Varimax rotation procedures (2 factors), we found that 9 items were loading significantly on the perceived camaraderie scale and 5 items were loading significantly on the perceived innovativeness scale (loadings of 0.6 and higher on the rotated component matrix). Two items (1 from each scale), however, were deleted because of cross-loadings; the deletion of the items was in line with deletion suggestions made by the reviewers and experts in the field because of face validity issues. Following the EFA, we conducted a CFA to confirm the factor structure. The analysis demonstrated that a two-factor model with perceived camaraderie and perceived innovativeness as separated but correlated factors (χ² [53] = 4435.99, p < 0.05, RMSEA = 0.05, CFI = 0.99, GFI = 0.98) was superior to a one-factor model with both scales collapsed (χ² [54] = 18,238.46, p < 0.05, RMSEA = 0.09, CFI = 0.98, GFI = 0.93), testifying to the distinctiveness of the two scales (Δχ² [1] = 13,802.47, p < 0.001). Internal consistency of the constructs was confirmed as the AVEs exceeded the threshold of 0.50 (Fornell & Larcker, 1981).

3 In line with suggestions by MacKenzie and Podsakoff (2012), the sample had the cognitive ability to respond to the questions, had experience thinking about the topic, questions were phrased in a clear and concise manner, the survey was short avoiding responders’ fatigue, and scale items were randomized to avoid monotony. Additionally, the survey was voluntary and conducted by a company external to the employees’ organization. It has personal relevance to them and thus they were motivated to respond honestly since the aggregated results can improve their workplace.

4 Prior research has empirically demonstrated that (1) CMV severely deflates rather than inflates interaction effects and (2) that interaction effects cannot be artificially created by CMV if they do not exist (Evans, 1985; Siemsen et al., 2010). Similarly, multi-level researchers established that CMV ‘does not pose a threat to the validity of significant cross-level interaction effects’ as CMV tends to suppress the identification of a true cross-level interaction (Lai et al., 2013, p. 264). In line with their findings, our multi-country research design already accounts for a potential remedy of CMV reducing our cross-level interaction effects as we use data from geographically dispersed organizations to increase the variance of organization-level predictors.

5 A variety of methods were used to ensure data quality resulting in a total of 68 participants being removed from the original N = 300 sample. 42 participants were removed based on Qualtric’s RelevantID Fraud and Duplicate Scores which flag likely fraudulent bot and duplicate responses (Qualtrics, 2020). Based on the ‘unusual comments’ validity indicator (Chmielewski & Kucker, 2020), 16 participants were removed due to their responses to an open-ended question on the industry their employing organization is in. Examples of disqualifying responses included ‘yes’, ‘no’, ‘none’, ‘USA’, and instances where Amazon/MTurk were indicated as the industry. One participant was removed due to completing the survey the day after other participants. Finally, 9 outliers were removed that were ±2 SD of the mean survey response time, given that short or long response times can indicate careless responding (Ward & Meade, 2018).

6 While our primary goal was to explore the organization-level effects on an individual-level relationship, prior secondary data has shown the similarity between the CAC countries on institutional, economic, and cultural factors, and the sample size of 8 on Level 3 is considered too small to model hierarchically (Peterson et al., 2012); we considered it as important to demonstrate that country-level differences did not significantly affect our overall results.