International Journal of Innovation Management and Organizational Behavior (IJIMOB)

Objective: The present study aimed to investigate the effect of similarity bias on auditors’ professional judgment and to examine the moderating role of self-esteem in this relationship.

Methods and Materials: This study was applied research in terms of purpose and descriptive-survey research in terms of data collection method. The statistical population consisted of auditors employed by the Audit Organization and audit firms affiliated with the Iranian Association of Certified Public Accountants. Based on the Krejcie and Morgan sampling table, 139 auditors were selected as the study sample. Data were collected using standardized questionnaires related to similarity bias, auditor professional judgment, and self-esteem. Similarity bias was examined through two questionnaires distributed with a one-month interval, differing only in the similarity-bias scenario. Auditor professional judgment was measured using the Jenkins and Haynes questionnaire utilized in the study by Puspa (2008), while self-esteem was measured using the Rosenberg Self-Esteem Scale. Data analysis was performed using descriptive and inferential statistics, including Cronbach’s alpha reliability analysis, AVE and HTMT validity tests, Kolmogorov–Smirnov normality testing, and structural equation modeling for hypothesis testing.

Findings: The findings demonstrated that similarity bias had a significant negative effect on auditors’ professional judgment (β = -0.361, t = 4.574, p < 0.001), indicating that increased similarity bias reduces the quality and objectivity of professional judgment. In contrast, self-esteem had a significant positive effect on auditor professional judgment (β = 0.271, t = 3.435, p = 0.001), suggesting that auditors with higher self-esteem demonstrate more accurate and confident professional judgments. However, the moderating effect of self-esteem on the relationship between similarity bias and auditor professional judgment was not statistically significant (β = -0.038, t = 0.822, p = 0.411). Therefore, the moderating hypothesis was rejected.

Conclusion: The results indicate that cognitive biases, particularly similarity bias, can significantly impair auditors’ professional judgment and potentially threaten audit quality and objectivity. At the same time, self-esteem contributes positively to auditors’ decision-making quality and professional confidence.

Objective: This study aims to predict employee innovative work behavior by evaluating the distinct and interactive contributions of psychological safety and proactive personality through the application of advanced machine learning algorithms.

Methods and Materials: A cross-sectional, quantitative design was utilized, drawing a stratified random sample of full-time employees from various dynamic corporate sectors across Indonesia. Data were collected via self-administered, translated questionnaires utilizing established Likert-scale instruments. To evaluate the predictive relationships, the dataset was partitioned into an training and testing split. Four machine learning models—Multiple Linear Regression, Support Vector Regression, Random Forest Regressor, and Gradient Boosting Regressor—were trained and evaluated using five-fold cross-validation, while SHapley Additive exPlanations (SHAP) values were computed to interpret feature importance and non-linear interactions.

Findings: The non-linear ensemble models outperformed traditional linear approaches, with the Gradient Boosting Regressor demonstrating the highest predictive accuracy on the testing set ( , , ). Feature importance analysis using SHAP values identified proactive personality as the dominant predictor of innovative work behavior (mean absolute SHAP value = ), closely followed by psychological safety (mean absolute SHAP value = ). Furthermore, the dependency plots revealed a critical non-linear interaction, indicating that high levels of proactive personality only consistently translate into innovative behavior when a baseline threshold of psychological safety is present in the work environment.

Conclusion: Organizations must concurrently hire for proactive traits and cultivate psychologically safe climates, as both are fundamentally intertwined in actualizing and predicting employee innovation.

Objective: This study aimed to identify distinct profiles of employee innovation behaviors using unsupervised machine learning and to evaluate the predictive roles of future time perspective and proactive motivation in determining membership within these empirically derived behavioral clusters.

Methods and Materials: A quantitative, cross-sectional design was utilized to collect data from a purposive sample of professionals employed in various knowledge-intensive sectors across Spain. Participants completed validated self-report questionnaires assessing three core dimensions of innovation behavior (idea generation, idea promotion, and idea realization), alongside measures of future time perspective and proactive motivation. The analytical framework employed a K-means clustering algorithm to partition the multidimensional innovation scores into distinct profiles, followed by multinomial logistic regression models to evaluate the predictive capacity of the psychological variables in determining cluster categorization.

Findings: The K-means clustering analysis ( ) successfully identified three distinct behavioral profiles: “Passive Innovators” ( , ), “Moderate Innovators” ( , ), and “High Innovators” ( , ). The multinomial logistic regression model demonstrated robust predictive power (Nagelkerke ). A one-unit increase in future time perspective significantly increased the odds of an individual belonging to the High Innovators cluster rather than the Passive Innovators cluster ( , ). Furthermore, proactive motivation emerged as an exceptionally strong differentiator; a one-unit increase in proactive motivation drastically elevated the likelihood of being classified as a High Innovator compared to the passive reference group ( , ).

Conclusion: Fostering a future-oriented mindset and cultivating intrinsic proactive motivation are critical psychological catalysts for transforming passive employees into highly engaged innovators capable of driving ideas from initial conception to full realization.

Objective: The objective of this study was to predict long-term employee creativity trajectories by developing a longitudinal deep learning model integrating evolving motivational states and job design characteristics.

Methods and Materials: This longitudinal study followed 611 full-time employees from knowledge-intensive organizations in Spain across four measurement waves over 18 months. Repeated measures of intrinsic and extrinsic motivation, psychological empowerment, and job design characteristics (including autonomy, task significance, skill variety, and feedback quality) were collected using validated instruments. Employee creativity was assessed using a multi-source composite index incorporating self-ratings, supervisor evaluations, and peer nominations. Long Short-Term Memory (LSTM) neural networks were employed to model non-linear temporal relationships and predict individual creativity trajectories. Model performance was evaluated using root mean square error, mean absolute error, explained variance, and trajectory similarity indices. Feature importance was examined using attention mechanisms and SHAP values, and unsupervised clustering was applied to identify distinct creativity development profiles.

Findings: The deep learning model explained 76.1% of variance in unseen creativity trajectories. Intrinsic motivation (β_importance = 0.31, p < .001) and job autonomy (β_importance = 0.24, p < .001) emerged as the strongest predictors of sustained creative growth. Four statistically distinct creativity trajectories were identified: accelerated growth (34.5%), stable growth (41.2%), stagnant (16.7%), and declining (7.6%). Employees in the accelerated growth cluster demonstrated significantly higher increases in intrinsic motivation and job autonomy over time (p < .001), whereas the stagnant and declining clusters exhibited significant decreases in psychological empowerment and increases in role overload (p < .01).

Conclusion: Employee creativity is a dynamic developmental capability shaped by continuous interactions between motivation and job design, and longitudinal deep learning provides a powerful framework for predicting creative growth and identifying intervention targets.

Objective: The objective of this study was to deploy a Random Forest machine learning algorithm to evaluate the predictive power and complex non-linear interactions of psychological diversity and team emotional intelligence on collective team creativity.

Methods and Materials: A cross-sectional quantitative research design was employed, collecting data from employees nested within established work teams in the Greek corporate sector. Team creativity was assessed via supervisor ratings, while emotional intelligence and psychological diversity (measured as the standard deviation of Big Five personality traits within teams) were self-reported and statistically aggregated to the team level. Data analysis utilized Random Forest regression, comparing its predictive performance against traditional Multiple Linear Regression using an training set (  teams) and a testing set (  teams), with hyperparameters optimized via grid search cross-validation.

Findings: The Random Forest model significantly outperformed Multiple Linear Regression in predicting team creativity on the testing set ( , , versus , , ). Variable importance analysis revealed that Team Emotional Intelligence was the paramount predictor (Importance Score , , ). This was followed by psychological diversity in Openness (Importance Score , , ) and Extraversion (Importance Score , , ), which both positively correlated with creative output. Conversely, diversity in Conscientiousness (Importance Score , , ) demonstrated a negative impact on team creativity.

Conclusion: High aggregate emotional intelligence and specific deep-level personality diversities interact in highly non-linear patterns to drive team innovation, underscoring the necessity of advanced ensemble learning techniques for accurate organizational and behavioral modeling.

Objective: The objective of this study was to autonomously identify latent patterns of organizational learning among corporate professionals and examine the differentiating roles of knowledge sharing and cognitive flexibility within these profiles.

Methods and Materials: A descriptive, quantitative cross-sectional design was employed, collecting data from a sample of corporate professionals in Egypt via a validated, self-administered online questionnaire. The analytical framework utilized unsupervised machine learning, specifically Principal Component Analysis for dimensionality reduction and K-means clustering to extract distinct learning profiles. Subsequently, Multivariate Analysis of Variance (MANOVA) and Tukey’s post-hoc tests were conducted to evaluate how knowledge sharing and cognitive flexibility differed across the autonomously generated clusters.

Findings: The K-means clustering algorithm identified an optimal three-cluster solution ( ): a Passive Learning Profile ( , ), an Adaptive Learning Profile ( , ), and a Proactive Learning Profile ( , ). The MANOVA results revealed a highly significant omnibus effect across the clusters, Wilks’ , . Univariate and post-hoc analyses demonstrated that individuals in the Proactive Learning Profile exhibited significantly higher levels of knowledge sharing ( ) and cognitive flexibility ( ) compared to those in the Adaptive Profile (  and , respectively; ) and the Passive Profile (  and , respectively; ).

Conclusion: Fostering a highly proactive and generative organizational learning environment fundamentally depends on deliberately cultivating both interpersonal knowledge exchange networks and individual cognitive adaptability.

Objective: The objective of this study was to elucidate the temporal dynamics of organizational innovation adoption by integrating empirical psychological metrics of risk perception and change readiness with a computational reinforcement learning framework.

Methods and Materials: This study utilized a cross-sectional design, sampling mid-to-senior level managers from the corporate sector in South Africa. Primary data regarding individual risk perception, baseline change readiness, and behavioral adoption intent were collected utilizing a structured, validated online questionnaire. The data analysis phase employed a novel hybrid methodology: inferential statistics via Structural Equation Modeling (SEM) were integrated with a mathematical Reinforcement Learning (RL) architecture. Specifically, a -learning algorithm was developed where the empirical psychometric scores directly parameterized the internal variables of synthetic agents, mapping readiness to the learning rate ( ) and risk perception to the environmental penalty weight ( ) and temporal discount factor ( ).

Findings: The empirical baseline analysis revealed a strong positive correlation between change readiness and the final innovation adoption rate ( ), while risk perception exerted a severe suppressive effect on the ultimate probability of adoption ( ). When mapped to the computational simulation, the fully optimized -learning model significantly outperformed static predictive models, successfully accounting for of the variance in actual innovation adoption trajectories ( ). The algorithmic simulation accurately demonstrated that agents with high change readiness exhibited an accelerated , quickly escalating their expected utility for adoption, whereas high risk perception effectively delayed transition by mathematically magnifying the operational penalty ( ) and minimizing the perceived future value of the new technology ( ).

Conclusion: The findings provides a highly accurate, dynamic computational model that predicts the temporal evolution of human behavioral adaptation during organizational technological transitions.

Objective: The primary objective of this study was to utilize an advanced artificial neural network approach to investigate the complex, non-linear predictive effects of work meaningfulness and job autonomy on radically innovative behavior among professional employees.

Methods and Materials: A quantitative, cross-sectional design was employed, utilizing a purposive sample of  Polish professionals operating within the high-technology, engineering, and manufacturing sectors. Data were gathered using adapted, standardized psychometric instruments measured on a five-point Likert scale, with all measures demonstrating robust internal reliability (Cronbach’s ). To accurately model intricate, non-linear behavioral relationships without relying on restrictive parametric assumptions, the data were analyzed using a Multilayer Perceptron artificial neural network. The dataset was systematically partitioned into training ( ), testing ( ), and holdout ( ) subsets to rigorously validate the computational model’s predictive accuracy and isolate predictor strength via sensitivity analysis.

Findings: The descriptive analysis confirmed that the sample ( ) reported significant positive correlations among all variables ( ). The artificial neural network demonstrated excellent model fit and explanatory power, successfully accounting for approximately of the total variance in radically innovative behavior ( ). The normalized sensitivity analysis revealed that job autonomy is the paramount predictor of radical innovation, yielding an absolute importance of and a normalized importance of . Work meaningfulness emerged as a highly critical secondary predictor, demonstrating an absolute importance of and a normalized importance of .

Conclusion: Structural job autonomy is the absolute foundational prerequisite for enabling radically innovative behavior, while the cultivation of work meaningfulness provides the indispensable psychological drive and intrinsic resilience required to sustain these complex, paradigm-shifting efforts.