Measuring Connection at Work: Development and Validation of the Organisational Connection Measure ()
1. Introduction
Research has long supported the relationship between employee perceptions and business outcomes such as employee performance (e.g., Christian et al., 2011; Harter et al., 2002; Judge et al., 2001). Indeed, organisations are increasingly reliant on surveying staff to identify ways to improve employee experience and so increase productivity and decrease attrition (WorkBuzz, 2022). However, current models of employee perceptions may lack breadth and overlook crucial elements of modern work that drive employee behaviour.
Employee engagement has a long-established relationship with individual performance (e.g., Christian et al., 2011). Measured through well-regarded and extensively validated measures such as the Utrecht Work Engagement Scale (UWES; Schaufeli et al., 2002), engagement is typically concerned with aspects such as individuals’ energy, involvement, and dedication to their specific role. A recent meta-analysis found engagement to be correlated with factors such as turnover intention, performance, and wellbeing (Mazzetti et al., 2023). However, engagement does not capture employees’ perceptions of their social relationships at work. Thriving is another construct that has been applied to work (e.g., Porath et al., 2012) and found to be related to important organisational outcomes such as task performance (TP; see Kleine et al., 2019). Thriving does cover social aspects of work, such as feeling valued, supported, and heard by coworkers (Peters et al., 2021); however, these insights merely provide an understanding of the satisfaction of an individual with their perceptions of how others behave around them, while not capturing the extent to which an individual feels part of and connected to a collective at work. Satisfaction with one’s relationship with coworkers and the broader psychosocial work environment has been found to be related to work performance (Shahid et al., 2011), counterproductive work behaviour (CWB; Agervold, 2009), and manager derailment (Gentry et al., 2007). In overlooking these aspects of employees’ social experiences at work, existing employee experience models lack insights into perceptions that are arguably very important to them. As a result, organisations developing interventions based on insights from existing models may be overlooking features of work that could potentially represent the greatest potential for positive change.
This paper makes the case for a novel model and measure of employee perceptions, the Organisational Connection Measure (OCM), developed to be highly predictive of factors relevant for organisational success, by first outlining its broad theoretical foundations and rationale before documenting the approach to model and measure development and, finally, presenting initial validation research.
Connection
Research across psychology and organisational science consistently positions connection as a fundamental human need for belonging and relatedness (Maslow, 1943; van Bel et al., 2009) which fosters experiences of affiliation, acceptance, and meaningful social ties that support wellbeing and behaviour (Allen et al., 2021; Baumeister & Leary, 1995; Ryan & Deci, 2000). Prior work characterises social connectedness as a subjective sense of being part of significant, supportive relationships and of having social ties that feel salient in daily life (Allen et al., 2021; van Bel et al., 2009). More broadly, connectedness has been described as the extent to which people experience their relationships and environments as personally meaningful and important, highlighting the appraised significance of social bonds rather than their mere presence (Allen et al., 2021; Schulze & Naidu, 2014).
Building on this foundation, Holt-Lunstad (2018a) proposed that social connection encompasses multiple dimensions, including the structure of a person’s social network, the functions relationships serve, and the emotional quality of those interactions. These aspects capture not only the presence of social ties, but also the perceived availability of support and the positive or negative nature of interpersonal exchanges; all of which have been shown to contribute to positive health and behavioural outcomes.
In the present work, workplace connection is understood as individuals’ perceptions of their connection to their role, their organisation, and the people around them at work. It reflects how supported and included a person feels in their work environment. Importantly, connection is viewed as measurable and malleable: Once understood, it can be strengthened through intentional behaviours and supportive organisational practices (Holt‑Lunstad, 2018b, 2024). This perspective underpins the development of the OCM, which seeks to capture meaningful variations in employees’ experiences of connection at work.
In their meta-analysis on the impact of social relationships on mortality, Holt-Lunstad et al. (2015) found that a lack of connection (i.e., social isolation or loneliness) has a significant detrimental effect on an individual’s wellbeing and health and increases risk of chronic illnesses, mental health issues, and mortality. In fact, weak social ties can increase the risk of early death by 26% - 32%, which is comparable to other well-known risk factors such as heavy smoking, obesity, and physical inactivity (Holt-Lunstad et al., 2015). In the workplace, a lack of connection is prevalent and impacts people’s lives inside and outside of work, and research has supported that fostering greater connection can help organisations to ensure their staff are healthy and productive (Holt-Lunstad, 2018b).
The development of this measure was informed by a desire to create a broad measure of aspects of people’s perceptions of their work life that ultimately influence business performance and success. Through this, it is anticipated that this measure will arm organisations with useful insights to improve their employees’ perceptions to lead to better business outcomes. As such, criterion validity was the primary concern when developing this measure, rather than a desire to align with established models.
Existing Literature
As part of the development of the OCM, a literature review was conducted to understand the domains of employee experience relevant to connection that ultimately contribute to business success. Papers have been included if they relate connection to aspects of individual performance, group performance, or retention. Individual performance includes subjective and objective performance, underperformance, derailment, and counterproductive work behaviour. Group performance includes aspects related to organisational performance, team productivity, or team cohesion. Retention includes related factors such as intention to quit. The specific interest of this model is in changeable states related to connection rather than stable psychological traits such as personality, so the model exclusively measures constructs that organisations can endeavour to improve, which theoretically should lead to greater organisational success. A decision was also made to overlook aspects relating to perceptions towards specific individuals, such as managers or leaders, as these aspects would be better captured through activities such as manager and leader appraisals.
In all, 27 different constructs that relate connection to performance or retention were identified in the literature review. Papers highlighting the predictive validity of these constructs in the workplace are outlined in Table 1 at the end of this section. The constructs are presented in three broad themes (role connection, organisational connection, and interpersonal connection) merely for ease of reading, although there is significant overlap between themes. This review aims to justify the inclusion of these specific constructs in the OCM model, rather than serve as a comprehensive systematic review of existing literature. Excellent reviews of this nature already exist; for example, Rovetta et al. (2025) on team and organisational identity, Istiningtyas et al. (2025) on social embeddedness of thriving at work, Blau et al. (2023) on organisational belonging, Iovoli et al. (2025) on interpersonal problems and mental health outcomes, and Pietromonaco and Collins (2017) on close relationships and health.
Role Connection
There are ten constructs that have been categorised under role connection, namely clarity, autonomy, burnout, recognition, wellbeing, stress/pressure, work-life balance, perceived productivity, job satisfaction, and motivation. These are constructs that pertain to how connected individuals feel to the actual job they are performing, how able they are to do this job, and the impact it has on them. They were particularly related to factors relevant to individual performance.
Organisational Connection
This category comprises six constructs related to how individuals feel about their connection to the organisation they work for. These are: compensation, organisational commitment, organisational support, organisational justice, organisational identity, and work environment. Many of these factors are particularly important considering employee retention and intention to quit.
Interpersonal Connection
Eleven constructs have been categorised as part of interpersonal connection, namely trust, psychological safety, team identity, social connection, collaboration, conflict, peer support, communication effectiveness, size of network, social roles, and satisfaction with relationships. These constructs broadly pertain to how an individual feels about the social aspects of work and how connected they feel to those around them. These constructs were especially relevant to factors related to team and group performance. For simplicity’s sake, trust has been listed as a single construct; however, findings encompass multiple dimensions, including interpersonal and organisational trust.
Table 1. Constructs relating aspects of connection to relevant outcomes in the literature.
Category |
Related Outcome |
Construct (Predictor) |
References |
Role Connection |
Underperformance |
Lack of clarity |
|
|
|
Low job autonomy |
Van den Heuvel et al., 2010 |
|
|
Burnout and excessive work demands |
Aboagye et al., 2019; Dyrbye et al., 2019;
Shahid et al., 2011; Van den Heuvel et al., 2010 |
|
|
Lack of recognition |
Hlengane & Bayat, 2013 |
|
|
Stress/pressure |
Pflanz & Ogle, 2006;
Shahid et al., 2011 |
|
|
Poor work-life balance |
Shahid et al., 2011 |
|
|
Lack of motivation |
Hlengane & Bayat, 2013 |
|
Counterproductive work behaviours |
Lack of job autonomy |
Browning, 2008; Marcus & Schuler, 2004; Penney et al., 2003; Vardi & Weitz, 2003 |
|
|
Burnout and high workload |
Baillien et al., 2011; Blau & Andersson, 2005 |
|
|
Stress/pressure |
Agervold, 2009; Hauge et al., 2009; Martinko et al., 2005; Vardi & Weitz, 2003 |
|
|
Poor work-life balance |
Scott et al., 2015 |
|
|
Low job satisfaction |
Hollinger & Clark, 1982; Mangione & Quinn, 1975; Spector & Fox, 2002 |
|
Derailment |
Perceived productivity |
Van Velsor & Leslie, 1995 |
|
Retention or intention to quit |
Clarity |
Ghosh et al., 2013; James et al., 2011 |
|
|
Autonomy |
Koyuncu et al., 2006 |
|
|
Burnout |
Malik et al., 2013 |
|
|
Recognition |
James et al., 2011;
Koyuncu et al., 2006;
Nasir et al., 2019 |
|
|
Work-life balance |
Nasir & Mahmood, 2016; Rasdi et al., 2018 |
Continued
|
|
Job satisfaction |
De Sousa Sabbagha et al., 2018; Lyu et al., 2022; McGilton et al., 2013; Nasir & Mahmood, 2016 |
|
|
Motivation |
De Sousa Sabbagha et al., 2018; Sarmad et al., 2016 |
Organisational Connection |
Performance |
Compensation |
Hlengane & Bayat, 2013 |
|
|
Organisational justice |
Hlengane & Bayat, 2013 |
|
|
Organisational support |
Hlengane & Bayat, 2013 |
|
Counterproductive work behaviour |
Organisational commitment |
Mathieu & Zajac, 1990 |
|
|
Perceived organisational support |
Alias & Rasdi, 2015; Ones et al., 2003 |
|
|
Organisational justice |
Akinsola & Alarape, 2019; Hershcovis et al., 2007; Jones & Martens, 2009 |
|
Retention and
intention to quit |
Perceptions of one’s work environment |
Bibi et al., 2017; Kundu & Lata, 2017; Nasir & Mahmood, 2016;
Warraich et al., 2019 |
|
|
Compensation |
Hussain & Rehman, 2013; Koyuncu et al., 2006;
Nasir et al., 2019; Sarmad et al., 2016 |
|
|
Organisational
commitment |
Atif et al., 2011; Nasyira et al., 2014 |
|
|
Organisational support (including through training and development) |
Haider et al., 2015; James et al., 2011; Janjua &
Gulzar, 2014; Nasyira et al., 2014; Saleem &
Affandi, 2014; Warraich et al., 2019 |
|
|
Perceptions of and alignment with
organisational identity and culture |
Haider et al., 2015; Hussain & Rehman, 2013; Matongolo et al., 2018 |
|
Team productivity |
Organisational support |
Salas et al., 2015 |
|
Organisational
performance |
Organisational support (through improved internal auditing and learning) |
Carmeli & Zisu, 2009 |
Interpersonal Connection |
Individual
performance and underperformance |
Poor communication |
Hlengane & Bayat, 2013 |
Continued
|
|
Satisfaction with one’s relationships with coworkers |
Shahid et al., 2011 |
|
Counterproductive work behaviour |
Trust in the organisation |
Thau et al., 2007 |
|
|
Social connection |
Agervold, 2009 |
|
|
Interpersonal conflict |
Chen & Spector, 1992; Frone, 1998; Hauge et al., 2009 |
|
Workplace derailment |
Satisfaction with one’s relationships |
Gentry et al., 2007;
Lombardo & McCauley, 1988; Van Velsor & Leslie, 1995 |
|
Retention and intention to quit |
Psychological safety |
Kim et al., 2021 |
|
|
Collaboration and conflict |
Belias et al., 2023 |
|
|
Communication effectiveness |
Hussain & Rehman, 2013; Steiner et al., 2020 |
|
|
Social roles |
Steiner et al., 2020 |
|
Team productivity |
Interpersonal trust |
Jarvenpaa & Leidner, 1999; Jarvenpaa et al., 2004 |
|
|
Collaboration |
Stashevsky & Koslowsky, 2006 |
|
|
Psychological safety |
Bergmann & Schaeppi, 2016 |
|
Team cohesiveness and productivity |
Organisational and interpersonal trust |
Cohen et al., 1996;
Hansen et al., 2002; Mach et al., 2010 |
|
|
Team identity and climate |
Casey-Campbell & Martens, 2009; Chen et al., 2008; Daspit et al., 2013; Hogg & Hardie, 1992 |
|
|
Collaboration |
Wang et al., 2010; West et al., 2009 |
|
|
Positive conflict |
Sullivan & Feltz, 2003; Sullivan & Shorts, 2011 |
|
|
Peer support |
Mäkikangas et al., 2017 |
|
|
Effective communication |
Holt & Sparkes, 2001; Mesmer-Magnus & DeChurch, 2009; Sullivan & Short, 2011 |
|
|
Size of network |
Schulte et al., 2012 |
Continued
Initial Development Process Preceding the Current Study
The current study presents the psychometric qualities of the novel OCM. The aim was to develop a psychometrically robust tool that can predict positive workplace outcomes. An iterative development process was used to refine the measure and ensure a balance between psychometric rigour and practical utility. Three studies were completed as part of the development of the OCM, with this paper being concerned with the final study demonstrating the validity and reliability of the final assessment. Each study built upon the last with a similar methodology but distinct occupational samples.
The first two studies have been documented in detail elsewhere (Thomas, 2025) and are summarised below as they relate to the present study. The initial item pool consisted of 108 items, with four items written for each of the 27 psychological components identified through the literature review. Items were authored to maximise content validity, avoid double-barrelling, and capture the breadth of each construct. Subject matter experts reviewed all items for clarity, redundancy, and conceptual alignment. Prior to analysis, inter-item correlations were inspected, and 24 items with correlations above 0.80 were removed, resulting in 84 items entering exploratory factor analysis (EFA). A parallel analysis supported an 8-factor solution, and an EFA using principal axis factoring with Oblimin rotation was conducted. Items with cross-loadings or loadings below 0.40 were removed, yielding 64 items. Although eight factors initially emerged, further theoretical inspection indicated conceptual overlap, and the structure was refined into six psychologically coherent dimensions. Specifically, Factor 1 and Factor 8 both focused on relationships and collaboration. Factor 6 and Factor 7 both reflected perceptions of effectiveness and ability to carry out the role. Although Factor 5 overlapped with relational themes, it was more closely aligned with psychological safety and trust. Based on factor loadings, predictive relationships with relevant outcomes (i.e., intention to quit, task performance, contextual performance, and counterproductive work behaviour), and coverage of the original conceptual domains, 54 items were retained for confirmatory factor analysis (CFA). Model fit indices, i.e., comparative fit indices (CFI > 0.95) and Tucker-Lewis indices (TLI > 0.95), and item performance guided further refinement, ensuring each factor retained strong theoretical coherence, internal consistency (ranging from α = 0.83 to 0.98), and predictive utility.
To strengthen underrepresented factors, 12 new items were added in the second study, resulting in 66 items tested in a subsequent CFA. Items were assigned to factors according to their highest loading and conceptual alignment, with no cross-loadings included. Skewness values ranged from −1.78 to 0.57. Most variables showed approximately normal distributions. As skewness values fell within the commonly accepted range of ±2 (e.g., Kline, 2016), no transformations were applied. All variables showed acceptable levels of kurtosis (−1.22 to 3.84), suggesting that the distributions slightly but did not significantly deviate from normality (Kline, 2016; Curran et al., 1996). The revisions made during the second study were therefore found to have strengthened the factor structure, improving both the theoretical and practical utility of the model. The first two studies also showed that the subdimensions exhibited differential predictive validity through differentiated patterns of relationships with predictive criteria: intention to quit, task performance, contextual performance, and counterproductive work behaviour, highlighting their unique contributions to workplace outcomes.
Through this iterative, evidence-based process, combining expert review, EFA/CFA outcomes, and predictive validity testing, the scale was reduced to a final, psychometrically robust set of 36 items representing six dimensions of organisational connection. Building on these prior studies, the present study sought to test the final version of the OCM, evaluating its robustness and predictive power in a more comprehensive manner.
2. Method
2.1. Participants
Given that the study focused on workplace perceptions, an occupational sample was utilised. All participants were working individuals who had voluntarily joined a research panel after completing a psychometric assessment with a large global psychometric publisher, undertaken for genuine occupational purposes such as recruitment or employee development. As such, all participants had experience in the working world.
Across the various studies undertaken, a total of 1,003 respondents participated. However, the present article focuses on the final sample of 307 respondents, of which 63% indicated being female, 34% male, and 3% other/no response. The majority were from the UK (58%), followed by South Africa (17%), USA (10%), Canada (7%), and Australia (5%), with 3% indicating other/no response. Their ages ranged from 19 to 74 years (M = 46.9, SD = 11.6).
2.2. Materials
Organisational Connection Measure (OCM)
For the current study, the final measure used was the 36-item OCM derived from prior research (Thomas, 2025). Respondents answered on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree) and were asked to think about their current job. Scores were calculated for each of the six factors (Appreciation, Cohesion, Belonging, Trust, Contribution, and Wellbeing), as well as an overall Connection score. Details of the factors, descriptions, and example items can be seen in Table 2.
Table 2. OCM factors, descriptions, and example items.
Factor |
Description |
Example Item |
Appreciation |
Fair reward for the work completed; feeling as though efforts are acknowledged and
appreciated and achievements celebrated. |
I feel valued for my work by my colleagues and peers. |
Cohesion |
Positive social relationships; individuals working together effectively towards the same goals; support for one another. |
I work in a friendly and
inclusive atmosphere. |
Contribution |
Access to resources required to be effective in role; clear expectations and autonomy to manage work and time. |
I have access to the information I need to
perform my job effectively. |
Belonging |
Satisfaction and fulfilment with job role and alignment to wider goals, values, and success. |
I am proud to be a member of my team. |
Trust |
Confidence in decision-making processes; feeling as though the work environment is a safe space for communication and that all voices matter. |
I can rely on the promises and commitments made by my superiors. |
Wellbeing |
Healthy work-life balance; ability to manage stress and pressure effectively without it
having a negative impact on health or
performance. |
My job promotes a sense of balance and wellbeing in my life. |
The Individual Work Performance Questionnaire (IWPQ; Koopmans et al., 2012)
The IWPQ is an 18-item measure of self-reported work performance, based on three scales: task performance (TP), contextual performance (CP), and counterproductive work behaviour (CWB). It has been found to be valid and reliable and correlates well with other measures of work performance (Ramos-Villagrasa et al., 2019). Participants responded to items on a 5-point Likert scale and were asked to consider the last three months of work.
Intention to Quit
Intention to quit (ITQ) was measured using a self-constructed single item scale asking: “I often think about leaving my job”, existing on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). Turnover intention is often used as a proxy for actual turnover, with evidence to suggest the two are related (Tett & Meyer, 1993).
2.3. Procedure
Participants were recruited via email and directed to an online survey platform. After providing informed consent, they completed the relevant measures for each study, followed by demographic questions. Participants completed the OCM, the ITQ question, and the IWPQ scale to enable the researchers to investigate the predictive validity of the items and scale. At the end of each study, participants received debrief information and were thanked for their participation.
2.4. Analysis
As exploratory factor analyses conducted in earlier studies helped establish an initial factor structure, the current study focused on confirmatory factor analysis to evaluate the finalised model, followed by correlations to examine relationships between study variables, and regression analyses to assess predictive validity. Prior to analysis, cases with missing data on key variables were removed. All analyses were performed using RStudio (Posit Team, 2025).
3. Results
3.1. Confirmatory Factor Analysis
To confirm the reduced 6-factor model and higher-order Connection model, a CFA was conducted using the diagonally weighted least squares (DWLS) estimator, which is recommended for ordinal data such as Likert-type responses due to its robustness and accuracy in estimating model fit (Flora & Curran, 2004). The 6-factor structure was specified based on the results of the preceding EFA, with items assigned to factors according to their highest loading and conceptual alignment. No cross-loadings were included. The aim of the CFA was to assess the model’s overall fit to the data and provide evidence for the factorial validity of the proposed structure.
Model fit was evaluated using a range of indices: the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardised root mean square residual (SRMR), in accordance with recommended thresholds (Hu & Bentler, 1999). Internal consistency reliability was assessed for each of the six factors using both Cronbach’s alpha and McDonald’s omega total. While Cronbach’s alpha remains a widely used measure of reliability, it assumes tau-equivalence and may underestimate reliability when this assumption is violated. Therefore, McDonald’s omega total was also reported, as it provides a more accurate estimate of reliability in the presence of congeneric items (Sijtsma, 2009). Values above 0.70 were considered acceptable, with higher values indicating stronger internal consistency. Results can be seen in Table 3.
Table 3. Confirmatory factor analysis and reliability statistics of the OCM.
Factor |
χ2/df |
p-value |
CFI |
TLI |
RMSEA |
Confidence Interval |
SRMR |
α |
ωt |
No. Items |
Connection |
1.078 |
>0.05 |
0.998 |
0.998 |
0.016 |
0.000 - 0.025 |
0.065 |
0.96 |
0.97 |
36 |
Appreciation |
0.856 |
>0.05 |
1.000 |
1.002 |
0.000 |
0.000 - 0.058 |
0.037 |
0.85 |
0.89 |
6 |
Continued
Cohesion |
0.379 |
>0.05 |
1.000 |
1.019 |
0.000 |
0.000 - 0.009 |
0.031 |
0.81 |
0.85 |
6 |
Contribution |
1.019 |
>0.05 |
1.000 |
0.999 |
0.008 |
0.000 - 0.065 |
0.046 |
0.80 |
0.85 |
6 |
Belonging |
0.254 |
>0.05 |
1.000 |
1.013 |
0.000 |
0.000 - 0.000 |
0.023 |
0.87 |
0.91 |
6 |
Trust |
0.441 |
>0.05 |
1.000 |
1.007 |
0.000 |
0.000 - 0.024 |
0.027 |
0.88 |
0.91 |
6 |
Wellbeing |
1.367 |
>0.05 |
0.997 |
0.995 |
0.035 |
0.000 - 0.078 |
0.048 |
0.86 |
0.90 |
6 |
|
Factor loadings |
|
Appreciation |
Collaboration |
Contribution |
Belonging |
Trust |
Wellbeing |
|
|
|
|
Item 1 |
0.677 |
0.472 |
0.696 |
0.863 |
0.782 |
0.797 |
|
|
|
|
Item 2 |
0.553 |
0.784 |
0.732 |
0.730 |
0.830 |
−0.5301 |
|
|
|
|
Item 3 |
0.883 |
0.780 |
0.668 |
0.750 |
0.810 |
0.716 |
|
|
|
|
Item 4 |
0.774 |
0.788 |
0.804 |
0.789 |
0.679 |
0.852 |
|
|
|
|
Item 5 |
0.795 |
0.650 |
0.644 |
0.731 |
0.794 |
0.930 |
|
|
|
|
Item 6 |
0.838 |
0.666 |
0.568 |
0.851 |
0.706 |
−0.7281 |
|
|
|
|
Connection |
0.916 |
0.949 |
0.996 |
0.937 |
0.954 |
0.925 |
|
|
|
|
R2 |
0.840 |
0.901 |
0.992 |
0.879 |
0.909 |
0.856 |
|
|
|
|
Note. 1Reverse-phrased statement. Item 1 - 6 refers to the items of the respective factors where Item 1 of Appreciation is not the same as Item 1 of Collaboration, etc.
All six factors as well as the higher-order Connection factor demonstrated excellent model fit (see Table 3), with all achieving very strong comparative fit indices (CFI ≥ 0.997) and Tucker-Lewis indices (TLI ≥ 0.995). The higher‑order Connection factor and all subdimensions (Appreciation, Cohesion, Belonging, Trust, Contribution, and Wellbeing) showed particularly strong fit, with CFI and TLI values at or above 0.997 and RMSEA values ranging from 0.000 to 0.035, all within the recommended thresholds for close fit. SRMR values were consistently low across models (0.023 - 0.065), further indicating excellent model-data correspondence. Internal consistency was high across all subdimensions, with Cronbach’s alpha values ranging from 0.80 (Contribution) to 0.88 (Trust), and McDonald’s omega total values from 0.85 to 0.91, indicating robust reliability. The overall Connection factor also demonstrated excellent internal consistency (α = 0.96, ω = 0.97), supporting the coherence and stability of the 36‑item measure.
3.2. Correlations
To explore the relationships between the latent constructs, inter-factor correlations were examined. These correlations provide insight into the extent to which the factors are distinct yet related, offering evidence for the conceptual structure of the model. Given the theoretical overlap between some constructs (e.g., collaboration and psychological safety), moderate to strong correlations were expected. The correlation matrix for the six factors and overall connection, as well as intention to quit and the IWPQ, is presented in Table 4.
Table 4. Correlation matrix for the connection factors, ITQ, and IWPQ (Consisting of TP, CP, and CWB).
|
M |
SD |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
1. Appreciation |
28.11 |
8.21 |
- |
|
|
|
|
|
|
|
|
|
2. Cohesion |
30.77 |
6.47 |
0.66*** |
- |
|
|
|
|
|
|
|
|
3. Contribution |
30.59 |
6.90 |
0.80*** |
0.69*** |
- |
|
|
|
|
|
|
|
4. Belonging |
31.30 |
7.80 |
0.76*** |
0.73*** |
0.80*** |
- |
|
|
|
|
|
|
5. Trust |
27.55 |
8.23 |
0.75*** |
0.79*** |
0.76*** |
0.72*** |
- |
|
|
|
|
|
6. Wellbeing |
28.94 |
8.03 |
0.71*** |
0.73*** |
0.73*** |
0.74*** |
0.75*** |
- |
|
|
|
|
7. Connection |
177.26 |
40.45 |
0.89*** |
0.86*** |
0.90*** |
0.89*** |
0.90*** |
0.88*** |
- |
|
|
|
8. ITQ |
2.97 |
1.57 |
−0.63*** |
−0.55*** |
−0.61*** |
−0.71*** |
−0.63*** |
−0.67*** |
−0.72*** |
- |
|
|
9. TP |
15.06 |
3.59 |
0.19** |
0.31*** |
0.28*** |
0.26*** |
0.21*** |
0.40*** |
0.31*** |
−0.13* |
- |
|
10. CP |
25.25 |
5.69 |
0.21*** |
0.18** |
0.31*** |
0.27*** |
0.14* |
0.20*** |
0.25*** |
−0.07 |
0.46*** |
- |
11. CWB |
6.17 |
4.11 |
−0.33*** |
−0.42*** |
−0.41*** |
−0.42*** |
−0.43*** |
−0.44*** |
−0.46*** |
0.47*** |
−0.36*** |
−0.16** |
Note. * = p < 0.05, ** = p < 0.01, *** = p < 0.001; ITQ = Intention to Quit, TP = Task Performance, CP = Contextual Performance, CWB = Counterproductive Work Behaviour.
The bivariate correlations in Table 4 show that the subdimensions were strongly and significantly intercorrelated (r = 0.66 - 0.80), suggesting that they share common variance, which is consistent with the presence of an overarching Connection factor. This is further supported by the very high correlations between each subdimension and the total Connection score (r = 0.86 - 0.90), demonstrating that each factor meaningfully contributes to the broader construct.
However, despite their shared variance, the subdimensions showed differentiated patterns of relationships with the predictive criteria, namely intention to quit, and IWPQ constructs Task Performance, Contextual Performance, and Counterproductive Work Behaviour. For example, Belonging demonstrated the strongest negative correlation with ITQ (r = −0.71), while Wellbeing had the strongest positive correlation with TP (r = 0.40) and Contribution with CP (r = 0.31). Wellbeing was most strongly negatively associated with CWB (r = −0.44); however, the relationships between all factors and CWB were of similar levels (r = −0.41 to −0.44).
3.3. Linear Regressions
Linear regression analyses were conducted to examine the relationship between Connection and the four outcome variables: ITQ, TP, CP, and CWB. Connection significantly predicted all four outcomes, as shown in Table 5.
Higher Connection was associated with lower intention to quit (B = −0.03, SE = 0.002, t(296) = −17.70, p < 0.001, 95% CI [−0.03, −0.02]), accounting for 51% of the variance in ITQ scores (R2 = 0.51). Connection positively predicted Task Performance (B = 0.03, SE = 0.00, t(296) = 5.61, p < 0.001, 95% CI [0.02, 0.04]), accounting for 10% of the variance in TP scores (R2 = 0.10), as well as Contextual Performance (B = 0.03, SE = 0.01, t(296) = 4.36, p < 0.001, 95% CI [0.02, 0.05]), accounting for 6% of the variance in CP scores (R2 = 0.06); and negatively
Table 5. Results from linear regressions showing connection predicting ITQ and IWPQ.
predicted Counterproductive Work Behaviour (B = −0.05, SE = 0.005, t(296) = −8.94, p < 0.001, 95% CI [−0.06, −0.03]), accounting for 21% of the variance in CWB scores (R2 = 0.21). These results suggest that stronger feelings of connection at work are associated with better performance and lower intentions to quit or performing of counterproductive work behaviours, with effect sizes ranging from small to large.
4. Discussion
The present study was the final study among a set of three to develop and validate a measure of workplace connection and examine its relationship with key organisational outcomes, including intention to quit, task performance, contextual performance, and counterproductive work behaviour. The findings provide strong support for the reliability and validity of the Organisational Connection Measure and its six subdimensions: Appreciation, Cohesion, Contribution, Belonging, Trust, and Wellbeing.
The CFA demonstrated excellent model fit across all subdimensions, with fit indices (CFI, TLI) at or above 0.995, RMSEA values at or near zero, and SRMR values all below 0.070. These results indicate that the Connection factors are psychometrically sound and capture distinct yet highly interrelated aspects of workplace connection. Reliability estimates were consistently strong, with Cronbach’s alpha and McDonald’s omega total values ranging from 0.80 to 0.97 across the dimensions, supporting the internal consistency of the measure.
Bivariate correlations revealed that while the subdimensions shared considerable variance, consistent with an overarching Connection construct, they also showed unique patterns of association with external criteria. In particular, Belonging demonstrated the strongest negative association with intention to quit, suggesting that employees who feel aligned to the organisation and its goals are less likely to consider leaving the organisation. Wellbeing, in turn, showed the strongest positive relationship with task performance, suggesting that employees who experience a healthy work-life balance have more energy and focus to complete their tasks well. Lastly, the finding that Contribution was most strongly related to contextual performance suggests that employees who are granted the autonomy and resources they need to take initiative and be effective in their work are willing to go above and beyond in their roles. These differentiated patterns suggest that while workplace connection operates as a coherent whole, its facets contribute uniquely to important organisational outcomes.
Regression analyses reinforced the utility of the measure in predicting outcomes, demonstrating that higher overall Connection significantly predicted lower turnover intentions and counterproductive work behaviours, and higher task and contextual performance. Notably, Connection accounted for a substantial 51% of the variance in intention to quit, highlighting its potential as a critical target for interventions aimed at improving employee retention. Although effect sizes for task and contextual performance were smaller, they were still statistically significant and meaningful, suggesting that fostering workplace connection can contribute to improved individual and organisational performance.
In a separate study of 178 respondents (Thomas, 2025), incremental validity was further supported in regression analyses: When turnover intention was first regressed on burnout alone, burnout strongly predicted intention to quit (β = 0.671, p < 0.001, R2 = 0.451). Adding Connection to the model substantially improved prediction (β = −0.538, p < 0.001, R2 = 0.611), with the part correlation of burnout dropping from 0.671 to 0.232 and Connection showing a unique part correlation of −0.400. These findings suggest that Connection captures variance in turnover intention not explained by burnout, supporting its discriminant and incremental validity relative to adjacent constructs.
We found strong support for a general factor of organisational connection, suggesting that a unifying sense of connection at work can be meaningfully captured. However, our findings also demonstrate that there is practical value in distinguishing between the six proposed sub-factors with a higher-order factor of Connection. Despite high intercorrelations among these sub-dimensions, each exhibited unique patterns of association with key work-related outcomes. This differentiation highlights the potential utility of the 6-factor model for organisational practice, particularly when seeking to diagnose and intervene in specific areas where connection may be lacking within teams or departments.
Overall, the results highlight the central role of workplace connection in shaping employee attitudes and behaviours. Employees who feel a greater sense of connection are more likely to remain with their organisation, perform better, and engage less in counterproductive behaviour, further underscoring the importance of organisational practices and leadership behaviours that cultivate a strong sense of connection among employees.
5. Limitations
This study relied on self-report measures rather than objective indicators, such as actual turnover, performance metrics, or multi-source assessments (e.g., 360-degree feedback). This may introduce common method bias and restrict the generalisability of findings to actual performance outcomes (Podsakoff et al., 2003). Future research could implement a latent common-method factor to assess its potential influence, as failing to account for common method bias may inflate the observed relationships among constructs. Nevertheless, previous studies have supported that both the self-report IWPQ and intention to quit have been related to work performance and turnover, respectively (Ramos-Villagrasa et al., 2019; Tett & Meyer, 1993), which supports their use in the present study.
Additionally, further validation of the Connection construct is needed, particularly its discriminant validity relative to related constructs. Although the present study focused primarily on internal structure and criterion relations, establishing that Connection is empirically distinct from related constructs such engagement, job satisfaction, person-job fit, and thriving, as well as psychosocial variables, such as facets of emotional intelligence, vocational interests, and preference for remote or hybrid work, remains an important next step. Discriminant validity testing using multitrait-multimethod approaches or structural equation modelling can help clarify the extent to which Connection captures a unique relational-motivational construct above and beyond related constructs. Incorporating more incremental validity analyses would further demonstrate whether Connection predicts meaningful outcomes (e.g., well-being, performance, retention intentions) above and beyond these established constructs. Including variables such as trait emotional intelligence (particularly facets like sociability, empathy, and well-being) and contextual preferences for remote work may be especially informative, as these represent individual differences that are likely to correlate with Connection but remain conceptually distinct from it. Together, these investigations will provide a more comprehensive validity framework and support the scale’s utility in both applied and research settings.
The sample was drawn primarily from Western populations, limiting confidence in generalisability to other cultural or occupational contexts. Although multiple English-speaking regions were included, replication with non-Western and more representative samples will be important. Future research should also examine the relationship between Connection and objective performance indicators (e.g., supervisor ratings, team metrics, absenteeism) and engagement, as well as potential demographic moderators such as age, gender, role seniority, and cultural background.
6. Conclusion
This research provides robust support for a comprehensive model of workplace connection, encompassing employees’ perceptions of their work, their organisation, and their colleagues. The findings also demonstrate the reliability, construct validity, and criterion validity of the connection measure, highlighting its relevance in predicting key work outcomes and its potential utility in both research and applied organisational contexts.