Organizational belonging – proposing a new scale and its relationship to demographic, organization, and outcome variables

Abstract

In view of its absence, a literature review covering Organizational Belonging (OB) and related constructs was done, and indicated the lack of a perceived OB scale, codified under one measure. OB was defined as “experiencing an acknowledgment of one’s talents, interests and experiences, and finding whole acceptance of one’s self expression of these.” Using a sample of 141 employees, twenty-seven OB items were adapted from three sources: prior quantitative work, prior qualitative work, and a survey on Diversity and Inclusion (American Institute for Certified Professional Accountants). Factor analytic support was found for a four-scale, 16-item OB measure. The four scales were labeled Be Myself (seven items), Acceptance (four items), Value Diversity (three items) and Connection (two items). Additional open-item analyses corroborated the validity of these scales. All four OB scales demonstrated sufficient reliability and distinctiveness from each other. These results validated the OB definition. After controlling for demographic and organizational variables, Be Myself had a significant impact on productivity and Acceptance had a significant impact on intent to stay. Future research testing this OB scale is needed, given the recent changes in the workplace, including remote work, retaining current staff, and hiring for unfilled positions.

Keywords:

• Organizational belonging
• authenticity
• acceptance
• value diversity
• connection

Introduction

When COVID first came in March 2020, over 47 million jobs were left voluntarily in 2021, according to study by CNN, and many have still not returned to the workforce (Romans, 2022). The impact of COVID on the United States (U.S.) workforce has been two-fold: retaining workers by giving them incentives to stay, including remote work wherever possible; as well as hiring new talent for unfilled job positions. One remote work downside has been many employees feeling less connected to their employer. Especially with some type of hybrid arrangement continuing for many, to help with retention, companies must find a way to increase their employees feeling “connected to the company’s mission, the company’s leaders and to their team” (Laker, 2022).

The search for new talent has been dubbed as a game of “musical chairs” (Laker, 2022) where many “employees swap jobs.” It isn’t just a question of hiring qualified new talent but also keeping them. Beyond expanding the talent pool (e.g., to nontraditional workers), and revising compensation and benefits (including work-life balance), De Smet, Dowling, Mugayar-Baldocci and Schaninger (2022) have argued that companies need to create “sticky” workplaces, which includes “listening to employees, anticipating and addressing their concerns, and a fostering psychological safety and a sense of community.”

Several recent practitioner articles have acknowledged the importance of creating employee feelings of organizational belonging. In a study of over 11,800 participants, Herbert (2020) found that fostering a sense of organizational belonging included: “having a connection, feeling like a valued team member; a perception they can be themselves at work; and a belief that their company is one where everyone can succeed to their full potential, no matter who they are.” In a survey of over 5,000 employees in Brazil, China, Germany, the UK, and the US, 82% of the respondents indicated they felt lonely at work, with 46% indicating they were likely to leave a job because of loneliness (Gurchiek, 2022). Keys to feeling a sense of organizational belonging included being able to freely voice one’s opinion, even if it differed from others, as well as feeling connected to others at work (Gurchiek, 2022).

Given this revitalized importance for studying organizational belonging (OB), the study’s purpose was to develop a valid and reliable scale for measuring OB and to also investigate its antecedents and outcomes. However, a literature review of existing OB-related research was first needed to document progress to date. Based on this literature review, we used the definition of OB offered by Belle, Burley, and Long (2015, p. 90), “experiencing an acknowledgment of one’s talents, interests and experiences, and finding whole acceptance of one’s self expression of these.”

The literature review below first briefly contrasts OB against related organizational constructs. It then systematically reviews prior empirical research on OB and related constructs, such as Workplace Belongingness, Employee Belongingness, Organizational Connectedness, and General Belonging. Individual study critiques of prior research are also included, as well as short section summaries, before looking at qualitative research to enlarge the item domain of OB to better measure its conceptual definition. Collectively, this builds the case for testing a new OB measure. Finally, prior research on antecedents and consequences of OB are reviewed, as a prelude to their being further studied.

Literature review

Prior conceptual and empirical work related to OB

OB is not organizational support, organizational culture, or organizational Commitment

OB is related to but distinct from Perceived Organizational Support (POS), Organizational Culture (OC), and Organizational Commitment (OrgC). POS is the employee’s perception that the organization values their contribution and cares about their well-being (Kurtessis et al., 2017), as demonstrated by fair employee treatment, supervisor support, organizational rewards, and favorable job conditions. OC has been defined as “those things that are valued and rewarded within your company—that is, the pattern of beliefs and expectations shared by members, and their resulting behaviors” (Chatman, Caldwell, O’Reilly, & Doerr, 2014). OrgC has been most often conceptualized/measured as having three attachment dimensions (Cohen, 2007): affective (attachment due to identification), normative (attachment due to obligation), and continuance (attachment due to the perceived costs of leaving). OB has a more innate foundation in the essential human need for belonging (Baumeister & Leary, 1995) than POS, OC, or OrgC.

OB rooted in the more general need for belonging

The human need for belonging is a fundamental human motivation (Maslow, 1968). Belonging refers to a relation, where people feel valued by another person, group, and/or organization, which involves creating a sense of identification with that person and/or entity, through forming social bonds (Baumeister & Leary, 1995). Belonging is a multidimensional construct because it involves our identity (e.g., age, gender, ethnicity, religion, sexuality, occupation) interacting with different people (and their identities), social contexts and places, including organizations (May, 2011).

Critiquing prior empirical work on OB

Knapp, Smith, and Sprinkle (2014, p. 274) defined OB as the “perception of an intimate association with the organization as demonstrated by a sense of: perceived insider status, psychological ownership, and organizational identification.” Perceived insider status (Stamper & Masterson, 2002) is the employee’s sense of inclusion within their organization. Psychological ownership is a feeling of possessiveness over some object, including an organization (Vandewalle, Van Dyne, & Kostova, 1995). Organizational identification is an employee defining him/herself in terms of their organization (Ashforth & Mael, 1989). All three of these components together comprise OB (Knapp et al., 2014). Prior empirical research on OB has worked with three multi-item scales.

For example, perceived Insider Status has been measured using a six-item scale (Stamper & Masterson, 2002, p. 883). However, closer inspection shows that the strong scale reliability is partially due to reverse-scored (R) item wording redundancies, e.g., “my work organization makes me believe I am included in it,” versus “I don’t feel included in this organization” (R). Van Dyne and Pierce (2004, p. 449) developed a seven-item Psychological Ownership scale, also with a high internal consistency reliability estimate partially due to item wording redundancies, e.g., “this is MY organization,” versus “I sense that this is MY company.”

Organizational Identification has been measured (Mael & Tetrick, 1992, p. 817) with a six-item scale, combining three items focusing on shared experiences, e.g., “when someone criticizes this organization, it feels like a personal insult,” and three-items on shared characteristics, e.g., “if a story in the media criticized the organization, I would feel embarrassed.” Edwards and Peccei (2007) developed a reliable six-item Organizational Identification scale, with three two-item components: self-categorization and labeling, sharing of organizational goals and values, and sense of organizational membership. For example, items comprising self-categorization are (Edwards & Peccei, 2007, p. 34): “my employment in (company name) is a big part of who I am” and “I consider myself to be a (company name) person.” It can be argued that these scales, via items that either over-paraphrase each other or by being antonyms, maximize internal consistency, but at the cost of narrower content/domain coverage, with potential lower construct validity (Stanton, Sinar, Balzer, & Smith, 2002).

Summary critique of prior OB research

Prior empirical research has been done using a variety of U.S. samples, including: nonprofit and manufacturing employees (Knapp et al., 2014); restaurant employees (Stamper & Masterson, 2002); clerical, office, supervisory, professional, and managerial employees (Van Dyne & Pierce, 2004); working part-time students (Mael & Tetrick, 1992); and health care employees (Edwards & Peccei, 2007). Aside from individual study critiques, why is it necessary to “cobble together” or combine three different constructs/scales, i.e., perceived insider status, psychological ownership, and organizational identification into an overall OB measure? When combining these three scales, it seems assumed that all are equally weighted (Knapp et al., 2014), but how do we know this is accurate? For example, in Knapp et al. (2014, p. 278), different means were reported, i.e., perceived insider status, M = 5.27, psychological ownership, M = 3.99, organizational identification, M = 4.88, which should have been tested for mean scale differences. Further skepticism about equal weighting occurs when Knapp et al. (2014) also found that when all three scales were entered simultaneously for explaining turnover intentions, organizational identification was not significant, unlike both perceived insider status and psychological ownership. Knapp et al. (2014, p. 281) speculated that “the nature of the relationship between organizational identification and outcomes operate through a different mechanism than the other two aspects of belonging.”

If this OB definition is still valid, collectively these empirical concerns necessitate the ongoing validation of a stronger composite measure of OB combining insider status, psychological identification, and organizational identification (Booysen, 2002). It is important to emphasize again that prior empirical work to date on OB has generally worked with these individual components, not an overall measure of OB. There has also been related scale research on the concepts of workplace belongingness, employee belongingness and organizational connectedness, as well as general belonging, which needs to be briefly reviewed.

Prior work on workplace belongingness, employee belongingness, and organizational connectedness scales

Jena and Pradham (2018) published a 12-item Workplace Belongingness scale, which they defined as (p. 454) “intending to capture the professionals’ belongingness towards their job and organization in both letter and spirit.” Their sample of 824 Indian respondents consisted of 69% male executives, and 72% professional graduate degree holders. The item content shows that it is combining heterogeneous statements together into one scale (p. 456). There are items reflecting organizational identification, e.g., “I refer as ‘we/us’ rather than ‘they/them’ when I refer my organization to outsiders,”; values, e.g., “I feel that there is a semblance between my organization and my own values and beliefs”; needs, e.g., “my personal needs are well met by my organization,” rewards, e.g., “accomplishments at work are adequately rewarded in my organization,” and a career perspective, e.g., “my career goals are well considered by my organization.” This item heterogeneity, combined with the sample used, would seem to limit their scale generalizability across other sample settings.

Studies by Randall, Lartey, and Tate (2020) and Gao, Low, and Gong (2022) used employee belongingness measures. Randall et al. (2020) investigated the relationship between enterprise social media (ESM) use and employee belongingness. They factor analyzed twenty-eight items to develop their scales for measuring both ESM use and belongingness. They found four factors labeled (Randall et al., 2020, p. 118) “Social ESM”; “Work ESM”; “Belongingness” and “Disengagement.” The four-item employee belongingness factor focused on feeling connected, however, no reliability estimate for this factor (scale) was reported. Gao et al. (2022) created a 12-item measure of employee belongingness based on Place Attachment Theory (PAT), which refers to the affective bond between people and a physical place, e.g., home, office. Gao et al. (2022) scale consisted of three components: place dependence (preference over alternatives); place identity (comfort); and social bonding (familiarity with people). This scale seems best suited for measuring how employees feel about their physical workspaces.

Organizational connectedness (Huynh, Metzer, & Winefield, 2012, p. 1058) is defined as “a positive state of well-being that results from an individual’s strong sense of belonging with other workers and the recipients of one’s service.” It has been measured using 15-items (Huynh et al., 2012, p. 1064) broken into four components (number of items): other workers (4) – sense of belonging with one’s work group; recipients (4)—being appreciated by one’s client; task requirement (3)—being really good at one’s work; and values (4)—feeling valued by one’s organization. The belonging and values components are central to OB (Herbert, 2020) but the other two components are less so. One common feeling across all workplace belongingness, employee belongingness, and organizational connectedness scales cited was a general belonging.

Prior empirical work on general belonging

Hagerty, Lynch-Sauer, Patusky, Bouwsema, and Collier (1992, p. 173) defined sense of belonging as “the experience of personal involvement in a system or environment so that persons feel themselves to be an integral part of that system or environment.” Hagerty and Patusky (1995) developed the 27-item Sense of Belonging Instrument (SOBI) consisting of two separately scored scales, SOBI-P (18 items, measuring psychological state) and SOBI-A (9 items, measuring antecedents), both with high internal consistency reliabilities. Abbreviated sample items presented for SOBI-P include (Hagerty & Patusky, 1995, p. 11): “wonder if I really fit,” “I don’t really fit,” and “describe myself as a misfit”; while abbreviated sample items for SOBI-A include: “important that I fit,” “working on fitting in,” and “make myself fit.” Davila and Garcia (2012) created a Sense of Belonging scale by adapting 7-items from the general belonging scale by Hagerty and Patusky (1995) to an organization. Sample items were (Davila & Garcia, 2012, p. 248) “I really don’t fit in this company,” “I could disappear for days and nobody at the company would miss me.”

More recently, Malone, Pillow, and Osman (2012) developed a 12-item, two factor General Belongingness Scale (GPS). The two factors, six-items each, are acceptance/inclusion and lack of rejection/exclusion. Sample items (Malone et al., 2012, p. 31) include: “when I am with other people, I feel included” and “I feel connected with others” (acceptance/inclusion) versus “with other people I feel like a stranger” and “I feel isolated from the rest of the world” (rejection/exclusion). The use of similar wording and antonym items would seem to play a big role in the strong psychometric data presented for both the SOBI and GPS scales (Stanton et al., 2002). However, a more recent 10-item Need to Belong scale does not seem to suffer these drawbacks (Leary, Kelly, Cottrell, & Schreindorfer, 2013). Yet, this is still a general belonging scale, not focusing on Organizational Belonging. Expanding the literature search on OB, several qualitative studies were found offering the potential for enlarging its item domain coverage.

Summary critique of prior related OB research

Again of a variety of samples have been utilized, including: Indian managerial executives (Jena & Pradham, 2018); U.S. corporate employees (Randall et al., 2020); employees of a Singapore bank (Gao et al., 2022); Australian volunteer ambulance workers and volunteer emergency service workers (Huynh et al., 2012); U.S. community college students, patients in mental health treatment, and Roman Catholic nuns (Hagerty & Patusky, 1995); Spanish contract service employees (Davila & Garcia, 2012); and U.S. college students (Malone et al., 2012; (Leary et al., 2013). There are domestic and non-U.S. samples across these studies. However, the measures used across these studies have narrow domain coverage, and do not adequately address the breadth of the OB construct (Belle et al., 2015).

Qualitative research enlarging the Item domain for measuring OB

Three qualitative studies on OB or general belonging were helpful (Belle et al., 2015; Filstad, Traavik, & Gorli, 2019; McClure & Brown, 2008). Belle et al. (2015) interviewed 10 teleworkers asking each the question “how do teleworkers experience organizational belonging?” and labeling each depending on their answers. Examples (Belle et al., 2015, p. 85–86), of the label given and their abbreviated quotes were: first responder—“having space for one’s expression of full empowerment and liberty to engage fully”; anonymous professional—“non-discriminatory, fact-based acknowledgement by others of one’s abilities,” and teamster—“it is akin to encountering mindful, persistent cohesion with others.” Filstad et al. (2019) had 51 students take one to four snapshot photographs with accompanying text explaining what “belonging at work” meant to them. Collective interpretation by the authors included the following themes: being part of something (on equal terms); building connection; and tension and negotiation as part of becoming. McClure and Brown (2008) interviewed a heterogeneous sample of twelve adults asking about their feeling of belonging as experienced at work. Six belonging themes were identified: (1) discovery of self within a job; (2) being invited and learning to be part of a group; (3) wanting to be included; (4) learning how to connect; (5) doing meaningful work; and (6) not-belonging, e.g., fear, rejection, and competition.

Through these three qualitative articles and the authors’ interpretation, new item generation ideas including freedom of expression, and appreciating and honoring differences surfaced, reinforcing newer, practitioner-oriented studies, i.e., Herbert (2020), Gurchiek (2022). These new item-testing ideas were combined with modifying existing items based on the earlier mentioned themes around connection, inclusion, and acceptance. This will be further discussed in the Measures section. Overall, by combining these different sources, the authors generated a new domain of items covering the OB construct for subsequent testing and scale development. Given the lack of prior research, this suggested the following general research question (RQ1) to be tested:

RQ1 – reliable and valid new OB scales can be measured.

Relationships of antecedents and Outcomes to new OB Scales

Having found new reliable and valid OB scales in RQ1, what are anticipated relationships to antecedent and outcome variables? Prior empirical OB research has only focused on using the components of OB, i.e., perceived insider status, psychological ownership, and organizational identification, separately, not in an overall OB measure. Working with each OB component, Knapp et al. (2014, p. 278) did not find any significant relationships of OB components to antecedent demographic (i.e., age, gender, part-time/full-time employment) or organizational-level variables (i.e., organizational tenure, salary, industry). Correlations showed that each OB component was negatively related to turnover intention.

Stamper and Masterson (2002) did not find significant relationships between organizational tenure, race, gender, or age, to perceived insider status. Race was coded as either white or other. Van Dyne and Pierce (2004) did not find age, education, gender, or organization tenure to be related to psychological ownership in their third study sample of professional employees (not reported for other samples). For this third sample, they did find a significant positive relationship of psychological ownership to supervisor-rated employee performance. Jena and Pradham (2018) also found a significant positive relationship between workplace belongingness and employee self-rated performance. For both samples used, Edwards and Peccei (2007) found their six-item organizational identification scale to be significantly negatively related to turnover intention. Both Davila and Garcia (2012) and Huynh et al. (2012) reported positive relationships between belonging/connectedness and one-item self-reports of intent to stay in the organization.

Based on this prior work, we would not expect significant relationships between demographic or organizational-level variables. However, it is important to note that independent sample t-tests on OB-related scales for groups of employees, e.g., by gender or race, were not reported. Prior research supports expecting positive relationships between OB to both productivity and intent to stay. This leads to RQ2:

RQ2 – positive relationships are expected between OB scales to productivity and intent to stay.

Related to RQ2 would be to test if the new OB scales will explain productivity and intent to stay beyond controlled for demographic and organization variables. This leads to RQ3:

RQ3 – OB scales will contribute to explaining productivity and intent to stay beyond controlled for demographic and organization variables.

Method

Subjects and procedure

Respondents were employees from two firms in the United States. One organization is a business technology company (Company A) with approximately 150 employees, and the second (Company B) an accounting firm with 45 employees. Both companies are members of the BDO Alliance Diversity Roundtable (BDO). The BDO is a nationwide association of approximately 20 independently owned local and regional accounting, consulting, and technology service-related firms with similar client service goals, i.e., expand services to their clients (https://alliance.bdo.com/). The BDO has affiliative ties to the American Institute for Certified Professional Accountants (AICPA), a nonprofit organization that created and distributes an annual AICPA Diversity and Inclusion toolkit for its members’ use through the BDO. The third coauthor’s company is a BDO member and secured permission to solicit companies to partake in this study. Of the ten or so firms participating in the Roundtable at the time, Companies A and B were the only respondents that agreed to have their employees surveyed.

An anonymous Qualtrics survey was used for data collection. There were two survey sections: a section measuring OB (27 items) and outcomes (two items), followed by a section measuring demographic and organizational variables (seven items). The demographic and organizational variable items were developed based on the third author’s knowledge of employee and organizational characteristics for Company A and in consultation with a professional contact at Company B. Temple University’s IRB approved this research under Exempt Review (Protocol #28794, 10/25/21). The data was collected from late Fall 2021 (Company A) through Winter 2022 (Company B), with email reminders sent to respondents in each company approximately every three weeks in that time period. In all, 130 out of 150 (87%) Company A employees and 30 out of 45 (67%) Company B employees filled out the survey for a total of 160 respondents. Respondents were told that the purpose of the survey was to measure their perceptions about organizational belonging and factors that could be associated with it. During this data collection period, approximately 77% of employees worked remotely at Company A. It is not known what percentage of employees worked remotely at Company B. The email reminders and longer data collection period likely increased the response rate.

In early July 2022, a follow-up open item survey was sent to Company A employees using an anonymous Qualtrics link. A small subsample of Company A employees, n = 42 responded. The purpose of this open item was to validate the initial OB survey results found. Although this validation sample came from the same population of the initial survey respondents, there was a six-month time gap (January to June) between the initial survey and this smaller sample open item follow-up. Thus, there is a very low probability that an employee would accurately remember their initial survey responses made earlier (Johns, 1994).

Initial sample survey measures

Demographic variables

Five demographic variables were measured on the survey: gender, race, religion, political scale, and age. Gender was measured by asking “what is your gender identity,” where 1 = woman, 2 = man, 3 = gender queer/non-binary, 4 = Agender, 5 = Other (fill in) 6 = prefer not to say. Race was measured by asking “what is your racial or ethnic identity,” where 1 = White, 2 = Black or African American, 3 = American Indian or Alaska Native, 4 = Asian, 5 = Native Hawaiian or Pacific Islander, 6 = Hispanic, 7 = Bi-racial, 8 = Other (fill in), 9 = Prefer not to say. Religion was measured by asking “what is your religious background,” where 1 = Christian, 2 = Muslim, 3 = Jewish, 4 = None, 5 = Other (fill in), 6 = prefer not to say. Political scale was measured using one item, “please indicate where you identify on the political scale,” where 1 = extremely conservative, 2 = conservative, 3 = moderate, 4 = liberal, 5 = very liberal, 6 = other (fill in). Age was measured by asking, “how old are you,” where 1 = 20s, 2 = 30s, 3 = 40s, 4 = 50s, 5 = 60s, 6 = 70s or older.

Organization variables

Two variables were measured: organization tenure and position level/role. Organization tenure was measured by asking “how long have you worked at your organization?” where 1 = less than one year, 2 = 1–4 years, 3 = 5–9 years, 4 = 10–20 years, 5 = more than 20 years. Position level/role was measured by asking “choose the position level that most aligns with your role in the organization,” where 1 = owner/partner, 2 = C-Suite, 3 = Vice President, 4 = Director, 5 = Manager/supervisor, 6 = Technical staff, 7 = Consultant, 8 = Sales, 9 = Administrative support, 10 = Other (fill in).

OB items

Twenty-seven items were used. Items were taken from three different sources. The first source was prior empirical work noted in the literature review: (items adapted for use), e.g., (feeling accepted; opinion valued), Hagerty and Patusky (1995) and Malone et al. (2012); (sense of belonging), Edwards and Peccei (2007), Huynh et al. (2012), and Van Dyne and Pierce (2004); (feeling included), Stamper and Masterson (2002); and (being an outsider), Malone et al. (2012). This represented ten items. The second source was based on the “themes” generated in qualitative studies, i.e., Belle et al. (2015), Filstad et al. (2019), McClure and Brown (2008), and Herbert (2020) which were translated into more specific items, e.g., being authentic; to engage fully; acceptance of oneself; cohesion with others; differences being appreciated and honored; and connecting with colleagues. This represented eight items. The third source for items was a survey by the American Institute for Certified Professional Accountants (AICPA, 2022). The AICPA has a toolkit for diversity and inclusion in firms, which includes items measuring attitudes toward diversity and inclusion, belonging, and discrimination. As a member of AICPA, the third coauthor had access to the toolkit and was allowed to adapt relevant nine sample diversity and inclusion survey items.

To the authors’ knowledge, these sampled AICPA survey items have never been utilized in published research. A six-point response scale was used for all items, where 1 = strongly disagree to 6 = strongly agree, with no neutral point. These items are presented in Table 2.

Outcomes

Two one-item variables were measured, productivity and intent to stay. Productivity was measured by asking “I feel productive in my position at work.” Intent to stay was measured by the item “I plan to stay at this organization for the foreseeable future.” Both variables used the same six-point scale, 1 = strongly disagree to 6 = strongly agree. Prior work has used similar items, e.g., Huynh et al. (2012).

Validation sample survey measures

Position level/role was measured by repeating the initial survey categories, where 1 = owner/partner, 2 = C-Suite, 3 = Vice President, 4 = Director, 5 = Manager/supervisor, 6 = Technical staff, 7 = Consultant, 8 = Sales, 9 = Administrative support, 10 = Other (fill in).

Open item

Respondents were asked “In a few sentences, please describe what it is that makes you feel that you are a part of, or belong to, this company (what factors contribute to your sense of belonging here)."

Data analyses

First, demographic and organization variables on each company sample were checked for differences before combining into an overall sample. The combined data was then used. Frequency analyses for nominal demographic and organization variables are reported first (Table 1). The nature of the research question asked dictated the statistical analysis to be used (Stevens, 1996). Exploratory factor analysis was used (Table 2) to reduce the 27 OB items into more usable scales to test RQ1. Means, standard deviations, and correlations were performed on continuous demographic, organization, OB scales, and outcome variables (Table 3). Correlation analysis was used to test RQ2. In addition, independent samples t-tests were also used to test RQ2. Hierarchical regression analyses were used to test RQ3. Any required pre-recoding of variables to run the above analyses will be reported. SPSS-PC (SPSS, 2021) was used for all data analyses. Following Hopkins, Marshall, Batterham, and Hanin (2008), findings at a p<.05 or p<.01value (two-tailed) will be reported as statistically significant, while findings at a p<.10 value (two-tailed) will be labeled as marginally significant. For the open item, the authors collectively analyzed each response and agreed on coding each respondent’s answers, ideally into at least one of the created OB scales.

Table 1. Frequencies and percentages – nominal demographic and organization variables.

Variable	n = 160
Gender
Woman	n = 63 (38%)
Man	n = 79 (49%)
Agender	n = 1 (1%)
Missing/Prefer not to say	n = 17 (12%)
Race
White	n = 118 (75%)
Black or African American	n = 5 (3%)
American Indian/Alaska Native	n = 0
Asian	n = 2 (1%)
Native Hawaiian/Pacific Islander	n = 0
Hispanic	n = 13 (8%)
Bi-racial	n = 2(1%)
Other (e.g., mix)	n = 2 (1%)
Missing/Prefer not to say	n = 18 (11%)
Religion
Christian	n = 91 (57%)
Muslim	n = 2 (1%)
Jewish	n = 8 (5%)
None	n = 35 (22%)
Other (e.g., agnostic, spiritual)	n = 6 (4%)
Missing/Prefer no to say	n = 18 (11%)
Position Level/Role
Owner/partner	n = 4 (2%)
C-Suite	n = 8 (5%)
Vice President	n = 6 (4%)
Director	n = 9 (6%)
Manager/supervisor	n = 28 (17%)
Technical staff	n = 32 (20%)
Consultant	n = 19 (12%)
Sales	n = 13 (8%)
Administrative support	n = 15 (10%)
Other (e.g., marketing, coordinator)	n = 6 (4%)
Missing	n = 20 (12%)

Table 2. Exploratory factor analysis for 27 organizational belonging items using a seven-factor extraction and Varimax rotation.

Thinking about your organization^a:	1^c	2^c	3^c	4^c	5^c	6	7
1. Important to me to feel sense of belonging at work^b	0.09	0.03	−0.09	0.07	−0.06	0.85	0.10
2. Employees treated fairly, regardless of differences (e.g., race, gender)	0.30	0.58	0.42	0.05	0.22	0.41	−0.19
3. People of all cultures and backgrounds are respected and valued	0.04	0.63	0.44	−0.04	0.09	0.42	−0.18
4. When I am with people from my organization, I feel included	0.19	0.75	0.03	0.22	0.19	0.05	0.23
5. I feel accepted by others in the organization	0.30	0.73	0.10	0.36	0.05	−0.12	0.20
6. I am comfortable talking about my background/cultural experiences^b	0.30	0.45	0.20	−0.04	0.44	0.22	0.26
7. Employees of different backgrounds interact well in our company	0.29	0.26	0.57	0.16	0.37	−0.02	0.06
8. Management shows a commitment to meeting employee disabilities’ needs	0.01	0.17	0.18	0.30	0.62	0.46	−0.16
9. Racial, ethnic, religious jokes are not tolerated	0.08	−0.04	0.82	0.05	0.01	0.08	0.11
10. The environment here allows for free and open expression of ideas, beliefs^b	0.50	0.22	0.50	0.28	0.01	0.09	0.18
11. I can be successful as my authentic self (who I am, regardless of position)	0.63	0.34	0.19	0.20	0.02	0.21	0.25
12. I feel a sense of belonging in this organization	0.35	0.32	0.11	0.78	0.01	0.13	0.05
13. I am recognized for my accomplishments in my organization^b	0.46	0.55	0.05	0.45	−0.05	−0.09	−0.07
14. My contributions in team meetings are valued	0.70	0.37	−0.01	0.19	−0.06	−0.11	0.22
15. I feel comfortable with being myself at work	0.66	0.18	0.33	0.20	0.01	0.08	0.36
16. There is transparent communication about important company matters^b	0.59	−0.10	0.20	0.50	0.09	0.19	−0.10
17. I can voice a contrary opinion without the fear of negative consequences	0.78	0.18	0.15	0.09	0.29	0.07	−0.03
18. When I speak up at work, I feel my opinion is valued	0.68	0.43	0.08	0.20	0.23	0.04	−0.08
19. I rarely feel like I am “the only one” (gender, sexual orientation, race, etc.)	0.14	0.03	−0.01	0.05	0.82	0.12	0.10
20. I often worry I don’t have things in common with others (Reverse-scored)^b	0.01	0.10	0.05	0.03	0.02	0.03	0.76
21. I feel like my colleagues understand who I truly am	0.55	0.04	0.25	0.35	0.21	−0.04	0.32
22. My company enables me to balance my work and personal life	0.72	0.06	0.22	0.03	0.11	0.08	−0.12
23. During an interview a job candidate meets culturally diverse employees^b	0.27	0.21	0.46	0.10	0.47	−0.07	−0.17
24. Everyone has access to equal employment opportunities in the company^b	0.49	0.20	0.48	0.15	0.26	−0.11	−0.15
25. I feel connected to others in my organization	0.17	0.20	0.18	0.82	0.13	0.03	0.09
26. The culture of my organization well-defined^b	0.63	0.13	0.51	0.20	0.10	0.06	−0.09
27. The norms (e.g., dress code, appearance, behavior) are clearly understood	0.31	0.14	0.61	0.24	−0.01	−0.16	0.08
N = 141. Initial Eigenvalues^b	10.77	2.03	1.78	1.32	1.22	1.11	1.08
Percentage of variance accounted for (%)	40	8	7	5	5	4	4

^a Responses using a 6-point response scale: 1 = strongly disagree to 6 = strongly agree;

* Factor loadings above 0.40 bolded

^b italicized items, #1, 6, 10, 13, 16, 20, 23, 24, 26, deleted due to either multiple factor loadings, insufficient items for a scale

^c Factor 1 = Be Myself – 7 Items, # 11, 14, 15, 17, 18, 21, 22

^c Factor 2 = Acceptance – 4 items, # 2, 3, 4, 5

^c Factor 3 = Diversity Valued – 3 items, #7, 9, 27

^c Factor 4 = Connection – 2 items, # 12, 25

^c Factor 5 = Safety – 2 items, #8, 19

Table 3. Means, standard deviations and correlations of continuous study variables.

Variable name	M	SD	1	2	3	4	5	6	7	8	9
1. Age^a	3.08	1.24	(NA)^e
2. Political Scale^b	3.09	0.91	−0.27**	(NA)
3. Organization Tenure^c	2.34	1.03	0.34**	−0.06	(NA)
4. Be Myself^d	5.05	0.62	0.08	−0.18*	0.02	(0.89)
5. Acceptance^d	5.35	0.68	0.07	−0.17*	0.01	0.59**	(0.84)
6. Diversity Valued^d	5.28	0.55	0.11	−0.26**	0.02	0.52**	0.44**	(0.71)
7. Connection^d	5.02	0.81	0.18*	−0.15	0.03	0.53**	0.48**	0.41**	(0.83)
8. Productivity^d	5.38	0.56	0.01	0.07	−0.08	0.45**	0.38**	0.30**	0.36**	(NA)
9. Intent to Stay^d	5.53	0.64	0.31**	−0.09	0.05	0.37**	0.39**	0.29**	0.35**	0.39**	(NA)

Note. N = 124.

* p<.05; ^**p<.01 (both two-tailed); ^aage, 1 = 20s to 6 = 70s; ^bPolitical Scale, 1= extremely conversative to 5 = extremely liberal; ^corganization tenure, 1 = less than 1 year to 5 = more than 20 years; ^dBe Myself, Acceptance, Valued Diversity, Connection, Productivity, Intent to Stay, 1 = Strong Disagree to 6 = Strongly Agree; ^ecoefficient alpha reliabilities in diagonal; NA: not applicable

Results

Initial survey sample

Comparing two initial survey samples before combining

Survey participants from the technology consulting company (N = 130) and accounting firm (N = 30) were compared on demographic, organizational and outcomes variables prior to combining. Results showed that participants were very similar on race (primarily White); religion (primarily Christian); political scale (M of 3.37 versus 3.34); age (M of 2.91 versus 3.06); organization tenure (M of 2.38 versus 2.26); productivity (M of 5.33 versus 5.37) and intent to stay (M of 5.38 versus 5.42). There was also a similar breakdown in position level/role. There was a notable difference in gender; Company A participants were primarily male (56%) and Company B participants were primarily female (70%). Overall, the similarities allowed these samples to be combined, into Table 1.

Frequency analyses

Table 1 reports the frequencies and percentages for nominal demographic and organization variables. Table 1 shows more participants were male, White, and Christian with a broad categorization of position level/roles.

Factor analysis of 27 OB items

Table 2 presents the results of the Exploratory Factor Analysis (EFA). The following criteria were used (Costello & Osborne, 2005): principal components method; eigenvalues greater than 1, scree test, and varimax rotation, to create more independent factors. The principal components method was used because there were no a priori assumptions about how the items would be related. The sample to items ratio was 5.2 to 1 (141: 27) which met the minimum recommended ratio of 5:1 (Suhr, 2006). There was some missing data, i.e., 19 out of 160 respondents (12%). Both the eigenvalue greater than 1 and scree tests suggested a seven-factor solution, with 73% of the variance accounted for.

There were double item loading complications (cross factor loadings of at least 0.40) on seven italicized items (#6, 10, 13, 16, 23, 24, 26), so these items were deleted. In addition, two italicized items (#1, 20) were the only items to load cleanly on a factor (#1 on Factor 6; #20 on Factor 7), so both were deleted. The more general “work” referent and not “this organization” may have contributed to item #1 not loading on Factor #4 (Bussing, 2002). Deleting nine items left 18 items, across five factors. These factors were then interpreted to mean (included items): Be Myself − 7 Items (# 11, 14, 15, 17, 18, 21, 22); Acceptance − 4 items (# 2, 3, 4, 5); Diversity Valued − 3 items (#7, 9, 27); Connection − 2 items (# 12, 25) and Safety − 2 items (#8, 19).

Beavers et al. (2013, p. 3) have argued that the size of the factor loadings is also important to consider, beyond sample size, in assessing the reliability of the factor analytic solution, “if a factor has four or more loadings of 0.60 or higher, sample size is not relevant versus the factors that have 10–12 items that load moderately (0.40 or higher), then a sample size of 150 or more is needed to be confident in the results.” Closer inspection of the factor loadings on each named factor in Table 2 shows that using this 0.60 loading criterion, Be Myself has 6 of 7 loadings; Acceptance has 3 of 4 loadings; Diversity Valued has 2 of 3 loadings; Connection has 2 of 2 loadings; and Safety has 2 of 2 loadings of at least 0.60 or greater. The overall strength of these loadings on each factor gives more confidence in the reliability of this factor solution.

Using these above-noted items to create scales, the following coefficient alpha reliability estimates for each factor were then found: Be Myself − 0.89; Acceptance − 0.84; Diversity Valued − 0.71; Connection − 0.83; and Safety − 0.54. Based on Hinkin (1995), a recommended alpha reliability threshold of 0.70 must be attained for scale formation. Using this criterion, the two-item Safety factor was deleted because of its poor reliability estimate. This resulted in four OB scales being formed using the number of items divided by the response scale for further analyses: Be Myself, Acceptance, Diversity Valued and Connection. There were no significant differences between any of the four OB scale means between Company A and Company B respondents. This allowed the two samples to be combined for the analyses below.

Means, standard deviations and correlations of continuous study variables

Table 3 shows the means, standard deviations, and correlations among continuous (measured) study variables. Age, political scale, and organizational tenure means are close to the middle of each scale with larger standard deviations. The four OB scales, Be Myself, Acceptance, Diversity Valued, and Connection, as well as the two outcome variables, Productivity and Intent to Stay, all have high means and lower standard deviations. There were no significant differences in outcomes between Company A and Company B, and using listwise deletion, the sample size was reduced from 141 to 124 due to missing data.

The four OB scales were correlated, with a range of r(122) = 0.41 to 0.59, with the strongest correlation between Be Myself and Acceptance, r(122) = 0.59, which means (0.59)² = 35% overlap. These results indicated that the scales were sufficiently distinct to be used separately (Stevens, 1996). The combination of factor analytic, reliability estimates, and correlational results support RQ1, reliable and valid OB scales can be measured. The correlation of r(122) = 0.39 between Productivity and Intent to Stay supported their distinctiveness.

General significant correlation relationships in Table 3 have been supported by prior research, i.e., the positive relationship between age and organization tenure, r(122) = 0.34 (Brimeyer, Perucci & Wadsworth, 2010); the negative relationship between age and political scale, r(122) = −0.27 (Malka & Lelkes, 2010); and the positive relationship between age and intent to stay, r(122) = 0.31 (Griffeth, Hom, & Gaertner, 2000). Showing these consistent correlations with prior research gives more confidence in the other correlations found in Table 3. These findings provide support for RQ2, positive relationships between OB scales to productivity and intent to stay.

Additional Independent T-test Results on OB Scales. Independent sample t-tests were done (all two-tailed), to explore the possible relationship of nominal variables (Table 1) to each OB scale. There are two gender differences; men were higher on Acceptance (M = 5.43) than women (M = 5.19), t(135) = −2.06, p < .05; and higher on Diversity Valued (M = 5.35) than women (M = 5.18), t(135) = −1.73, p < .10. In order to test for race, all nonwhite respondents needed to be combined, and even with combining these categories, the sample size was still small, n = 24 compared to Whites (n = 118). However, there were no significant differences on any OB scale for white versus nonwhite groups. For religion, Christian was the dominant group (n = 88), and all non-Christians needed be combined together (n = 48). There was one significant difference, Christians had higher Connection (M = 5.07) than non-Christians (M = 4.76), t(134) = 2.09, p < .05. To test for differences in position level/role, two groups were created: management, n = 55 (39%) versus non-management, n = 85, (61%). There were two marginally significant differences: management was higher on Be Myself (M = 5.13) than non-management (M = 4.92), t(137) = 1.75, p < .10, and on Connection (M = 5.11) than non-management (M = 4.83), t(137) = 1.85, p < .10.

Testing the third research question

RQ3 stated OB scales will contribute to explaining productivity and intent to stay beyond controlled for demographic and organization variables. Hierarchical regression was used to test this research question, and as noted above, race, religion, and position level/role needed to be combined into two groups (Stevens, 1996). A three-step process was used, first controlling for the set of demographic variables (Step 1) and then organizational variables (Step 2) before entering all four OB scales together (Step 3). This final step will determine the relative “importance” of these scales when competing for explaining an outcome (Stevens, 1996). Due to missing data, the sample size was reduced from 124 to 120. However, the respondents/predictors or N/K ratio of 120/11 used exceeded the recommended 10:1 ratio (Stevens, 1996) for regression analyses.

Prior to regression analyses, using case-wise deletion, independent sample t-tests were done for complete-data versus missing-data respondents across all variables shown in Table 4. The results showed three significant t-test results. Complete data respondents had higher means (M) than missing data respondents on three of the four OB scales: Be Myself, complete data M = 5.05, incomplete data M = 4.66, t(141) = 2.28, p < .05; Acceptance, complete data M = 5.35, incomplete data M = 5.00, t(140) = 2.01, p < .05; and Connection, complete data M = 5.02, incomplete data M = 4.48, t(140) = 2.54, p < .05.

Table 4. Final hierarchical regression models for incrementally testing the contributions of organizational belonging scales for explaining productivity and intent to stay.

Outcomes	Productivity^h					Intent to stay^h
	b	SE	R^2	Chge R^2	F	b	SE	R^2	Chge R^2	F
Step 1: Demographic variables
Gender^a	−0.02	0.10				−0.27**	0.11
Race^b	−0.09	0.13				0.14	0.15
Age^c	0.01	0.04				0.20**	0.05
Political Scale^d	0.10+	0.06				0.01	0.06
Religion^e	0.04	0.11				0.19	0.12
			0.02		0.40			0.11**		2.92*
Step 2: Organization variables
Organization Tenure^f	−0.02	0.05				−0.03	0.05
Position Level/Role^g	0.19+	0.10				0.27*	0.11
			0.03	0.01	0.88			0.14*	0.03	2.64*
Step 3: Organizational belonging scales
Be Myself^h	0.31**	0.11				0.14	0.12
Acceptance^h	0.10	0.09				0.25**	0.10
Diversity Valued^h	0.03	0.11				0.02	0.12
Connection^h	0.10	0.07				0.12	0.08
			0.28**	0.25**	30.88**			0.35**	0.21**	5.19**
(Adjusted R^2)			(0.21)					(0.28)

Note. N = 120. b is unstandardized regression weight; SE: standard error; +p<.10; ^*p<.05; ^**p<.01; all two-tailed.

^a Gender, 1 = woman, 2 = man; ^brace, 1= White, 2 = Nonwhite; ^cage, 1 = 20s to 6 = 70s; ^dpolitical scale, 1= extremely conversative to 5 = extremely liberal; ^ereligion , 1= Christian, 2 = Non-Christian; ^forganization tenure, 1 = less than 1 year to 5 = more than 20 years; ^gposition level/role, 1 = management, 2 = non-management; ^hbe myself, acceptance, diversity valued, connection, productivity, intent to stay, 1 = strong disagree to 6 = strongly agree.

The final regression models for explaining Productivity and Intent to Stay are shown in Table 4. Total variance explained (R²) and overall F values with successive steps, as well as changes in variance explained by each step are reported. At the end of each model, total variance explained, an overall F value, and the Adjusted Variance (R²) for shrinkage are reported.

Five demographic variables were entered first for explaining Productivity: Gender, Race, Age, Political Scale and Religion. Collectively, these variables did not explain a significant amount of Productivity variance, R² = 0.02, F(5, 114) = 0.40, p = .85. One variable, Political Scale, was marginally significant, b = 0.10, t(118) = 1.79, p < .08, such that liberals perceived higher Productivity. In the second step, two Organization variables were added, Organization Tenure and Position Level/Role. Collectively, these variables explained a non-significant additional proportion of Productivity variance, R² = 0.01, and the overall F(7, 112) = 0.42, p = .88. One variable, Position Level/Role, was marginally significant, b = 0.19, t(118) = 1.84, p < .07, such that non-management respondents perceived themselves as more productive. Finally, in the third step, all four OB scales were entered separately. Collectively, these variables explained a significant proportion of the Productivity variance, R² = 0.25, with an overall F(11, 108) = 3.88, p < .01. Be Myself had a significant positive impact on Productivity, b = 0.31, t(118) = 2.96, p < .01, such that higher perceptions led to higher Productivity. The model explained 28% of the Productivity variance, and 21% adjusted for shrinkage.

Again, five demographic variables were entered first for explaining Intent to Stay: Gender, Race, Age, Political Scale and Religion. Collectively, these variables did explain a significant amount of the Intent to Stay variance, R² = 0.11, F(5, 114) = 2.92, p < .05. Two variables, Gender and Age, were each significant: Gender, b = −0.27, t(118) = −2.59, p < .01, such that women had a higher intent to stay; and Age, b = .20, t(118) = 4.23, p < .01 such that older employees had a higher Intent to Stay. In Step 2, two Organization variables were added, Organization Tenure and Position Level/Role. Collectively, these variables explained a non-significant additional proportion of the Intent to Stay variance, R² = 0.03, and the overall F(7, 112) = 2.64, p < .05. One variable, Position Level/Role was significant, b = 0.27, t(118) = 2.37, p < .05, such that non-management respondents had a higher intent to stay. Finally, in the third step, all four OB scales were entered separately. Collectively these variables explained a significant proportion of the Intent to Stay variance, R² = 0.21, with an overall F(11, 108) = 5.19, p < .01. Only Acceptance had a significant positive impact on Intent to Stay, b = 0.25, t(118) = 2.57, p < .01. Overall, the model explained 35% of the variance in Intent to Stay, and 28% when adjusted for shrinkage. Collectively, these regression results provide partial support for RQ3 – OB scales will contribute to explaining productivity and intent to stay beyond controlled for demographic and organization variables.

Validation sample

Frequency breakdown by position level/role

Using the same position level item as the initial survey, the following frequencies (percentages) were found: owner/partner, n = 0 (0%); C-Suite, n = 2 (5%); Vice President, n = 3 (7%); Director, 5 (12%); Manager/supervisor, n = 6 (14%); Technical staff, n = 5 (12%); Consultant, n = 10 (24%); Sales, n = 5 (12%); Administrative support, n = 4 (10%); Other (fill in), n = 2 (5%). A comparison of this frequency breakdown to Table 1 showed general consistency. When comparing management/non-management collapsed categories, for the initial sample it was 39%/61% and for the validation sample, 38%/62%.

Coding responses matched to four OB scales

The authors met online for approximately two hours to collectively code answers given by the 42 respondents. There were 59 coded responses for the 42 respondents, as some wrote more phrase/sentence-to-scale matches. Two respondents had to be discarded because of no scale match leaving an n = 40. Across these respondents there were 57 codable responses, with the following scale match frequencies (percentages) and sample item examples were found: Be Myself − 14 (25%), “knowing my feedback is heard (good or bad) from my manager” and “I feel my voice and opinions are sought out by others in the company”; Acceptance − 12 (21%), “I feel that I am included and involved in things that help the company be successful” and “having a voice in our wonderful caring community regardless of position”; Diversity Valued − 3 (5%), “there is a culture of mutual respect” and “a culture that encourages open communication and collaboration”; and Connection − 28 (49%), “we are all in it together, one voice!” and “I like the fact that the people I interact with look after one another.” Overall, this content analysis of open items further validated the four OB scales (Hinkin, 1995).

Discussion

Overall results

Most important was finding initial support for four distinct reliable and valid OB scales – Be Myself, Acceptance, Diversity Valued, and Connection. These scales are new to the OB literature and were primarily based on combining prior quantitative (e.g., Hagerty & Patusky, 1995; Stamper & Masterson, 2002) and qualitative (e.g., Belle et al., 2015; Herbert, 2020) studies, with support from a content analysis of open item responses from a small validation sample. Once establishing these scales, other important results included: (1) showing several demographic differences on these OB scales; (2) finding that all four OB scales were significantly positively correlated with productivity and intent to stay variables; and (3) when these OB scales competed against each other (in the final step of the regression analyses), showing the significant impact of Be Myself on productivity and Acceptance on intent to stay, after first controlling for demographic and then organizational variables.

Finding high mean levels for all four OB scales speaks well for the organizational cultures of both Companies A and B. Follow-up t-tests showed that males were significantly higher on two of the four scales, i.e., Acceptance and Diversity Valued. There were no significant differences on any OB scale for race, Christians had higher perceived Connection than non-Christians, and management level employees perceived higher Be Myself and Connection than non-management employees. It may be easier to “speak freely” when you have more authority. The diverse t-test and regression results add to the credibility of OB scale distinctiveness.

Study strengths and limitations

Perhaps the biggest concern is the validity of the new four OB scales, Be Myself, Acceptance, Diversity Valued and Connection. The sample to items ratio of 5.2–1 used for the exploratory factor analysis met the minimum recommended ratio of 5:1 (Suhr, 2006). However, the strength of the factor loadings on the 16 retained items was very strong, supporting the reliability of the factor analytic solution (Beavers et al., 2013). Follow-up scale formation showed good internal consistency estimates for all four scales, including the two-item Connection scale, and scale inter-correlations showed the four OB scales were distinct. Of the four scales, Diversity Valued had the lowest internal consistency (alpha = 0.71) and the fewest matched open item responses. There were several diversity-related items asked (Table 2), e.g., “meeting culturally diverse employees” (#23) and “access to equal employment opportunities” (#24), which had to be deleted due to double loading complications. However, the demographic breakdown of the overall sample showing it to be primarily White and Christian (Table 1) may have contributed to these complications, and the weaker open item response frequency. Given the need for diversity, equality and inclusion (Herbert, 2020) as part of OB, these items should be retained for future testing, as they are part of the belonging construct (May, 2011).

Future research with larger, more diverse samples is needed to further validate these scales. The research design was limited by the overall homogeneity of the samples, despite being from two different organizations. In addition, although survey anonymity did not allow for identifying which employees worked remotely versus in-person, since such a large percentage (77%) of Company A employees were working remotely during the initial data collection (although this percentage was not known for Company B employees), it must be assumed that many who filled out the survey were working remotely. Thus, this new OB scale would seem to be applicable to the remote workplace. This data was collected during the pandemic, and research (Yang et al., 2022) has shown that remote work has made collaboration more difficult during the pandemic, as well as leading to a decline in relatedness needs being met, e.g., feeling part of the group (Schade, Digutsch, Kleinsorge & Fan, 2021). However, De Smet, Dowling, Mugayar-Baldocci, and Spratt (2022) argued that “it’s not about the office, it’s about belonging” citing their McKinsey international survey findings about pandemic-era employees who left their jobs primarily due to not feeling valued or having a sense of belonging. By uneasy contrast, surveyed employers (i.e., chief talent officers/managers) overlooked these relational elements, instead citing better compensation and work-life balance as the main reasons for why they thought employees quit.

Due to survey response anonymity, the research design is based on cross-sectional self-reported data. Spector (2019) has argued that cross-sectional research designs can still provide evidence for variable relationships, and that inflated results due to self-report common method variance is often overstated, particularly when relevant control variables (such as gender, age, race, political scale, religion, organization tenure, position level/role used here) rule out spurious relationships (Spector, 2021). A one-factor test showed that 30% of the variance was captured by the first factor, which is not inconsistent with prior studies (Spector, 2019). Both outcomes, productivity and intent to stay, were one-item, which did not allow for computing a reliability estimate. However, prior OB research has also used one item measures of outcomes, such as intent to stay (e.g., Davila & Garcia, 2012; Huynh et al., 2012). Recently, Matthews, Pineault and Hong (2022) have defended the use of such one-item measures. However, multi-item self-report measures of productivity and intent to stay, allowing for reliability estimates, are needed to confirm these initial results. Collecting other source data, such as supervisor-rated productivity and employee turnover records, would be ideal.

The research design allowed for controlling a greater number of demographic and organizational level variables before testing for the impact of OB scales on outcomes, compared to previous research (e.g., Knapp et al., 2014; Stamper & Masterson, 2002; Van Dyne & Pierce, 2004). However, small sub-samples within several variables necessitated aggregating specific categories together into a more general category for t-tests and regression analyses, i.e., under race, the nonwhite race category (combined all other racial categories), under religion, non-Christian (combined all other religions), and under position level/role, two broadly-aggregated categories, i.e., management versus non-management. Combining smaller sub-samples into a larger category has been done, e.g., race (Stamper & Masterson, 2002). Future research, using larger, more diverse samples, would not necessitate such aggregating. Despite these limitations, a promising 16-item OB Measure was developed, composed of four scales: Be Myself, Acceptance, Diversity Valued and Connection.

Implications for research and practice

Future research is needed comparing OB and outcome (e.g., performance, retention) impact for in-person, hybrid, and remote employees, as well as studying generational differences. Having large enough samples to specifically isolate work location, i.e., in-person, hybrid and remote, to then compare them on OB dimensions is critical (De Smet, Dowling, Mugayar-Baldocci & Spratt, 2022). In addition, Gurchiek (2022) noted generational differences in likeliness of leaving because of loneliness: Generation Z (54%); Millenials (52%); Generation X (42%); and Baby Boomers (31%). Herbert (2020) found that 78% of employees who felt like they belonged rated their well-being favorably versus 28% who felt they didn’t belong rating their well-being favorably. From a more extreme perspective, Boccio and Macari (2013) argued that fostering a sense of belonging can prevent workplace suicide.

In terms of practice, perhaps the most powerful OB scale for employees is Be Myself. This represents the freedom to voice a contrary opinion without fear of negative consequences and being your authentic self, regardless of employee diversity (e.g., race, gender identity, culture and ethnicity, health, disability, generation). As noted, missing sample respondents were found to be lower on Be Myself, Acceptance and Connection. There were also mean sub-sample differences on OB scales, e.g., management level employees perceived higher Be Myself and Connection than non-management employees; males were significantly higher on two of the four scales, i.e., Acceptance and Diversity Valued and Christians had higher perceived Connection than non-Christians. Given the sensitivity of testing for any employee individual differences in perceived OB, using an anonymous survey technique helped to surface such differences. However, finding initial group differences in employee perceptions of OB raises follow-up ethical concerns, for example, how to address these differences without singling out employees? First, it is important for any organization to make each employees feel safe enough so that data can be collected from ideally all employees when measuring OB. Then, follow-up from a trustworthy pro-active Human Resource Department could supplement a survey with other more personalized data collection techniques, including focus groups and interviews. Finally, different feedback forums such as department meetings and general town halls would allow for all employees to participate and voice their opinions. Any demographic differences in perceived OB, for example as found in this study for gender and position level, presents an opportunity for Human Resource Departments to work on programs that are targeted to the needs of these specific groups.

Conclusion

Given the challenges faced by many organizations today to not only retain current staff but to also hire additional employees (Laker, 2022), increasing employees’ feelings of belonging to their organization is critical, i.e., a “sticky” workplace (De Smet, Dowling, Mugayar-Baldocci & Schaninger, 2022). This seems to be regardless of whether an employee is working in-person, remotely or hybrid (De Smet, Dowling, Mugayar-Baldocci & Spratt, 2022). A review of the literature revealed that there was not an existing comprehensive scale for measuring OB, but instead a construct measured by combining three variables: insider status, psychological identification, and organizational identification, without research investigating if each facet should be equally weighted. Related measures such as workplace belongingness, employee belongingness, organizational connectedness and general belonging were also found lacking. Study results re-affirmed the definition of OB by finding initial empirical support for a new four-scale, 16-item measure, labeled Be Myself (seven items), Acceptance (four items), Diversity Valued (three items) and Connection (two items). Two OB scales had significant impacts on productivity (Be Myself) and intent to stay (Acceptance). As such, this study adds to the OB and workplace health literature, where additional study is needed, including future research further testing the validity of the OB scale using more diverse samples.

Disclosure statement

No potential conflict of interest was reported by the author(s).