One of the most critical processes of any organization is the ability to integrate new hires, increasing their capacity to do the job while also assimilating them into a company’s culture and social fabric. Organizational socialization is the process by which individuals learn the knowledge, skills, and behaviours needed to be successful in a new role (Bauer & Erdogan, 2010; Saks & Ashforth, 1997b). Socialization - often referred to as onboarding - is the process whereby employees go from being perceived as outsiders to insiders (Bauer, et al., 2007). Research has shown that employee socialization plays an increasingly prominent role in an individual’s adjustment to their job, work team, and overall organizational culture (Ashforth, Sluss, & Saks, 2007). Adequate socialization has become even more essential as employees are considerably more itinerant today, holding an average of 12.4 jobs throughout their lifetime (U.S. Department of Labour, 2021). Given that workers are trained into new roles more often, new hires must be adequately socialized so they are prepared for their roles.

Although implementing any form of socialization is connected with positive work outcomes, how you socialize your employees is also important (Bauer & Erdogan, 2010; Saks & Ashforth, 1997b). While some organizations implement a structured and systematic approach to socialization, others incorporate informalized sink or swim tactics, hoping that a new employee stumbles into the knowledge and relationships needed to succeed. The problem with this approach is that it engenders role ambiguity, stress, and lower organizational attachment (Bauer, et al., 2007; Chong, et al., 2020; Saks, et al., 2007). Alternatively, formalized socialization tactics can help reduce uncertainty and ambiguity in one’s role by instilling feelings of competence, providing social support in the form of mentors and role models, and providing necessary information regarding one’s role (Chong, et al., 2020).

A sink or swim socialization strategy is not a new concept. It is an ideology of socialization that has been perpetuated and romanticized in popular culture, as phrases like sink or swim, make it or break it, thrown in the deep end, or trial by fire are widespread and often recycled within an organization’s senior management. Within the socialization literature, sink or swim is primarily mentioned in association with the organizational socialization tactic (OST) model. This model was created by Van Maanen and Schein (1977) and includes six dimensions that describe various tactics undertaken by organizations to socialize their employees and facilitate adjustment to their roles. Jones (1986) later placed the six tactics on a bipolar continuum from structured (institutionalized) to unstructured (individualized) socialization and developed a scale to measure these constructs. Unfortunately, the OST scale has some practical and psychometric issues that limit its utility for academics and practitioners (e.g., Allen and Meyer, 1990a; Ashforth and Saks, 1996; Filstad, 2011; Kowtha, 2018; Saks, et al., 2007; Wingerter and Ahn, 2020). Therefore, the goal of this research was to create an alternative and accessible measure of employee socialization that will improve upon the tenuous reliability of the OST scale.

Sink or Swim Socialization

The idiom sink or swim can refer to any situation where someone must succeed through their efforts or else fail completely (Merriam-Webster, n.d.). A review of the literature found three main themes associated with a perception of sink or swim socialization: [1] shifting of responsibility to the employee (responsibility), [2] a lack of perceived effectiveness of a socialization program (effectiveness), and [3] limited perceptions of support (support).

Employee-Shifted Responsibility

If organizations take an unstructured and apathetic approach to socialization, the locus of responsibility is naturally shifted to the employee rather than the organization. This idea is recognized in the literature, as Van Maanen and Schein (1977) described sink or swim tactics as one “by which individuals must learn how to perform the new role on their own” (1977, p. 34). Ashforth et al. (2007a, b) described a sink or swim socialization approach as one that “encourages-almost mandates-that newcomers find their own way” (2007a, p. 14). These perspectives view a sink or swim approach as an unintended consequence resulting from a lack of formality and structure in a socialization program. Thus, an apathetic approach to socialization ensures that the onus of responsibility for getting adjusted falls on the employee rather than the organization.

Organizational leaders may also hold views that cause them to intentionally place the responsibility of getting socialized on the employee. This is not completely unfounded as unstructured socialization tactics have been shown to promote creativity and innovative role behaviours in certain contexts (Bauer & Erdogan, 2010). Organizational leaders can often withhold their support to push new workers to be innovative in how they perform their jobs. Additionally, leaders may also develop a sink or swim culture to weed out poor performers and help to identify high performers (Guimaraes & Igbaria, 1992). Finally, leaders increasingly expect new hires to learn more proactively, with the mindset that new employees should know how to work or learn their roles independently (Korte, et al., 2015). Consequently, leaders avoid participating in the adjustment of the new employee, expecting that they will take proactive steps to set goals, seek out information, and develop relationships. While some benefits of informal socialization practices exist in certain contexts, this approach can also foster a sense of abandonment and role ambiguity, which can significantly undermine positive adjustment to the organization.

Perceived Socialization Effectiveness

Employees left to sink or swim might also perceive that their company’s socialization and training program has left them ill-prepared for their roles. Informal socialization programs provide limited clarity into training content, an unclear time frame or order of events, and less access to social connections to acquire information (Perrot, et al., 2014). This lack of structure can lead to lower role clarity and higher role ambiguity, which has regularly been connected with negative outcomes like burnout and lower job satisfaction and job performance (Madlock & Chory, 2014; Papastylianou, et al., 2009; Saks, et al., 2007; Tubre & Collins, 2000). Ambiguity in one’s role creates stress, which can divert mental energy away from focal job responsibilities, reducing one’s mental capabilities to perform (Tuten & Neidermeyer, 2004). Organizations may not have any form of a socialization program, having employees start their roles almost immediately and forcing them to learn on the job (Schnepf, et al., 2016). One study that surveyed hospitality staff in New Zealand found these training tactics to be highly ineffective and commonly misused (Poulston, 2008). “Show as you go” methods or ineffective buddy systems were consistently used as cheap substitutes for formal and systematic training systems. Workers have displayed a preference for formalized socialization tactics, given that they provide structure, guidance, and information, that reduces newcomer anxiety and stress (Gruman & Saks, 2011). Overall, this lack of clarity will cause employees to perceive their organization’s socialization program as ineffective.

Socialization Support

A key component of sink or swim is that an employee will feel as though they do not have the necessary social support required to learn their roles. The literature consistently references sink or swim socialization as an isolating experience with employees learning the information necessary for their role, mostly on their own. Van Maanen and Schein (1977) referenced a sink or swim approach where individuals “are left to their own devices” (p. 63) to learn and adapt to their roles. Allen and Shanock (2013) described a sink or swim environment as a “lack of such [structured] activities may leave newcomers with a sense that they are on their own” (p. 354). This method of training implies that these employees lack the support needed to be properly socialized.

Being left to sink or swim can lead an employee to perceive that they lack the support of both an organization and its leaders. By not engaging employees in a socialization program, a hiring organization may be signalling to their new hires that they do not value their contributions (Allen & Shanock, 2013). Training investments have been connected with perceived organizational support (POS), as it is a signal to employees that they are worthy of such an investment (Johlke, et al., 2002). This lack of a signal may cause employees to feel less psychologically secure in their roles and less willing to reach out to other organizational insiders (Perrot, et al., 2014). Organization leaders are also essential to developing a positive transfer climate whereby employees may feel safe employing skills they are learning in practice (Salas, et al., 2012). The more employees perceive their post-training environment and supervisor to be supportive, the better their learned skills are practiced and maintained within the workplace (Rouiller & Goldstein, 1993).

Organizational Socialization Tactics

Van Maanen and Schein’s (1977) model of socialization tactics is one of the more theoretically advanced and popular models of socialization and includes a typology of six tactics. Collective (vs. individual) socialization tactics involve putting newcomers into a group and moving them through a standard set of onboarding activities, rather than leaving them on their own. Formal (vs. informal) socialization tactics involve the process of separating new hires from existing employees for a specified period, rather than having no formal separation. Sequential (vs. random) tactics relate to the degree to which an organization sets a predetermined sequence of distinct training steps leading to realizing the role, rather than having unknown or ambiguous stages. Fixed (vs. variable) socialization refers to the degree to which a timetable is associated with specific steps, with variable including no set timetable. Serial (vs. disjunctive) tactics include assigning a mentor to groom new employees into a similar position, whereas disjunctive assigns no such mentors. Finally, investiture (vs. divestiture) seeks to confirm a new hire’s identity and individual characteristics rather than deny and strip them away for assimilation purposes.

Jones (1986) later condensed these six socialization tactics into three broad factors that exist on a continuum, which he termed social, content, and context. On one side of the continuum is institutionalized socialization tactics, consisting of collective, formal, sequential, fixed, serial, and investiture. These tactics represent a formalized approach whereby an organization employs a structured socialization program for incoming employees (Jones, 1986). Institutionalized socialization tactics are effective at helping employees learn their new roles and enhance their organizational attachments. On the other side of the continuum is individualized socialization tactics, made up of individual, informal, random, variable, disjunctive, and divestiture. Alternatively, this is a sink or swim style of socialization, as it is characterized by its informality and absence of structure (Jones, 1986). These tactics are expected to encourage newcomers to question the status quo and take an innovative approach to how they adjust to their roles.

Critique of Jones’ (1986) Model of Socialization

Theoretical limitations

One theoretical concern with the OST scale is that it tends to take an organizational perspective of socialization, focusing on the tactics they employ rather than an employee’s perspective of their socialization experience. It is important to gather insight into an employee’s perspective of the quality of their socialization experience, as it provides context into important outcomes at an organization, like their attachment, engagement, performance, and satisfaction. Furthermore, sub-factors relating to social support tactics, widely considered to be the strongest predictor of employee adjustment, have demonstrated several inconsistencies (Saks, et al., 2007). More specifically, the tactic investiture-divestiture has been generally misrepresented in the literature. Van Maanen and Schein originally defined this construct as a form of “identity verification,” yet Jones’s (1986) conceptualization of this factor more closely resembles a measurement of “social support” (Filstad, 2011). This is problematic as researchers continue to use these definitions interchangeably, resulting in limited clarity of the construct. This factor has also been shown to load onto both institutionalized and individualized factors, placing doubt into the 1-factor theoretical conceptualization (Ashforth, Sluss, & Harrison, 2007). Additionally, the construct of serial-disjunctive focuses on the assignment of formal mentors but fails to account for informal support networks from other organizational insiders. While being assigned formal mentors is no doubt important, informal support networks from peers are equally as important, as positive work relationships significantly increase skill transfer, commitment, and job satisfaction (Allen, et al., 2017; Chiaburu, et al., 2010; Chiaburu & Marinova, 2005; Rhoades, et al., 2001).

Psychometric limitations

The OST scale has also demonstrated various psychometric issues. First, the sub-dimensions of the OST scale have historically displayed fairly inconsistent reliabilities across studies as individual factors have regularly displayed alphas below 0.70 (e.g., Allen and Meyer, 1990a; Ashforth and Saks, 1996; Filstad, 2011; Kowtha, 2018; Saks, et al., 2007; Wingerter and Ahn, 2020). This has been compounded by the fact that multiple researchers have elected to use shortened versions scale, challenging its already tenuous reliabilities and making it difficult to ensure conceptual consistency of the socialization tactics across studies (e.g., Ashforth and Saks, 1996; Cable and Parsons, 2001; Nasr, et al., 2019; Saks, et al., 2007; Saks and Ashforth, 1997a; Wang, et al., 2017). The tactic, formal-informal has been particularly unstable as even an attempt to adapt and improve upon this factor still resulted in an alpha of 0.63 (Ashforth, Sluss, & Harrison, 2007; Filstad, 2011). Finally, attempts to re-conceptualize the OST model into 3-factors have also continued this trend of poor reliabilities, specifically with the context factor demonstrating historically poor alpha’s (e.g., Filstad, 2011; Saks & Gruman, 2011; Wang, et al., 2017).

Practical limitations

The OST scale also has several practical limitations. First, the OST scale was developed in 1986 and has remained mostly unchanged since then; therefore, the items may not reflect the tactics that organizations currently employ. For example, socialization has become significantly more decentralized given the advancement of digital technologies in employee training (Wiseman, et al., 2022). This has been heightened by the COVID-19 pandemic, as organizations have been forced to adapt to the increasing use of home offices. Traditional tactics of socialization where employees move through group-based in-person training experiences may not be relevant to a changing work landscape. Second, for organizations and practitioners, the OST scale may be overly specific and complex for their needs. The 6-factors of the OST scale use academic language and target very particular socialization tactics. For example, the OST scale uses factor labels that include investiture (vs. divestiture) or serial (vs. disjunctive), terminology that is not regularly used in a practitioner setting. Multiple researchers have elected to use the OST scale as 1-factor, demonstrating their limited need for the precision of a 6-factor model.

Therefore, the goal of this study was to develop an alternative measure to the OST scale, based on a conceptual model of sink or swim. This new measure was designed to offer a simple and practical scale for academics and practitioners. This research accomplished this across two studies. Study 1 involved generating items for the scale, followed by an initial content validation assessment. Study 2 used a longitudinal survey design to assess the factor structure, psychometric properties, and predictive validity of the SSS.

Study 1: Scale Construction and Content Validation

The purpose of study 1 was to generate an initial set of items to make up the three factors of the SSS. These items were developed through a content analysis of the relevant literature, the use of both quantitative and qualitative data from an industry survey, and a re-evaluation of the original OST scale. The initial item pool was run through an item-matching task to assess content validity and improve the conceptual consistency of the sub-factors.

Methods

Item Generation

Three strategies were used to generate the initial 36-item pool. The first approach followed a deductive strategy, where the academic literature was scanned for mentions of the phrase sink or swim and items were generated and sorted into one of the three scale sub-factors, resulting in 13 items. The second strategy involved intercorrelations between a pilot 2-item measure of sink or swim and each of the items in the OST scale. This was done to assess which OST items had some level of conceptual relevance to a perception of sink or swim. Seven items from the OST scale were reworded and included as they demonstrated strong correlations. The third strategy included a content analysis of open-ended question response data from an industry survey, which focused on topics relating to organizational training (termed ‘onboarding’). One independent coder pulled relevant comments that were then rewritten into items and reviewed and assigned into one of the three sub-groups of sink or swim, resulting in the contribution of 16 items. The initial item pool was reviewed by three faculty members at a Canadian University to ensure conceptual relevance, redundancy of item content, and face validity, resulting in 36 items.

Participants

Participants for the industry survey were recruited via convenience sample through a partnership with a Canadian technology association. This project surveyed 57 employees and 29 managers within technology firms across Nova Scotia, with some participation from US companies. The respondents were 43.0% (n = 37) female and 48.8% (n = 42) male, with an average age of 37.3 years old. Managers and employees each received a separate link with a unique survey with open-ended questions specific to their employment level (staff vs. manager).

Content Validity Assessment

A total of 18 graduate-level psychology students across multiple disciplines were recruited to act as subject matter experts (SMEs) for an item-matching task. These SMEs were provided with a document with the definitions of the three constructs—responsibility, effectiveness, and support—and a list of the initial pool of 36 items. Each SME was asked to sort the items into their respective constructs and provide additional feedback. Items with an agreement index above the minimum threshold of 75% on the correct construct were retained. Any items that fell below the 75% agreement index were either reworded based on feedback and retained or removed.

Results and Discussion

An initial item pool of 36 was derived from three main sources: deductively through a literature review, correlations with the OST measure, and a content analysis of open-ended question data. Seven items were automatically removed as they fell below the 75% agreement index threshold as recommended by Hinkin (1998). Two items were moved from the effectiveness factor to the responsibility factor, as they indicated over a 75% agreement index on this unintended construct. Another two items close to the 75% threshold were reworded based on SME feedback and retained. SME feedback also highlighted 6 items that demonstrated redundancy, so these were removed. No identifiable patterns of unexpected loading were found within the error agreement indices across each of the factors. This demonstrated that each of the 3 sub-dimensions demonstrated a suitable level of conceptual independence, with no two factors illustrating a pattern of undue conceptual overlap. A total of 23 items were retained for participation in the next stage of the study. Each of the subcategories support and responsibility was represented by 8 items with the effectiveness subcategory being represented by 7 items. These steps ensure that the items represent a reasonable measure of sink or swim socialization as represented by the three hypothesized sub-categories.

Study 2: Psychometric Assessment and Scale Validity

Factor Structure

Overall, the literature suggests the existence of three distinct elements of sink or swim socialization: [1] a shift of responsibility, [2] limited effectiveness, and [3] a lack of socialization support. The responsibility factor represents an organization’s failure to take responsibility for socialization resulting in a perceived breach of an employee’s psychological contract (Tekleab, et al., 2013). The effectiveness involves a process whereby an employee evaluates the quality of their company’s socialization program (Jones, 1986). Finally, the support factor reflects an employee’s perception of support from both their organization and its leaders during its socialization program. It is expected that the new measure will validate this 3-factor model of the perceptions of sink or swim socialization.

Hypothesis 1

There are three distinct factors related to sink or swim socialization: (1) responsibility, (2) effectiveness, and (3) support.

Construct Validity

To determine if the SSS measures what it intends to, this study ran analyses for both convergent and discriminant validity. The OST scale was used to assess convergent validity, which evaluates the extent to which a scale correlates with other similar measures. It is predicted that a perception of sink or swim will demonstrate convergent validity with a perception of individualized socialization tactics, given its unstructured approach. Both the SSS and OST scale focus on the level of formality of a socialization program, thus it is expected that they will share significant variance. Perceived supervisor support (PSS) and perceived organizational support (POS) also served as measures of convergent validity, as sink or swim is expected to measure a lack of support more pointedly at the time of socialization. Supervisors often act as “socializing agents,” acting as a source of information and support that employees need for a successful socialization experience (Takeuchi, et al., 2021). Similarly, perceiving one’s organization as supportive ensures that employees are comfortable in seeking out formal sources of information within their organization, resulting in a positive socialization experience. Thus, the perception of sink or swim socialization will be negatively related to PSS and POS.

Hypothesis 2a:

Sink or swim socialization is positively correlated with individualized socialization.

Hypothesis 2b:

Sink or swim socialization is negatively correlated with perceived supervisor support (PSS).

Hypothesis 2c:

Sink or swim socialization is negatively correlated with perceived organizational support (POS).

Discriminant validity assesses the degree that a scale does not correlate with dissimilar measures. Both proactive personality and self-efficacy were expected to demonstrate discriminant validity with the SSS. Proactive personality is defined as the degree to which a person takes control to influence their work environment (Bateman & Crant, 1993). Although a proactive personality may help employees benefit from the ambiguity of a sink or swim environment, one’s disposition operates independently of a company’s socialization tactics. General self-efficacy is defined as an individual’s belief in their abilities (Chen, et al., 2001). Similarly, while self-efficacy may help employees cope with ambiguity in the workplace, it is a dispositional trait, one that is independent of an organization’s socialization tactics. Therefore, sink or swim socialization is not expected to share similar variance to an employee’s proactivity or general self-efficacy.

Hypothesis 3a:

Sink or swim socialization demonstrates a non-significant or small correlation with employee proactivity.

Hypothesis 3b:

Sink or swim socialization demonstrates a non-significant or small correlation with employee self-efficacy.

This study also evaluated the temporal stability of the scale through an evaluation of test-retest reliability (Guttman, 1945). It was expected that each of the three factors of sink or swim socialization will indicate test-retest reliability across time 1 and 2 data.

Hypothesis 4:

Sink or swim socialization demonstrates a significant positive correlation between the surveys at time 1 and time 2.

Predictive Validity

The socialization literature has found that formalized socialization systems tend to be connected with better employee outcomes, including positive organizational attitudes and less intention to leave (Bauer & Erdogan, 2010; King, et al., 2005). The 3-factors of sink or swim imply that an organization has taken an unstructured or informal approach to socialization. A higher perceived level of sink or swim means that employees believe that their organization’s socialization program is ineffective and unsupportive. These negative attitudes regarding an organization’s socialization systems are likely to lead to a decreased perceived competency in one’s role (Saks & A. Gruman, 2014). Those who have less clarity in their role will be less satisfied and committed to their organization, and more likely to leave (Saks, et al., 2007). Perceptions of sink or swim will also be considered a breach of an employee’s psychological contract, motivating employees to offer less of their affective resources to the organization (Delobbe, et al., 2016). Based on these findings, this study predicts that perceptions of being left to sink or swim within a socialization program will lead to adverse employee outcomes.

Hypothesis 5a:

Sink or swim socialization is negatively related to employee job satisfaction.

Hypothesis 5b:

Sink or swim socialization is negatively related to employee affective commitment.

Hypothesis 5c:

Sink or swim socialization is positively related to an employee’s intention to quit.

Methods

Procedure

To examine the factor structure, construct validity, and predictive validity of the SSS, study 2 implemented a repeated measure longitudinal survey study. The average time elapsed between the first and second time point was 12.19 days (SD = 5.10). No participants under the age of 18 were included in this study. Participants were required to either be part-time or full-time workers and have been hired by their current organization in the last 2 years.

Participants

The 278 participants included within the first time point were recruited on Prolific across three weeks. Following this, 200 participants (71.9%) returned to complete the second survey. The participants were made up of 156 females (56.1%) and 115 males (41.4%) participants with an average age of 32.7 (see Table 1). The majority of the participants were Caucasian (n = 243, 87.4%), had an undergraduate degree (n = 126, 45.3%), and had over 10 years of working experience (n = 135, 48.6%).

Table 1 Demographic Information from Time 1 and Time 2 Samples
Full size table

Measures

The participants indicated their agreement level with each item on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree). Participants were initially asked a series of demographic-related questions related to age, ethnicity, gender, and occupation (see Tables 2 and 3). Quinn and Sheppard’s (1974) job satisfaction index was used to assess the employee’s overall satisfaction with their job (Eisenberger, et al., 1997; time 1 and 2 Cronbach’s α = 0.92). Affective commitment, which represents the emotional attachment that one has towards an organization, was measured using an 8-item scale created by Allen & Meyer (1990; time 1 α = 0.92, time 2 α = 0.91). A 3-item scale developed by Kelloway et al. (1999) was used to measure turnover intentions (time 1 and 2 α = 0.92). POS was measured by the 8-items in the Survey for Perceived Organizational Support (SPOS) as developed by Eisenberger et al. (1986; time 1 α = 0.92, time 2 α = 0.93). The SPOS was adapted to represent PSS, by replacing the word organization with supervisor across the eight items (Eisenberger, et al., 2002; Rhoades, et al., 2001; time 1 α = 0.93, time 2 α = 0.94). Jones’ (1986) 30-item measure was used to assess the employees’ perception of socialization tactics within their organization. Scores are combined so that a high score indicates institutionalized socialization and a low score indicates individualized socialization. The alphas of each of the 6 tactics were as follows: collective-individual (α = 0.64), formal-informal (α = 0.36), investiture-divestiture (α = 0.77), sequential-random (α = 0.80), serial-disjunctive (α = 0.77), and fixed-variable (α = 0.74). Bateman and Crant’s (1993) 10-item proactive personality scale was used to measure the employee’s proactivity (α = 0.90). Finally, the 8-item general self-efficacy scale (Chen, et al., 2001) was used to measure self-efficacy (α = 0.94).

Results

Factor Structure

Exploratory Factor Analysis (EFA)

To test the factor structure and refine the scale, an EFA with a principal axis oblique factor rotation (Promax) was conducted on the initial 24-item SSS (time 1). This rotation was chosen as the factors within the scale indicated intercorrelations. After a scan of the inter-item correlations, it was found that all items correlated reasonably well with each other and that none of the coefficients were excessively large (< 0.90). A total of 3 factors were found to have eigenvalues of greater than one. These three factors fell in line with the theoretical organization of the socialization construct as predicted in hypothesis 1 (see Table 4). All items intended to represent the support sub-factor demonstrated medium to strong factor loadings from 0.54 to 0.92. The factor loadings onto the hypothesized effectiveness and responsibility sub-factors also demonstrated medium to strong coefficients, ranging from 0.36 to 0.88. Within the hypothesized effectiveness subfactor, only items 1, 2, 5, 6 and 7 demonstrated factor loadings above 0.4 with no appreciable cross-loadings, so these items were retained. Items 3 and 4 did not demonstrate a high enough loading on the effectiveness factor over and above the loading on the support factor; therefore, these items were removed. Within the hypothesized responsibility sub-factor, items 1, 2, 4, 5, 7 and 8 demonstrated loadings on the appropriate factor. Item 3 indicated uninterpretable cross-loadings and item 6 indicated loadings onto a factor that could not be explained; therefore, these items were removed.

Table 2 Descriptive Statistics and Correlations for Time 1 Variables
Full size table
Table 3 Descriptive Statistics and Correlations for Time 2 Variables
Full size table
Table 4 Factor Loadings of First EFA on Initial 23-Item Pool (Time 1 Data)
Full size table

A second EFA was run using the surviving items from the first. This EFA also demonstrated 3 factors with eigenvalues of 1 or more. Support item 1 and effectiveness item 6 indicated cross-loadings on incorrect factors and were removed (see Table 5). A third EFA with the remaining 17 items demonstrated adequate loading coefficients (< 0.4) while mapping well onto the 3-factor sink or swim solution. Each of the sub-factors for support, effectiveness, and responsibility also demonstrated excellent reliability with a Cronbach’s α of 0.90, 0.88, and 0.90 respectively. Therefore, the three-factor model as predicted in hypothesis 1 was supported following three EFA analyses, with six items being removed.

Table 5 Factor Loadings of First EFA on Reduced 19-Item Pool (Time 1 Data)
Full size table

Confirmatory Factor Analysis (CFA)

A follow-up CFA was then conducted using the time 2 data (N = 200) to further assess the goodness of fit of the model and compare it to alternatives. Across all 17 items, none of the factor loadings fell below the 0.40 threshold (Hinkin, 1998; see Table 6). Five model fit indices were used to assess the overall fit of the factor structure: the model chi-square, standardized root-mean-square residual (SRMR; Hu and Bentler, 1999), the comparative fit index (CFI; Bentler, 1990), the Tucker Lewis Index (TLI; Tucker and Lewis, 1973), and the root mean square error of approximation (RMSEA; Browne and Cudeck, 1992). In the SSS, fit indices for the 3-factor model indicated a statistically significant chi-square test with a value of, χ2 (116, N = 200) = 280, p < 0.001, and an SRMR value of 0.047. The indicators CFI (0.937, 90% CI [0.07, 0.10]), TLI (0.926, 90% CI [0.07, 0.10]), and RMSEA (0.084, 90% CI [0.07, 0.10]) taken together indicate an adequate to good fit of the measurement model (Hu & Bentler, 1999). Although the 0.084 RMSEA score was beyond the 0.06 threshold, the CFI, TLI, and SRMR scores all indicated either close to or below their recommended thresholds based on the recommendations of Hu and Bentler (1999). The 3-factor model was also compared to a unidimensional model with all the factors combined and a two-factor model, with responsibility and support combined. A review of the fit indices (see Table 7) and a chi-square difference test indicated that the hypothesized 3-factor model provided a significantly better fit than the 1-factor, χ2diff(3) = 295, p < 001, and 2-factor models, χ2diff(2) = 113, p < 001. Finally, a CFA was completed on the support factor to assess its potential as a standalone measure. Fit indices indicated a strong fit due to a statistically significant chi-square test, χ2 (14, N = 200) = 43.3, p < 0.001, an SRMR of 0.029, a CFI of 0.971 (90% CI [0.06, 0.12]), a TLI of 0.956 (90% CI [0.06, 0.12]), and an RMSEA of 0.087 (90% CI [0.06, 0.12]). The results of both the EFA and CFA stages indicated that the 17-item SSS was best fit by the 3-factor model, providing support for hypothesis 1.

Table 6 Factor Loadings from CFA on the 17-Item Scale (Time 2 Data)
Full size table
Table 7 Fit Indices for 1–3 Factor Models in CFA (Time 2 Data)
Full size table

Psychometric Properties

Construct Validity

The construct validity of the SSS was assessed using bivariate correlations. The measures that were expected to hold convergent validity with the sink or swim measure were socialization tactics, POS, and PSS scale. The three sub-factors of sink or swim had significant negative correlations with each of the variables of socialization tactics, POS, and PSS (see Table 8). These results demonstrate that a higher perception of sink or swim is connected with individualized socialization, confirming hypothesis 2a. Effectiveness was found to have the strongest connection with institutionalized socialization, r = − 0.72. In addition, the more an employee perceived that they had been left to sink or swim, the less they perceived support from either their supervisor (PSS) or organization (POS), confirming hypothesis 2b and 2c. Support was found to be most strongly connected with both POS and PSS, with coefficients of r = − 0.60 and r = − 0.66 respectively.

Table 8 Bivariate Correlations with Variables for Convergent Validity (Time 2 Data)
Full size table

The scales that were expected to hold discriminant validity with the SSS were the proactive personality scale and the self-efficacy scale. Bivariate correlations between each of the 3 subfactors of sink or swim and proactivity and self-efficacy indicated that responsibility had a small significant positive correlation (r = 0.12, p = 0.05) with proactivity. Also, support indicated a small significant negative correlation with self-efficacy (r = − 0.18, p < 0.001). The remaining correlations indicated no other significant correlations (see Table 9). Overall, hypotheses 3a and 3b were supported, as the two significant correlations were only small effects.

Table 9 Bivariate Correlations with Variables for Discriminant Validity (Time 1 Data)
Full size table

Test-Retest Reliability

The test-retest reliability of the SSS was assessed using a bivariate correlation. A significant strong positive correlation between each of the three subfactors across time 1 and 2 (see Table 10) was found, with no coefficient falling below r = 0.77, p < 0.001. These findings indicate that the SSS indicates strong test-retest reliability, providing support for hypothesis 4.

Table 10 Bivariate Correlations for Test-Retest Reliability across Time 1 and Time 2 Data
Full size table

Predictive Validity

Three linear regression analyses were run to assess whether the SSS was predictive of job satisfaction, affective commitment, and turnover intentions (see Table 11). The SSS was found to account for 33.3%, 23.5%, and 22.1% of the variance of job satisfaction (F(3, 189) = 31.43, p < 0.001), affective commitment (F(3, 188) = 19.29, p < 0.001), and intention to quit (F(3, 188) = 17.80, p < 0.001), respectively. Support was found to be the only significant predictor for job satisfaction (β = − 0.64, p < 0.001), affective commitment (β = − 0.36, p < 0.001), and intention to quit (β = 0.64, p < 0.001). Taken together, sink or swim was found to be a significant overall predictor of each of the three outcome variables, providing support for hypothesis 5a, 5b, and 5c.

Table 11 Linear regression of SSS (Time 1) for job satisfaction, affective commitment, and intention to quit (Time 2)
Full size table

General Discussion

The primary goal of this research was to develop and validate a measure of the perception of being left to sink or swim within an organization’s training program. Much of the current research on training has focused on the various socialization tactics and their impact on employee outcomes. Instead, this study aimed to operationalize a form of training by default, or sink or swim and better our understanding of its unique role. Developing a measure of sink or swim socialization provides both researchers and practitioners with a valid alternative to the OST scale.

The factor structure of the SSS was supported across the two studies. In study 1, items were generated by pulling from both academic and practitioner language, ensuring its generalizability to both of these domains. Then, SMEs helped verify the content validity of the items through an item-matching exercise. The results of both the EFA and CFA analyses provided additional support for the 3-factor sink or swim model as hypothesized. A series of EFAs on the SSS highlighted factor loadings of items that represented the categories of responsibility, effectiveness, and support. The CFA confirmed that a 3-factor solution was a significantly better fit for the model than either a 2-factor or 1-factor solution. More specifically, an employee’s socialization experience can be broken down into three distinct categories: (1) the responsibility to socialize oneself, (2) the effectiveness of the program, and (2) the support one feels during the process. Additionally, the strong factor loadings of the support items were particularly notable as over half of the items that make up this factor were pulled directly from an industry survey. Thus, employees more often considered support when interpreting problems within their organization’s socialization program. This trend can also be seen within the socialization literature, as informal or formal support mechanisms are regularly connected with positive adjustment outcomes (Saks, et al., 2007). Having a supportive workplace provides a level of psychological safety, which prompts employees to seek out the required materials, skills, and knowledge required in a new role. Therefore, although a 3-factor solution to sink or swim was confirmed, the items included within the support sub-factor can act as a standalone construct.

The results of this study also highlighted the strong psychometric properties of the SSS. For one, the scale indicated strong convergent validity. The strong negative correlation between each of the sink or swim factors and institutionalized socialization tactics was an expected result, as the informal and unstructured nature of an individualized approach to training is implied by the characteristics of the 3-factor model of sink or swim (Van Maanen and Schein, 1977). Surprisingly, the effectiveness factor demonstrated only small negative correlations with PSS and POS. Looking closely, the effectiveness factor holds more conceptual differences to PSS and POS than the support factor given that it is focused on the tangible qualities of a socialization program. The assessment of the discriminant validity of the SSS was also supported. One surprising finding was the small significant positive correlation between responsibility and proactivity. Considering proactive individuals tend to take more control over their early work environment, they may be more sensitive to violations in a psychological contract, and thus quicker to claim responsibility for their own early training experience (Ashford & Black, 1996). Interestingly, individual self-efficacy also had a small significant negative correlation with support. Individuals with low self-efficacy tend to be more anxious about their work-related capabilities and thus more likely to seek information and support from their peers (Major & Kozlowski, 1997); therefore, it is possible that these workers may be more sensitive to a toxic socialization environment. Finally, an assessment of the test-retest reliability of the sink or swim measure demonstrated strong correlations between time 1 and time 2 data, indicating a good degree of temporal stability over time.

The SSS also demonstrated predictive validity due to its relationship with all of affective commitment, job satisfaction, and intention to quit. These results show that an employee’s perception of sink or swim is predictive of their self-reported commitment to their organization. These results are supported in the literature, as individualized socialization tactics are similarly connected to adverse employee outcomes (Bauer, et al., 2007; Saks, et al., 2007). These results suggest that leaving an employee to sink or swim can have a deep and early impact on the formation of the psychological contract. A lack of organizational investment in a socialization program can signal an unwillingness to invest in a new worker, giving them less confidence in the company culture (Delobbe, et al., 2016). Thus, employees who see their socialization program as ineffective reciprocate with less effort and loyalty to their organization, resulting in adverse outcomes.

Implications for Theory and Practice

This research was successful in developing a psychometrically valid and reliable measure of sink or swim. This measure does not hold some of the same limitations seen within the oft-used Jones (1986) measure of OST. First, this scale takes more of an employee-centered perspective, focusing on an employee’s global perception of a socialization program rather than individual organizational tactics. This can provide a more nuanced understanding of what socialization experiences are most effective at driving positive adjustment outcomes. For example, rather than evaluating whether one’s organization uses formal mentors, the SSS evaluates an employee’s perception of support more generally. This gives researchers and practitioners a deeper understanding of the affective processes of employees that lead to positive workplace outcomes. Furthermore, the SSS benefits from a simplified factor structure. The strong reliabilities of each of these factors indicate that they do not have the same theoretical inconsistencies found within the 6-factors of the OST scale. In simplifying the factor structure of the scale and pulling the items from academic and practitioner sources, the SSS is a more practically useful scale for both of these domains. For academics, the SSS offers a simplified factor structure that does away with some of the obsolete tactics included within the OST scale. For example, formal group-based training activities are not as common in a post-COVID-19 workplace context as organizations continue to implement digital training technologies to cut costs and accommodate those working from home. For practitioners, the scale does away with the academic terminology included within the OST scale (e.g., individualized and institutionalized), and incorporates accessible terminology relevant to an everyday workplace. This ensures the generalizability of the SSS to an industry setting, giving practitioners a tool that can be used to make improvements to their socialization programs.

The results of this study have further highlighted the importance of feeling supported during a socialization program. The majority of the support items were generated directly from employees through the industry survey, demonstrating that this is top of mind during socialization. Furthermore, research has regularly connected support socialization tactics (i.e., mentors and positive climate) as singularly important to promoting positive adjustment outcomes in workers (Mornata & Cassar, 2018; Perrot, et al., 2014; Saks & Gruman, 2011). It is particularly troubling that the OST scale’s measurement of support has proven to be tenuous at best, as it has often been represented more as a form of “identity verification” during socialization. Additionally, the OST scale does not account for supportive tactics outside of formal mentorship. Support can be felt from a variety of sources which include a supervisor, co-workers, or from a company culture more generally. Given the strong performance of the support scale items, it can operate as a standalone measure of how supported an employee feels during a socialization program. For academics, a standalone measure of socialization support can offer a tool for understanding its importance in driving positive adjustment outcomes. For practitioners, a measure of socialization support will add insight into how best to improve upon their existing training programs.

Finally, this study has provided additional insight into a highly contextual and complicated construct. It has been well established that how you train your employees holds significant weight within an organization’s ability to ensure a productive and satisfied workforce. However, some managers continue to employ sink or swim techniques perpetuating a harmful misconception that this is the only way to learn. It does not help that this phenomenon has consistently been romanticized by an endless stream of idioms like sink or swim, trial by fire, make it or break it, or thrown in the deep end. Although there is no doubt that employees must engage in some form of learning on the job, organizations continue to misuse sink or swim techniques, resulting in employees who feel unsupported and ill-prepared for their new roles. In line with previous research, this study has illustrated the important role that interpersonal support and a structured socialization program can have in driving positive adjustment outcomes (Ashforth & Saks, 1996; Bauer & Erdogan, 2010; Saks, et al., 2007). Additionally, this trend will continue as workplaces adjust to the new normal of a post-COVID-19 pandemic context. As digital training technologies have improved and more workers have transitioned to home offices, organizations are seeking cost-efficient virtual training options that promote individualized learn-on-your-own solutions. This research demonstrates that while these solutions can be effective, the pre-pandemic requirements of training, in supporting and effectively preparing new hires are still a necessity for promoting high-performing and satisfied workers. The SSS can be beneficial to both researchers and practitioners, as they look to evaluate the impact and effectiveness of post-pandemic socialization experiences.

Limitations

When reviewing the findings within this research, it is important to take into account some limitations. First, this study relied solely on self-report data. Organizational psychology researchers have raised concerns about self-report data based on the common method variance (CMV), stating its potential to raise the magnitude of the relationships that exist between predictors and work outcomes (Muntz & Dormann, 2020). To account for this issue, a longitudinal design at two time points was used; however, CMV and endogeneity concerns may remain considering the relatively short time between each wave of data.

A second limitation was the relative homogeneity of the sample. For one, the vast majority of the participants were Caucasian. Within the socialization literature, research has shown racial identity theory and ingroup identity play a significant role in a newcomer’s training experience and social network formation (Ashforth, et al., 2008; Mollica, et al., 2003); therefore these diverse experiences may not have been effectively represented. Additionally, the majority of the sample had more than 10 years of working experience. Previous experience is a significant moderator between socialization tactics and adjustment outcomes, with more experienced workers relying less on a structured training program (Kowtha, 2018). Although age did not demonstrate a correlation with any of the three factors, it is possible that the analyses may have been more representative of an increasingly mature and resilient workforce.

Finally, this research incorporated a sample of employees who had entered their organization within the last two years. Although this study intended to survey participants who had been hired in the previous year, an inability to reach a sufficient sample necessitated an adjustment. It is possible that a portion of the sample may not have experienced a socialization program for almost two years, meaning the participants in the sample may have significantly different memories of their experiences. Additionally, the unique challenges faced by workers during COVID-19 could have impacted the ecological validity of the study. Although the majority of the sample still specified that they had some form of in-person training, their socialization experiences may have been impacted by policies and protocols around social distancing.

Future Research

Future research should replicate the results of this study while using a longitudinal design with more time between measurements. For example, researchers could look to evaluate the impact of sink or swim training over time by measuring its effects directly after training and at additional time points during employment. Additionally, the OST scale has oft been used when connecting socialization tactics with adjustment outcomes. Future research should look to use the SSS to evaluate how this reconceptualization of a socialization experience interacts with psychological variables. More specifically, it would be valuable to better understand how the factors of responsibility, effectiveness, and support interact with driving positive or negative adjustment outcomes. In addition, although the 3-factor model was deemed the best fit in this study, it still demonstrated only an adequate fit overall. Future research should investigate the psychometric potential for a standalone measure of socialization support.

Furthermore, this research has regularly connected sink or swim socialization experiences with adverse outcomes. However, there is research connecting informal socialization experiences with positive outcomes like innovative and creative behaviours (Bauer & Erdogan, 2010). For example, organizations like Google have successfully implemented iterations of a sink or swim design to help develop an increasingly entrepreneurial and proactive workforce (Johnson & Senges, 2010). Future research should investigate the situational antecedents that may allow sink or swim socialization experiences to be successful in certain contexts. Additionally, the SSS can also be used to assess how employee’s socialization experiences are changing following the COVID-19 pandemic. The use of decentralized learn-on-your-own forms of training continues to rise as more employees work from home and organizations look for inexpensive means of providing training from a distance. The SSS is a tool that can be used by researchers to assess how employees are adjusting to these newfound forms of socialization. This can help organizations ensure the quality of their socialization programs even as they adjust the methodology.

This research also found that responsibility and support both had a small positive and negative significant relationship with proactivity and general self-efficacy respectively. While previous research has looked at the moderating impact of individual differences, future research should look more closely at the causal relationship between an individual’s general disposition and their perception of an organization’s socialization tactics (Ashford & Black, 1996; Jones, 1986). Additionally, previous research has found that work experience has a moderating impact on the relationship between socialization tactics and workplace outcomes (Kowtha, 2018). Future research should replicate these findings with sink or swim to better understand if more experienced workers are more resilient and resistant to the adverse effects of these socialization experiences.

Conclusion

In closing, this study was able to operationalize a model of perceived sink or swim within the workplace and develop a 17-item scale to measure it. This scale was designed to operate as a simple and practical alternative to the OST scale. Additionally, psychometric assessments of the scale supported its validity and reliability in measuring its intended construct. This included an assessment of predictive validity, which illustrated the relationship that the perception of sink or swim has with adverse workplace outcomes. The results of this study also demonstrated the strength of the support factor, as future research should seek to evaluate its capacity to operate as a standalone construct. Sink or swim is not a new concept. One would be hard-pressed to find an individual in most organizations who have not come in contact with it in some form over their career. However, researchers have yet to meaningfully evaluate the impact that it can have on workers and how it can best be used to facilitate employee adjustment. With the development of a practical measure of perceptions of sink or swim, this study has provided researchers and practitioners with a valid tool to better understand the socialization experiences of employees.

Appendix A - Final 17-Item Sink or Swim Scale (SSS)

Please indicate the extent of your agreement with each item based on your training experience in your current organization.

Support

  1. 1.

    My work team didn’t help me get up-to-speed in my role.

  2. 2.

    Nobody checked in on me during my first few months of employment.

  3. 3.

    I wasn’t formally introduced to members of the organization when I started.

  4. 4.

    My co-workers didn’t have time for me when I started my role.

  5. 5.

    My supervisors were not available to help when I had a problem during training.

  6. 6.

    I received limited support in my role when I started at this organization.

  7. 7.

    I didn’t know who to talk to when faced with a problem during training.

Effectiveness

  1. 1.

    I did not go through an organized training program when I started in my role.

  2. 2.

    There was no training roadmap to help prepare me for my position.

  3. 3.

    My organization had little structure in its training program.

  4. 4.

    My organization did not provide insight into how long each stage of the training process should take.

Responsibility

  1. 1.

    Sink or swim training is used as a way to screen out poor performers at this organization.

  2. 2.

    My organization relies on a sink or swim approach to teaching new hires how things work.

  3. 3.

    It is expected that I find my own way within my organization’s training.

  4. 4.

    This organization believes that it is the responsibility of a new hire to learn how to fit in.

  5. 5.

    My organization uses a learn-on-your-own approach to training.

  6. 6.

    Senior members of this organization want new hires to rely on themselves in their training.