Perceived supervisor support (PSS) is an important factor in employees’ health and well-being, and organizational outcomes. This study aimed to develop a Japanese version of the Survey of Perceived Supervisor Support (SPSS-J) and to evaluate its validity and reliability for assessing PSS in Japanese workers. The SPSS-J, including its subscales, SPSS-J (Sufficiency of Positive Aspects [SPA]) and SPSS-J (Minimality of Negative Aspects [MNA]), was developed based on the Japanese version of the Survey of Perceived Organizational Support (SPOS-J). An online survey was conducted with 1,946 participants, followed after two weeks by a test-retest survey with 589 participants. Structural validity was assessed using confirmatory factor analysis (CFA), and convergent validity was evaluated through Pearson’s correlations. Internal consistency and test-retest reliability were determined using Cronbach’s α coefficients and intraclass correlation coefficients (ICCs). CFA confirmed that the two-factor model provided a generally good fit, supporting structural validity. Both SPSS-J (SPA) and SPSS-J (MNA) demonstrated expected correlations with SPOS-J and its subscales, as well as possible antecedents (e.g., job autonomy and strengths-based performance appraisal) and outcomes (e.g., work engagement, organizational affective commitment, and turnover intention). The SPSS-J (SPA) consistently showed stronger correlations than the SPSS-J (MNA). Its internal consistency was high, with Cronbach’s α coefficients of .92 for SPSS-J (SPA) and .87 for SPSS-J (MNA). The test-retest reliability was also high, with ICCs of 0.77 and 0.74, respectively. Our study indicates that the SPSS-J is a reliable and valid tool for assessing PSS. The SPSS-J (SPA) may offer a practical alternative as a shortened version, reducing respondent burden.
Securing a workforce capable of maintaining long-term health and productivity is a critical challenge for organizations. Globally, developed nations are experiencing a reduction in their working-age population due to lower birth rates and an aging demographic. Furthermore, employees with health issues may experience reductions in health-related work productivity [1, 2]. Consequently, many organizations worldwide are investing in initiatives to support and promote employee health and well-being [3, 4]. Similarly, the government-led “Health and Productivity Management” initiative has been implemented in Japan, with an increasing number of organizations participating each year [5, 6].
Support from workplace supervisors is crucial for the success of health initiatives in organizations. One commonly used indicator for assessing such support is the Perceived Supervisor Support (PSS) scale. PSS, a concept developed based on Perceived Organizational Support (POS) [7], refers to the extent to which employees believe their supervisors value their contributions, provide assistance, and care about their well-being [8]. Other questionnaires, such as the globally used Job Content Questionnaire [9] and the New Brief Job Stress Questionnaire, which is commonly employed in the Japanese workforce [10], also assess supervisor support. However, these tools primarily assess support related to job tasks. In contrast, PSS emphasizes not only work-related support but also employees’ health and well-being, making it a particularly valuable indicator for organizational initiatives aimed at promoting employee health and well-being.
Many previous studies have shown that high PSS contributes to various outcomes of organizational health initiatives. Specifically, high PSS has contributed to positive psychological states, such as work engagement, job satisfaction, and organizational commitment. High PSS has also been related with lower levels of negative psychological states, including work-related stress, emotional exhaustion, and burnout [11–13]. High PSS has also been shown to mitigate work productivity losses and reduce employee turnover, which represent critical outcomes of organizational health initiatives [14–18]. These effects can be explained by the theoretical foundation of PSS in social exchange theory [8, 19]. Specifically, the more employees perceive support from their supervisors, the more their sense of reciprocity strengthens, motivating them to exert greater effort to meet their supervisors’ expectations.
Previous research has also identified several antecedent factors of PSS, including job autonomy and strengths-based performance appraisal [20, 21]. Studies have reported that PSS may improve when employees perceive higher levels of job autonomy and strengths-based performance appraisal. Enhancing PSS is therefore essential for promoting employee well-being and organizational success. For instance, promoting job autonomy and implementing strengths-based performance appraisal may be practical approaches to achieve this goal.
Perceived Supervisor Support is commonly evaluated by adapting the Survey of Perceived Organizational Support (SPOS). First introduced by Eisenberger et al in 1986, it was designed to assess employees’ perceptions of organizational support [7]. Over time, shortened versions of the SPOS (e.g., 16-item or 8-item) have been developed or used to suit research purposes [22]. Kottke et al introduced the Survey of Perceived Supervisor Support (SPSS) in 1988 by modifying the wording of SPOS items to replace “organization” with “supervisor,” resulting in a 16-item scale specifically designed to measure PSS [8]. However, the use of the original SPSS scale has been limited in subsequent research. Instead, many studies have assessed PSS by selecting from the SPOS items and substituting “organization” with “supervisor” to suit the specific research purpose [11–17]. In studies that simultaneously measure POS and PSS, researchers often use the same set of items for both constructs [14, 23]. Although PSS has been adapted to fit research objectives, previous studies have primarily measured the PSS by only evaluating its internal consistency within study samples, and, to our knowledge, few studies have conducted a systematic examination of its validity.
The Japanese version of the SPOS (SPOS-J) was developed by Odagami et al and its validity and reliability were evaluated within a Japanese context [24]. Notably, the SPOS-J differs from the original SPOS in its factor structure: while the original version demonstrated a single-factor structure [7], the SPOS-J was shown to have a two-factor structure. Consequently, a new 8-item shortened Japanese version was created, which differs from the widely used 8-item version of the original SPOS. To develop a Japanese version of the SPSS (SPSS-J), it is necessary to adapt the SPOS-J by replacing the term “organization” with “supervisor,” following the procedures outlined in previous studies [8, 11–17]. Additionally, before applying this scale, its validity and reliability need to be systematically evaluated.
This study aimed to develop the SPSS-J based on the SPOS-J and to evaluate its validity and reliability for measuring PSS.
Participants
The survey was conducted online from January 23 to January 26, 2023, targeting registered monitors of Cross Marketing Inc. (Tokyo, Japan), which has 5.0 million registrants. Invitations to participate were emailed to 41,235 individuals, with 5,882 having completed the screening questions to assess their eligibility. Due to budget limitations, the target sample size was set at 2,000 responses, and participants were selected on a first-come, first-served basis. Recruitment was closed once the final count of 2,046 participants was achieved. In the present study, we excluded participants whose employment status might have affected the PSS evaluation, such as those who were self-employed, family employees, or working side jobs at home. After applying these exclusion criteria, 1,907 participants remained in the baseline survey.
A follow-up survey was conducted two weeks later to assess test-retest reliability. The follow-up targeted participants from the 2,046 respondents of the baseline survey, with a goal of recruiting 600 participants. Recruitment was closed after reaching 618 respondents. The same exclusion criteria were applied as in the baseline survey, resulting in a final sample of 581 participants in the follow-up analysis.
The study was approved by the Ethics Committee of Medical Research, University of Occupational and Environmental Health, Japan (R4-065). All participants gave their informed consent via an online form accessible on the survey website.
Measures
SPSS-J scale
As in previous studies evaluating PSS, we developed the SPSS-J (8 items) by replacing the subject of the SPOS-J (8 items), which has demonstrated validity and reliability, from “organization” to “supervisor” [8, 11–17, 24]. We also developed two subscales for the SPSS-J: SPSS-J (Sufficiency of Positive Aspects [SPA]) and SPSS-J (Minimality of Negative Aspects [MNA]), each consisting of four items based on the subscales of the SPOS-J, SPOS-J (SPA) and SPOS-J (MNA) [24]. The specific items for both the original and Japanese versions are listed in Supplemental Table 1 ( https://jstagedata.jst.go.jp/juoeh).
Responses to each item were scored on a 7-point Likert scale ranging from 0 = “strongly disagree” to 6= “strongly agree,” and the mean scores of SPSS-J (8 items) and its subscales, SPSS-J (SPA) and SPSS-J (MNA), were used as the indicators in this study.
Relevant Variables
In this study, we selected the following antecedents and outcomes of PSS based on previous research, in addition to SPOS-J (8 items and its subscales). As antecedents, we included job autonomy and strengths-based performance appraisal. As outcomes, we included work engagement and organizational affective commitment as positive indicators, and turnover intention as a negative indicator.
Possible Antecedents
Job autonomy was assessed using Japanese translations of four items from previous research [20]. The following items were included: “I have the freedom to decide what I do on my job,” “It is basically my own responsibility to decide how my job gets done,” “I have a lot of say about what happens on my job,” and “I am given a lot of freedom to decide how I do my own work.” Responses were scored on a 4-point Likert scale ranging from 1= “strongly disagree” to 4= “strongly agree,” and the mean score was used as the indicator. The Cronbach’s α coefficient was 0.86. Strengths-based performance appraisal was measured using a Japanese translation of four items from previous research [21]. The items included: “Appreciation was expressed for my unique qualities,” “Attention was paid to discovering my unique qualities in relation to my work,” “Attention was paid to how I can further my talents,” and “Attention was paid to how I can make better use of my strengths in my work.” Responses were scored on a 5-point Likert scale ranging from 1= “not true at all” to 5= “very true,” and the mean score was used as the indicator. The Cronbach’s α coefficient was 0.94.
Possible Outcomes
Work engagement was measured using the 9-item Japanese version of the Utrecht Work Engagement Scale, which has been validated for its consistency, reliability, and both factor and construct validity [25]. Each item was scored on a 7-point Likert scale ranging from 0 = “never” to 6 = “always.” The mean score was used as the indicator in this study. The Cronbach’s α coefficient was 0.96. Organizational affective commitment was measured using three items from The Japan Institute for Labour Policy and Training Human Resource Management checklist regarding affective commitment [26]. The items included: “I feel as if this company’s problems are my own problems,” “I am proud to be a member of this company,” and “I strongly identify with being a member of this company.” Responses were scored on a 5-point Likert scale ranging from 1= “not true at all” to 5= “very true,” and the mean score was used as the indicator in this study. The Cronbach’s α coefficient was 0.84. Turnover intention was measured using one question from the Japanese version of the Turnover Intention Scale [27]: “How often have you seriously considered quitting your current job?” Responses were scored on a 6-point Likert scale ranging from 1=“not at all” to 6=“very often.”
Demographic characteristics
Participant characteristics were collected in the first survey. Respondents were asked to provide information on gender (male or female), age, annual household income (<3.00 million; 3.00–4.99 million; 5.00–9.99 million; and ≥10.00 million Japanese yen), and final education (junior high or high school; vocational school or college; and university or graduate school). Age was categorized into six groups: 20–29 years; 30–39 years; 40–49 years; 50–59 years; 60 years and older. Job type was classified according to the Japan Standard Industrial Classification (Rev. 4th, December 1997) into the following categories: professional and technical; administrative; clerical; sales; service; security; agriculture, forestry, and fishing; transportation and communication; production process and labor work; and others.
Statistical analysis
Structural validity was assessed using confirmatory factor analysis (CFA) to evaluate whether the two-factor model identified in the SPOS-J study was also valid for SPSS-J. To determine model fit, we used the χ ² goodness-of-fit statistic, root mean square error of approximation (RMSEA), standardized root mean square residual (SRMR), and comparative fit index (CFI). Acceptable model fit was judged by the following criteria: RMSEA < 0.06, SRMR < 0.08, and CFI > 0.95 [28].
Convergent validity was assessed by first examining the correlations between the SPSS-J (including its subscales) and the SPOS-J, on which the SPSS-J was based. Next, Pearson’s correlation coefficients were calculated to explore the relationships between the SPSS-J (8 items and its subscales) and other relevant variables, including possible antecedents (job autonomy and strengths-based performance appraisal) and outcomes (work engagement, organizational affective commitment, and turnover intention). The cutoff points for correlation coefficients vary across studies, but we defined them as follows: coefficients < 0.10 indicate negligible, 0.10–0.39 indicate mild, 0.40–0.69 indicate moderate, 0.70–0.89 indicate strong, and ≥ 0.90 indicate very strong correlation [29].
Reliability was confirmed through internal consistency and test-retest reliability. Internal consistency reliability was evaluated using Cronbach’s α coefficients for the SPSS-J (8 items) and its subscales. To evaluate test-retest reliability, we decided not to conduct formal normality tests in this study, as previous experience with SPOS-J indicated that normality tests in large samples are often overly sensitive to even slight deviations from normality, which can lead to significant findings [30]. Nevertheless, to confirm the normality of the SPOS-J, we visually inspected histograms and normal quantile-quantile plots, which suggested that the scores closely approximated a normal distribution. Based on this visual assessment, we assessed test-retest reliability using Pearson correlations, the intraclass correlation coefficient (ICC), and the standard error of measurement (SEM), treating the SPSS-J (8 items) as a parametric variable. Statistical significance was determined as P < 0.05 (two-tailed). All statistical analyses were performed using IBM SPSS Statistics, Version 27.0 and Stata Statistical Software, Release 16 (NY, USA).
Participant characteristics
Table 1. presents the characteristics of the participants in the baseline and follow-up surveys. In the baseline survey, 54.3% were male, and 52.9% of participants held a college degree or higher, with the most prevalent job types being clerical, and professional and technical. The characteristics of participants in the follow-up survey did not show a significant difference from those in the baseline survey.
| Baseline | Follow-up* | |
|---|---|---|
| Total | 1,907 | 581 |
| Gender, male | 1,036 (54.3%) | 311 (53.5%) |
| Age, years | ||
| 20–29 | 297 (15.6%) | 69 (11.9%) |
| 30–39 | 331 (17.4%) | 90 (15.5%) |
| 40–49 | 456 (23.9%) | 152 (26.2%) |
| 50–59 | 420 (22.0%) | 132 (22.7%) |
| 60– | 403 (21.1%) | 138 (23.8%) |
| Annual household income (JPY) | ||
| <3.00 million | 275 (15.8%) | 90 (17.0%) |
| 3.00–4.99 million | 506 (29.0%) | 148 (28.0%) |
| 5.00–9.99 million | 826 (47.4%) | 253 (47.9%) |
| ≥10.00 million | 135 (7.7%) | 37 (7.0%) |
| Final Education | ||
| Junior high or high school | 473 (24.8%) | 140 (24.1%) |
| Vocational school or college | 426 (22.3%) | 136 (23.4%) |
| University or graduate school | 1,008 (52.9%) | 305 (52.5%) |
| Job type | ||
| Professional and technical | 441 (23.1%) | 138 (23.8%) |
| Administrative | 161 (8.4%) | 49 (8.4%) |
| Clerical | 510 (26.7%) | 158 (27.2%) |
| Sales | 180 (9.4%) | 57 (9.8%) |
| Service | 232 (12.2%) | 58 (10.0%) |
| Security | 28 (1.5%) | 11 (1.9%) |
| Agriculture, forestry, and fishing | 6 (0.3%) | 2 (0.3%) |
| Transportation and machine operation | 58 (3.0%) | 22 (2.8%) |
| Production process jobs | 163 (8.5%) | 53 (9.1%) |
| Others | 128 (6.7%) | 33 (5.7%) |
JPY: Japanese Yen
*Follow-up sample was used only to examine the test-retest reliability.
Structural validity
Confirmatory factor analysis showed revealed the following fit indices for the two-factor model: χ ² = 182, P < 0.001; RMSEA = 0.067 [90% CI: 0.058–0.076]; SRMR = 0.031; and CFI = 0.984. While the RMSEA was slightly above the criteria, the other indices met the criteria. These results indicate that the two-factor model provided a generally good fit to the SPSS-J data.
Convergent validity
Table 2. presents correlations of the SPSS-J and its subscales, SPSS-J (SPA) and SPSS-J (MNA), with the SPOS-J, as well as possible antecedents (job autonomy and strengths-based performance appraisal) and outcomes (work engagement, organizational affective commitment, and turnover intention). The SPSS-J (8items) showed a strong positive correlation with the SPOS-J (8items) (r = 0.81). The SPSS-J (SPA) exhibited a strong correlation with the SPOS-J (SPA) (r = 0.76), which was higher than its correlation with the SPOS-J (MNA) (r = 0.41). Similarly, the SPSS-J (MNA) had a moderate correlation with the SPOS-J (MNA) (r = 0.68), which was higher than its correlation with the SPOS-J (SPA) (r = 0.37).
Both the SPSS-J (SPA) and the SPSS-J (MNA) demonstrated moderate to mild correlations with possible antecedents and outcomes. Notably, turnover intention had a negative correlation, whereas all other variables exhibited positive correlations. Between the two subscales, the SPSS-J (SPA) generally demonstrated stronger correlations with antecedents and outcomes compared to the SPSS-J (MNA).
| Mean | SD | 1 | 1–1 | 1–2 | 2 | 2–1 | 2–2 | 3 | 4 | 5 | 6 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.SPSS-J (8 items) | 3.07 | 1.09 | ||||||||||
|
1–1. SPSS-J (SPA) |
2.88 | 1.31 | .85* | |||||||||
|
1–2. SPSS-J (MNA) |
3.25 | 1.25 | .84* | .44* | ||||||||
| 2. SPOS-J 8items | 2.94 | 1.13 | .81* | .70* | .67* | |||||||
|
2–1. SPOS-J (SPA) |
2.64 | 1.40 | .70* | .76* | .41* | .81* | ||||||
|
2–2. SPOS-J (MNA) |
3.24 | 1.38 | .61* | .37* | .68* | .81* | .31* | |||||
| Possible antecedents | ||||||||||||
| 3. Job autonomy | 2.51 | 0.76 | .34* | .39* | .18* | .32* | .39* | .12* | ||||
| 4. Strengths-based performance appraisal | 2.65 | 1.01 | .60* | .64* | .37* | .62* | .66* | .34* | .46* | |||
| Possible outcomes | ||||||||||||
| 5. Work engagement | 3.47 | 1.38 | .39* | .47* | .20* | .43* | .54* | .15* | .35* | .47* | ||
| 6. Organizational affective commitment | 2.81 | 0.95 | .55* | .61* | .32* | .57* | .62* | .30* | .42* | .72* | .56* | |
| 7. Turnover intention | 2.98 | 1.35 | -.42* | -.41* | -.30* | -.42* | -.41* | -.27* | -.25* | -.34* | -.38* | -.39* |
Note; * P<.001. SPSS-J: Japanese version of Survey of Perceived Organizational Support,
SPOS-J: Japanese version of Survey of Perceived Organizational Support, SPA: Sufficiency of positive aspects,
MNA: Minimality of negative aspects
Reliability (Internal and test-retest)
Table 3. shows Cronbach’s α coefficient, Pearson’s correlation coefficients, ICCs, and SEM for the SPSS-J and its subscales, SPSS-J (SPA) and SPSS-J (MNA). Cronbach’s α coefficients of the SPSS-J (SPA) and SPSS-J (MNA) indicated high internal consistency (0.92 and 0.87, respectively). For test-retest reliability, Pearson’s correlation coefficients and ICCs indicated moderate to strong reliability for both the SPSS-J (SPA) and the SPSS-J (MNA).
| Mean | SD | Min | Max | Cronbach’s α | r | ICC (95% CI) | SEM (95% CI) | |
|---|---|---|---|---|---|---|---|---|
|
SPSS-J (8 items) |
3.07 | 1.09 | 0 | 6 | 0.88 | 0.83 ( P<.001) | 0.83 (0.80-0.85) | 0.56 (0.52-0.60) |
| SPSS-J (SPA) | 2.88 | 1.31 | 0 | 6 | 0.92 | 0.77 ( P<.001) | 0.77 (0.74-0.80) | 0.75 (0.69-0.80) |
| SPSS-J (MNA) | 3.25 | 1.25 | 0 | 6 | 0.87 | 0.74 ( P<.001) | 0.74 (0.70-0.78) | 0.77 (0.71-0.82) |
Note: Mean, SD, Min, Max, and Cronbach’s α were calculated using all responses in the baseline survey.
SPSS-J: Japanese version of Survey of Perceived Organizational Support,
r: Pearson’s correlation coefficient for test-retest reliability,
ICC: intraclass correlation coefficients, SEM: standard error of measurement,
CI: Confidence interval, SPA: Sufficiency of positive aspects, MNA: Minimality of negative aspects
We developed the SPSS-J (8 items) and its subscales, SPSS-J (SPA) and SPSS-J (MNA), based on the SPOS-J, whose validity and reliability have been previously established. To evaluate the SPSS-J, we assessed its construct validity, convergent validity, internal consistency, and test-retest reliability for the total SPSS-J (8 items) and its subscales in Japanese workers.
Structural validity
Our results indicate that the two-factor structure identified in the SPOS-J study is also valid for the SPSS-J. The good fit indices observed in the CFA demonstrate satisfactory structural validity, which supports the appropriateness of adopting this two-factor model for the SPSS-J.
Convergent validity
A strong positive correlation was observed between the SPSS-J (8 items) and the SPOS-J (8 items) (r = 0.81), supporting the validity of the SPSS-J. PSS and POS are based on the same foundational theory, and PSS is considered a potential antecedent of POS. Consequently, a positive correlation between the two has been reported in previous studies [14, 31–33]. For instance, a study by Eisenberger et al using the same eight items of SPOS and SPSS reported a moderate correlation coefficient (r = 0.60) between POS and PSS [14]. Similarly, a meta-analysis by Boyer et al also reported a moderate correlation coefficient (r = 0.60) [31]. One possible reason for the stronger correlation coefficient observed in our study may reflect contextual differences, such as variations in the specific questions used in the Japanese version of the scale. Cultural differences may also have played a role. PSS and POS are said to correlate because workplace supervisors are often seen as representatives of the organization [14]. Vertical-collectivist societies like Japan, compared to horizontal-individualist societies common in Western countries, emphasize interdependence and alignment with group interests. As a result, employees in such cultures may identify more strongly with their leader’s goals and view their supervisor’s support as a reflection of the organization’s support [34]. This cultural tendency may explain the stronger correlation between SPSS-J and SPOS-J observed in this study. These findings provide additional insights into the relationship between SPSS-J and SPOS-J and highlight the importance of considering cultural and contextual factors in future research. Additionally, the SPSS-J subscales demonstrated validity. Specifically, the SPSS-J (SPA) showed a stronger correlation with the SPOS-J (SPA) than with the SPOS-J (MNA), and the SPSS-J (MNA) similarly exhibited a stronger correlation with the SPOS-J (MNA) than with the SPOS-J (SPA). These findings indicate that the SPSS-J subscales accurately measure their intended constructs, further supporting their validity.
We examined the correlation between the SPSS-J and possible antecedents (job autonomy and strengths-based performance appraisal) as well as possible outcomes (work engagement, organizational affective commitment, and turnover intention), and found results that align well with previous studies. Specifically, job autonomy showed mild correlations with the two SPSS-J subscales (r = 0.39, r = 0.18) in this study. Previous research has reported a mild positive correlation between job autonomy and PSS [20], and our findings are consistent with this. Strengths-based performance appraisal has been shown to have a moderate positive correlation with PSS [21]. In our study, we observed moderate positive correlations (r = 0.64) with the SPSS-J (SPA) and mild positive correlations (r = 0.37) with the SPSS-J (MNA), confirming the results of earlier research.
Many previous studies have reported a moderate positive correlation between PSS and work engagement [12, 13]. In this study, the two subscales showed moderate and mild positive correlations (r = 0.47, r = 0.20), respectively, which are consistent with previous studies. Furthermore, previous studies have demonstrated a positive correlation between PSS and organizational affective [23, 31]. A meta-analysis by Boyer et al reported a moderate correlation (r = 0.45) [31]. In our study, moderate and mild positive correlations were obtained (r = 0.61, r = 0.32) for the two subscales. Regarding turnover intention, previous studies have reported a negative correlation with PSS, with direct or indirect effects through POS, job performance, and self-efficacy [14–17]. In our study, moderate and mild negative correlations were found (r = –0.41, r = –0.30) for the two subscales. These results indicate that the SPSS-J and its subscales show findings consistent with previous studies, supporting convergent validity.
In this study, the SPSS-J (SPA) showed higher correlation coefficients with all the relevant factors compared to the SPSS-J (MNA). This result is consistent with the SPOS-J study [24]. To control for acquiescence bias, which is a tendency for respondents to agree with items regardless of their content, it is common practice in affective scales to include an equal number of negatively and positively worded items; however, it has been reported that negatively worded items tend to be less clearly understood, leading to inconsistent or inaccurate responses and reduced validity [35]. Indeed, when the original SPSS was developed by Kottke et al inconsistencies in responses to some negatively worded items were observed, prompting their revision into positively worded items [8]. In contrast, the SPSS-J (SPA), which uses positively worded items, likely benefits from clearer and more consistent interpretation, producing more stable and reliable responses. This improved reliability may explain the stronger correlation coefficients observed with the SPSS-J (SPA).
Reliability
Regarding reliability, the SPSS-J and its subscales demonstrated high internal consistency, with Cronbach’s α coefficients ranging from 0.87 to 0.92. Although direct comparisons are not feasible due to differences in item composition, the reliability of the SPSS-J appears comparable to that of the 8-item version of the SPSS reported in a study by Eisenberger et al [14]. To the best of our knowledge, no studies have previously examined the test-retest reliability of the SPSS. In this study, both the SPSS-J (SPA) and the SPSS-J (MNA) showed moderate to strong reliability based on Pearson’s correlation coefficients and ICCs. These findings suggest that the SPSS-J has high reliability.
Practical implications
When organizations implement measures to promote employee health and well-being, supervisors play a critical role. Employees who perceive strong support from their supervisors are more likely to engage in these measures, which can positively influence productivity and reduce turnover intentions [14–18, 36]. To effectively implement health initiatives, and by extension, the Health and Productivity Management practices widely adopted by organizations in Japan, it is crucial to use the SPSS-J to evaluate and enhance PSS over time.
Previous research has demonstrated that improving employees’ POS can enhance organizational performance, which may lead many organizations to prioritize POS as a strategic goal [32, 33]. However, since PSS is considered one of the key antecedents of POS [14], organizations should focus on enhancing employees’ perceptions of support from their supervisors rather than directly targeting POS. Supervisors interact more directly and frequently with employees, making PSS a practical and effective avenue for improvement. Importantly, while PSS and POS are related, they are not interchangeable: POS reflects organizational-level support, whereas PSS depends heavily on individual supervisors’ behaviors. For example, even when organizational support is strong, insufficient supervisor support can undermine employee participation in health initiatives, work engagement, and productivity [18, 36, 37]. This distinction underscores the importance of focusing on PSS, particularly in medium- and large-sized organizations where employee POS tends to be lower than in smaller organizations. Research has shown that enhancing PSS in such settings plays a particularly meaningful role in improving POS [38]. Therefore, evaluating and strengthening PSS is crucial for fostering POS and achieving better organizational outcomes as part of a comprehensive support strategy.
Regarding the SPSS-J, our study confirmed that the SPSS-J (SPA) possesses sufficient validity and reliability. While it is generally recommended to include an equal number of positively and negatively worded items, it may be possible to reduce the number of items to lessen respondent burden. Similar to the SPOS-J [28], using only the four items of the SPSS-J (SPA) could be a feasible approach in such cases.
Limitations and Future directions
This study has several limitations. First, the use of self-reported measures may have introduced common method bias, potentially affecting the results. The true relationships between variables might be weaker than those observed in this study. Future research should replicate the findings using objective measures, such as turnover based on personnel data and work engagement assessed through peer evaluations. Second, the generalizability of the findings is limited by the characteristics of the sample. Data were collected from a single internet company for convenience, which may not represent the entire Japanese workforce. For example, this sample included a higher proportion of highly educated individuals compared to the overall Japanese workforce [39], which might have influenced the results. Third, as our study used a cross-sectional design, it did not investigate causal relationships. Future research should explore the causal links between PSS and its possible antecedents and outcomes. Finally, this study did not assess criterion validity or responsiveness, as these aspects were beyond the scope of the current research. Future studies should address these aspects to enhance the robustness of the findings.
The SPSS-J was developed based on the validated SPOS-J, and its validity and reliability for measuring PSS were examined. The findings indicate that the SPSS-J is a reliable and valid tool for research and surveys, particularly targeting Japanese workers. Among its subscales, the SPSS-J (SPA) may offer a practical option for reducing respondent burden while still maintaining validity and reliability.
This study was financially supported by the University of Occupational and Environmental Health, Japan, and the HASEKO Corporation.
The authors declare no conflicts of interest associated with this manuscript.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Conceptualization: TM, KO, TN, AI, and KM
Formal analysis: KO, AI
Funding acquisition: KM
Investigation: KO, TN, KM
Supervision: KM
Writing-original draft: TM
Writing-review & editing: TM, KO, TN, AI, and KM
記事関連データはJ-STAGE Dataで利用できます。（産業医科大学雑誌 J-STAGE Data） / The data analysis file and all annotator data files are available in J-STAGE Data. (J-STAGE Data from Journal of UOEH)
