Perceived systems intelligence and performance in organizations

Purpose – This study aims to introduce the perceived systems intelligence (SI) inventory, developed based on the earlier published self-report SI inventory (Törmänen et al. , 2016). It can be used together with earlier managerial level tools for building a learning organization and included in general 360-style evaluations in personneldevelopment. Design/methodology/approach – The inventory is validated with con ﬁ rmatory factor analysis with a model based on the self-report SI inventory, using data from full-time used employees and managers in the USA and UK.Perceived SIfactor scoresarecorrelated with theperceived study performance ofthe individual. Findings – The perceived SI inventory is found to have good factorial validity, and it correlates strongly with evaluations of perceived study performance. Managers perceived high in performance are also found to score high in perceived SI.Perceived SIdoes notdependongender, age, organization size orindustry. Originality/value – The perceived SI inventory is the ﬁ rst personnel level peer evaluation tool suggested for developing learning organizations. The new inventory makes peer evaluations possible and provides a newgrassrootslevel tool forpersonneldevelopment programs inlearning organizations. of her environment. perceives herself as a part of a the in ﬂ uence of the whole upon herself as well as her own in ﬂ uence upon the observing own interdependence


Introduction
Systems intelligence (SI) is a concept describing our abilities to succeed in complex situations in organizational settings and in our everyday life. Saarinen and Hämäläinen (2004, p. 4) originally defined SI as: [. . .] intelligent behavior in the context of complex systems involving interaction and feedback. A subject acting with Systems Intelligence engages successfully and productively with the holistic feedback mechanisms of her environment. She perceives herself as a part of a whole, the influence of the whole upon herself as well as her own influence upon the whole. By observing her own interdependence in the feedback intensive environment, she is able to act intelligently. SI has been suggested as an essential competence needed in positions of leadership Saarinen, 2006, 2007a) and has been applied for various fields such as organizational development (Luoma et al., 2008(Luoma et al., , 2011, knowledge management (Sasaki, 2017), personal growth (Hämäläinen et al., 2014;Saarinen and Lehti, 2014), therapy discourse (Martela and Saarinen, 2013), design thinking (Harviainen et al., 2021;Jumisko-Pyykkö et al., 2021) and engineering disciplines (Hämäläinen et al., 2018). For a history of the SI perspective, see Törmänen (2021).
SI draws from Peter Senge's seminal management book The Fifth Discipline (Senge, 1990), which introduced what Senge calls the five disciplines of the learning organizationsystems thinking, personal mastery, mental models, building shared vision and team learning. SI seeks to operationalize Senge's vision of the learning organization and integrate it conceptually. To this effect, the original SI inventory of 2016 offered a self-evaluation questionnaire (Törmänen et al., 2016). The more recent organizational SI inventory focused on SI in organizations . The current paper takes the SI perspective yet one more step further, extending it to the peer level.
While Senge's work has been hailed as "almost synonymous with the idea of learning organization" (Örtenblad, 2018), quantitative models of the learning organization have often bypassed key aspects of Senge's vision, especially regarding insights that are difficult to operationalize or control with managerial structures. In her recent overview, Bui highlighted Senge's constructs of personal mastery and systems thinking as ideas that "have powerfully shaped new ways to see the world and act upon" (Bui, 2020). However, in research literature seeking to operationalize the learning organization the two "disciplines" that Bui specifically singles out play in fact only cameo roles. For example, in the arguably most popular measurement scale for the learning organization, Dimensions of the Learning Organization Questionnaire (Marsick and Watkins, 2003), only two out of seven dimensions are focused on the individual. Traditionally, learning organization tools have been top-down with the agency at the leadership of the organization. Recently, however, interest has been rising for also developing bottom-up and multilevel tools and frameworks (Chou and Ramser, 2019). Rigby and Ryan (2018), in their visionary overview of human resource development (HRD), go as far as to allude to a "Copernican Turn" as an emerging organizational paradigm. In contrast to "'Pre-Copernican' approaches that rely on institutional levers and 'command and control' systems that management can activate to drive the desired behavior", the "Copernican Turn" looks at individuals and their motivational and emotional factors. "For HRD to succeed, tools are needed that tap into the worker's internal frame of reference".
In the HRD literature the need that Rigby and Ryan dramatize has indeed been noted for yearsthe need to link with "key individual variables" such as intention, goals, commitment and satisfaction (Peterson, 2004), "perceived investment in employee's development" (Lee and Bruvold, 2003) and "learning opportunities that nurture human expertise in organizations" (Shuck et al., 2014, p. 239). This emphasis of quintessentially human factors is particularly pronounced when scholars meet head-on the fact, painfully obvious to practitioners, that an HRD program might be excellent as a plan and yet fail to deliver desired outcomesbecause employees do not perceive the benefits. Key phenomena such as employee turnover intention or employee engagement hinge on how employees experience their peers, managers and organization, but managers and researcher alike have had a difficult time in conceptualizing and operationalizing the developmentally relevant TLO "human, all too human" parameters. Still, on the abstract level, the situation remains clear enough. As Chalofsky put it in his introduction to the authoritative Handbook of Human Resource Development, what is needed in the field is "a holistic approach to human development," one that draws from "inner growth that is realized through interaction of self, context, and life experiences" (Chalofsky, 2014, p. xlv).
The call for more humanly-tuned grassroots-level informed conceptualizations is burning in the context of learning organizations, where the scholarly discourse tends to favor managerial structures that operates from top to bottom. Here the SI approach takes a sharp departure in favor of employees, individuals and human experience. The idea is to approach development through a discourse that does not speak about structures that only managers command but in terms of a discourse employees can understand.
For Senge, systems thinking never was not the cold and objectivistic model building of complex systems from without. As Senge makes clear in his 1993 key paper in Human Resource Development Quarterly, "systems thinking" amounts to "seeing relationships" that calls for such deeply human from-within qualities as "genuine caring" and "compassion" (Senge, 1993). It is here where the SI perspective is particularly mindful of Senge's thinking for the benefit of organizational learning as a process. The SI approach takes caution neither to reify employees to objects nor systems to external entities. The SI perspective draws insight from modern relational, systems inspired empirical infant research and from its way of understanding the mother-infant dyad as the paradigmatic context of development (Hämäläinen and Saarinen, 2007b;Saarinen and Hämäläinen, 2010). In the mother-infant context, the baby is an active partner in the dyadic, bidirectional and co-creational system in which intra-and inter-subjective processes of development are intertwined (Beebe and Lachmann, 2005;Seligman, 2017).
The guiding intuition of the SI perspective concerns the factors that make humans succeed with and within wholes. With factors such as "systemic perception", "attunement", "positive engagement" and "effective responsiveness", the SI perspective wishes to take the context seriously as a key determinant of successful action. The environment is brought to bear on the subject bidirectionally. In organizations, a key contextual factor is created by one's fellow employees.
There is a long tradition in organizations to develop ways for managers and employees to get feedback of their actions, to avoid self-deception and to get a more realistic view of one's behaviors for the benefit of right-directed development. Indeed, the use of multisource feedback, often called 360-degree feedback, in its various forms, is an established organizational practice (Maylett, 2009;Church et al., 2019). The fact that leaders' perceptions of their behaviors, along with the employees' perceptions generally, might differ from those of others is a source of lively discussion. How the "rater bias" (Holzbach, 1978) and the "self-other agreement" affects various organizational outcomes is an issue that has relevance both theoretically and in practice (Atwater and Yammarino, 1997;Fleenor et al., 2010;Halverson et al., 2005;Lee and Carpenter, 2018). When it comes to leadership, the issue becomes particularly relevant with results such as Jacobsen and Bøgh Andersen (2015) that indicate that at least in some cases, employee-perceived leadership appears to be a more useful metric than the leader's own self-evaluation.
In this paper, we are interested if SI can also be evaluated by peers as perceived competence, especially vis-à-vis perceived job performance. We extend the notion of SI to perceptions of one's colleagues in one's organization and introduce a perceived SI measurement scale, adapting the factor structure introduced by the SI inventory of Törmänen et al. (2016). We study the relationship between the learning organization and organizational performance (Kim and Lu, 2019) by correlating perceived SI with perceived job performance, with a particular focus on people in a leadership position.

Perceived systems intelligence
The new inventory is tested and validated by administering it through a web-based questionnaire to people working in the UK and USA. The subjects (N = 569) included employees and managers in various organizations. The subjects were asked to evaluate two of their colleaguesone they felt was a "top performer" in their organization and one they felt was an "average performer." The subjects were asked to evaluate the person in question regarding behaviors and features that related to SI, as well as how the person succeeded in her work performance. The large sample of 1,138 peer-evaluated colleagues allowed analyzing the results obtained in different professions, organizational positions and age groups, as well as for men and women.

Method
Creating the perceived systems intelligence inventory We chose to develop the perceived SI inventory by adapting the self-report SI inventory developed in Törmänen et al. (2016) to a peer-evaluation context. Recognizing that some of the items might be more difficult to address as perceived items instead of self-report items, using the exact same factor structure allows for comparison and contrasting of the selfreport and peer perceived SI measurements, and thus significantly widens the applicability of the inventory. Therefore, it was considered desirable to retain as much of the already identified eight-factor SI structure as possible.
The perceived SI inventory was created by revising the phrasings of the original inventory. The original "I [. . .]" format ("I contribute to the shared atmosphere in group situations") was changed to a "My colleague [. . .]" format ("My colleague contributes to the shared atmosphere in group situations"). The resulting inventory of 32 items was screened in a pilot study, whose results were used to ensure that the new perceived versions were understood properly and that their answer distributions were not heavily skewed. Pilot results indicated that direct modification produced a well-functioning inventory with wellbehaving items. The resulting inventory, as used in this study, is included in Appendix.

Data and samples
To gather a large, well-sampled set of colleague evaluations, individuals in the academic crowdsourcing platform Prolific.ac were invited to evaluate their own colleagues. The participants received monetary compensation of £1.00-£1.25 for participating in the study. Each participant was asked to evaluate two of his or her colleagues, one who they considered to be among the top 5% of performers in their organization and another who they considered a typical, average performer.
The subject filled the 32-item perceived SI inventory for both colleagues and gave an estimate of how well they felt the colleagues performed in his/her work. The SI questions were answered on a seven-point Likert scale from "never" to "always." Performance was asked with the question "On a scale from 0 to 10, how well do you feel your colleague performs in his/her work?", with the subject answering on a Likert scale. The data set was gathered during a number of smaller-scale questionnaire rounds. In later rounds, the following two additional performance questions were posed with the same scale: "On a scale of 0-10, how much does this colleague help other people succeed?" and "On a scale of 0-10, how much does this colleague contribute to a positive work climate?" The subjects also described how well they knew the colleague ("We are close friends"; "Quite well"; "Only a little") and how long they had been colleagues ("Less than 1 year"; "1-3 years"; "3-10 years"; "Over 10 years"). The questionnaire also asked for the colleague's gender, age and position in their organization, and the participant's gender and age.

TLO
The subjects were selected from a pool of participants who worked full-time, were residents of the UK or the USA and at least 25 years of age. In total, 569 people participated, resulting in 1,138 perceived SI evaluations. The summary statistics from the data set have been collected to Table 1. The sampling strategy resulted in a roughly even split of men and women for both participants and colleagues, with a wide distribution of age groups, and with 44% of the colleagues being in managerial or supervisor positions.

Construct validity of perceived systems intelligence
The self-report SI inventory uses factor scores of the eight SI factors to highlight systems intelligent behavior and to give suggestions and recommendations to the individual. To study whether the same eight-factor model can be used with perceived SI data, we replicate the confirmatory factor analysis process described by Törmänen et al. (2016) using our full data sample (N = 1,138).
Construct validity is assessed by fitting a confirmatory factor analysis model using structural equation modeling (Bollen, 1989) to the data set, with the same 32-item 8-factor structure as the self-report SI inventory. Using the implementation of the R "sem" package (Fox et al., 2020), the resulting model has a chi-squared value of 1,831.4 with 436 degrees of freedom (p < 0.001).
The eight-factor model has a good model fit as indicated by recommendations by Hu and Bentler (1999), with structural fit indices root mean square error of approximation 0.053 and standardized root mean squared residual 0.057. Additionally, the model outperforms a simple single factor model with all items loading to a single SI factor (x 2 = 2,241.3, df = 464, x 2 difference p < 0.001), showing that a multifactor structure is clearly preferable for describing the perceived SI data set.
These results indicate good construct validity for the eight-factor perceived SI inventory and suggest that it can be used with the same eight factors as the self-report SI inventory.

Perceived systems intelligence and perceived work performance
We calculate perceived SI factor scores as weighted averages of the structural equation model coefficients produced in the previous step. Table 2 shows cross-correlations between  Table 2.
Correlations between SI factor scores and peer evaluated work performance TLO the SI factors and correlations between SI factor scores and perceived colleague performance. Additionally, the table shows the correlation between the average of all eight SI factor scores and perceived performance. All correlations are statistically significant at a level p < 0.001. Table 3 shows the average scores for the SI factors and performance questions.
Correlations are high, both within the SI factors and between SI factors and perceived work performance. This implies that participants see that the factors of perceived SI are closely linked to how they perceive high performance in a work environment. Especially the perceived SI factors of Systemic Perception and Effective Responsiveness have very strong correlations with perceived performance.
The strong link between factor scores and work performance is further highlighted in the scatter plot of Figure 1, which shows the average perceived SI factor score on the vertical axis and perceived performance on the horizontal axis. In general, colleague evaluations follow the regression line quite closely; in only a few cases do the two variables significantly differ.

Alternative perspectives to performance
In addition to the main performance evaluation question, some participants were also asked to answer the following two extra performance questions: "On a scale of 0-10, how much does this colleague help other people succeed?" and "On a scale of 0-10, how much does this colleague contribute to a positive work climate?" Table 2 shows also correlations between SI and these two questions.
In general, the two alternative perspectives show similar results to perceived performance; correlations between perceived SI factors and the performance evaluations are high. Some of the most significant differences related to the more interpersonal factors of perceived SI (Attunement and Positive Engagement), which correlate strongly with contributing to a positive work climate. On the other hand, the more general systems thinking factors of systemic perception and effective responsiveness have a stronger link to direct perceived performance.
Distribution of perceived systems intelligence factor scores for top, high and lower performing individuals Figure 2 presents histograms of perceived SI scores for managers and non-managers for each SI factor. The histograms are shown as cumulative bar charts with the evaluated Perceived systems intelligence colleagues split into three color-coded groups based on their perceived work performance evaluation (0-5, 6-8 and 9-10).
As the figures show, higher perceived performance and higher SI scores go closely hand in hand in all of the eight perceived SI factors. Especially for systemic perception and effective responsiveness, the two most strongly correlating factors in Table 2, nearly all managers perceived as top performing score in the upper end of the perceived SI subscales.
There are also clear differences between managers and non-managers. Comparatively, a larger portion of managers and supervisors belong to the top and high-performance categories. As shown in Table 3, managers and supervisors score higher in all eight factors of perceived SI and all three performance questions. The largest SI differences, effective responsiveness and systemic perception are also distinct visually in the statistical peaks of the distributions in Figure 2. Figure 3 shows answer distributions to each questionnaire item as violin (distribution) plots for colleagues in supervisory or managerial positions. The figures are laid out so that each perceived SI factor is in its own row. These figures are especially useful to studying which of the questionnaire items seem to be particularly important to top-performing managers, and in which items also lower evaluations are common.

Profiles of systems intelligent perceived managers
Essentially, these figures show a more fine-grained view of the distributions shown in Figure 2. Some items, such as "My colleague easily grasps what is going on," seem to be strong requirements of high performance; there are virtually no top-performing managers that receive low marks on the question.

Relationship between background variables and perceived systems intelligence
We use one-way analysis of variance to study whether different answers to background variables are linked to different levels of total perceived SI (calculated as the average of the eight perceived SI factor scores). The rejection of an analysis of variance null hypothesis implies that participants are not sampled from a common population, but rather that perceived SI evaluations and the background variable correlate in some way. Table 4 shows results for various background variables, including a broad industry grouping of technological, educational, manufacturing and other companies. The analysis of variance tests were carried out with Type III sums of squares using the R "car" package (Fox and Weisberg, 2019).
As Table 4 shows, only a few background variables are linked to the perceived SI assessment. Most notably, colleagues in manager/supervisory position differ greatly from other colleagues, and colleagues that the participant is closely acquainted with are evaluated differently from those the participant is less acquainted with. Additionally, colleague age group and acquaintance time with a colleague have smaller links, though at the level that  The analysis of variance results imply that perceived SI evaluations are not strongly dependent on many common background variables, such as gender, age, organization size or industry, and therefore, perceived SI appears to have generic applicability and similar behavior in most common cases. In applying perceived SI, care should likely only be taken on avoiding close acquaintances from evaluating each other.

Discussion
The perceived SI uses colloquial phrases as opposed to structural or semi-theoretical discourse. As a result, the inventory proposed here supports approaches that seek to develop an organization from people-centered perspectives and look for opportunities at developing the organization from the "bottom up" instead of the more classic "top down" approach. As such, it can serve frameworks such as the multilevel model of organizational learning proposed by Chou and Ramser (2019), which emphasizes the upwards helping, psychological empowerment and voice behavior of employees. As Chou and Ramser note, there are few tools and frameworks available for the "bottom up" and individual-level development of learning organizations. The perceived SI inventory helps to fill this niche as likely the first peer evaluation tool in the field.
The question of leadership in a learning organization was raised already early in the seminal paper by Bass (2000). Empirical evidence in the current paper suggests a strong correlation between high perceived SI scores and perceived performance for managers. The availability of a peer evaluation tool for SI can help the leader in finding improvement opportunities, and to recognize and avoid destructive "systems of holding back" (Hämäläinen and Saarinen, 2007a;Sasaki et al., 2015).
Many aggregates and measurements in organizational and leadership research call for assessments that are difficult for employees to peer evaluate. Based on the results presented here, the perceived SI inventory seems to be one way of evaluating relevant-to-all behavior in a way that is easily discussable and directly comprehensible while at the same time relating to perceived performance. The development could be aided by the stages of SI identified by Jones and Corner (2012). Note that the theory of empathetic leadership by Kock et al. (2019) parallels in many respects SI; empathetic leadership has positive effects on job satisfaction and follower performance. The items of SI call out to action and behaviors that anybody irrespective of her position can perceive, appreciate and potentially improve. This indicates that the SI perspective is a useful framework for an organization to adopt for its developmental discourse on any level of the organization. Not only does SI correlate with individual work performance but also it relates to how much a person supports others and the workspace environment; a systems intelligent worker is also one who builds a positive atmosphere in the workplace.
The factors of SI seem to be clearly visible to one's colleagues and at the same time, they closely relate to how employees perceive others' performance. However, the majority of the SI items pointedly talk about the "soft" aspects of performance. This highlights the human dimension called out by HRD professionals and by Rigby and Ryan with their "Copernican Turn." Integrating the SI perspective with more reifying performance indicators, a leader will likely be encouraged to pay attention to SI-related skills in herself and in her personnel. The way employees perceive their peers' performance is certainly a concern for a leader. An employee's performance might be objectively good, but if peers perceive it as weak, selfgenerated problems can be predicted for the whole. Versatile, well-rounded and humanly rich development thinking will carry an organization further than the narrow performance focus that employee experience as reducing them to objects.
As the perceived SI inventory is a rather lightweight instrument of 32 Likert-scale items, it can have wide applicability for different organizational development and improvement purposes. For example, the perceived SI inventory could be included as a component of a 360-degree feedback questionnaire, providing a viewpoint to how an employee's organizational and systems skills are perceived. The evaluation could also be relatively easily repeated later with the same individuals to study changes in perceived SI factor scores.
The perceived SI inventory provides many opportunities for developing organizations and teams. For example, it can be used as an organization-wide or team-wide intervention, where all members give perceived SI evaluations on their closest colleagues, helping individuals identify their strengths and weaknesses on the eight SI factors. The results could be further connected to the self-report SI inventory of Törmänen et al. (2016), highlighting possible differences between self-perceptions and colleague perceptions of SI. Figures 2 and 3 show two suggestions to how SI factor score distributions can be visualized and allow for the easy showing of how an individual's own scores are positioned next to the overall distribution.
In the results reported here on SI, it is very rare for there to be a person who scores low in perceived SI and still is perceived to generate good performance or vice versa. Managers receive higher evaluations in all perceived SI factors and based on the analysis-of-variance test, and managerial position is the strongest background variable affecting an individual's perceived SI evaluation.
Thus, the perceived SI inventory can have the potential to serve as a powerful tool for leadership development. The inventory and its factors can be used as part of coaching programs, for example, by focusing on any gaps an individual has in his/her SI capabilities or they can be taken as concepts and vocabulary for more informal dialogue within the workplace, optionally supported by other tools such as design games (Hämäläinen et al., 2020).

Conclusions
Based on confirmatory factor analysis results, the perceived SI inventory has good factorial validity using the original self-report SI inventory factor structure. Thus, the eight SI factors can thus be used to describe perceived SI.

TLO
The results show that all perceived SI factors correlate strongly with subjective evaluations of work performance. When studying top-performing individuals and especially top-performing managers, the effect is especially pronounced with the perceived SI factors Systemic Perception and Effective Responsiveness.
When comparing managers to other employees, managers tend to have higher perceived SI. perceived SI is also strongly linked to perceived performance in managers, and thus seems to be equally or even more important for managers than for other employees. Thus, results indicate that managers and leaders could benefit from coaching and tools that help them develop their SI capabilities.
Based on the relationship between the perceived SI factors and background variables perceived SI seems to be generic; there are only minor differences between genders, age groups or between different industry sectors. This result is maintained for both employees and managers. Indeed, it seems that systems intelligent behavior and leadership is similar no matter the place.