Organizational constraints and performance : an indirect effects model

Purpose – Surprisingly, most studies have failed to demonstrate a strong correlation between organizational constraints (conditions at work that make doing a job difficult) and job performance. The purpose of this paper is to challenge the view that constraints are a direct barrier on performance and take an alternative approach whereby constraints have an indirect effect via decreased motivation and increased workload. Further, differential effects of various constraints are examined. Design/methodology/approach – Qualitative and quantitative data were collected from 660 engineers licensed in the state of Florida using a single online survey. Findings – Qualitative results showed that the most commonly experienced constraints were from coworkers and organizational rules and procedures. Constraints identified as having a greater detrimental effect on motivation are from the supervisor, and organizational rules and procedures. Quantitative results supported an indirect effects model that includes an indirect path via motivation, and a path via workload, which had a curvilinear component. Originality/value – This is one of few studies to explain the relationship between constraints and performance, rather than simply estimate it. The use of mixed methods allows us to gain an in-depth understanding of constraints, and the convergence of findings across the methods increases confidence in this study’s results.


Introduction
Organizational constraints are defined as the aspects of the immediate work environment that inhibit the translation of motivation and abilities into effective performance (Peters and O'Connor, 1980).Common examples of constraints include insufficient information or equipment, and interruptions or inadequate help from other people.Constraints were originally hypothesized to limit the maximal level of performance, thus having a stronger effect on workers high in ability and motivation (Peters and O'Connor, 1980;Guzzo and Gannett, 1988).Whereas early laboratory studies supported the performance-inhibiting role of constraints, research in field settings revealed only small and often non-significant correlations (Gilboa et al., 2008;Spector and Jex, 1998;Villanova and Roman, 1993;Pindek and Spector, 2016a).The inconsistent results between laboratory and field suggest perhaps a more complicated mechanism is in play, but very few studies have examined how constraints impact performance, and whether these effects are uniform for different constraints.Recently, Pindek and Spector (2016a) examined possible explanations for the low correlation between constraints and performance in a series of studies that focused on methodological and perceptual moderators as explanations for the weak relationship between organizational constraints and performance.The current study argues that it is also imperative to consider the specific constraints that employees face and examine their effects on more proximal outcomes (i.e.motivation and workload), that in turn affect performance.This is in contrast to the traditional view of constraints as a direct performance inhibitor.
The current paper therefore has two main contributions: First, combining qualitative and quantitative methods on a large sample of employees, allows us to gain a deeper understanding of the impact that various constraints have on employee performance.Second, challenging the original conceptualization of constraints as a direct barrier to performance, and examining motivation and workload as possible mediators, helps bridge the gap in the literature between the strong theoretical link and the surprisingly weak empirical correlation between constraints and performance (mean r ¼ −0.07, k ¼ 17, n ¼ 8,764, when performance was not self-rated; mean r ¼ −0.16, k ¼ 5, n ¼ 1,460, when performance was self-rated; Pindek and Spector, 2016a).This goes beyond Pindek and Spector's (2016a) tests of attenuating factors for the constraints-performance relationship.
Part 1an in-depth look at constraints The majority of prior literature viewed constraints as a general, homogenous construct, which has typically been examined using quantitative methods.However, the organizational constraints construct was not initially designed as being homogeneous.In their conceptual work, Peters and O'Connor (1980) described eight distinct variables they included as situational elements relevant to performance.These variables were later captured by the 11-item organizational constraints scale (OCS) (Spector and Jex, 1998) that has been used in over 70 percent of constraints studies included in a recent meta-analysis (Pindek and Spector, 2016b).Constraints captured by this instrument include poor or lacking equipment, rules, inadequate training or instructions, the supervisor, and coworkers.There have been few previous efforts to examine the differential relationships that various constraints have with other variables.Perhaps the most prominent attempt by Liu et al. (2007) examined the cultural differences in prevalence rates between Americans and Chinese when separating interpersonal constraints from more task-oriented constraints.This is one possible way to discuss how different types of constraints can vary in their prevalence and effects.
Another lens through which differential effects for constraints may be considered is based on work by LePine et al. (2005).They introduced a two-dimensional stress model, which is focused on the appraisal of stressors as a hindrance (a threat, de-motivator or "bad stress") vs a challenge (a motivator or "good stress"), having differential correlations with strain, motivation and performance.A recent study that applied a challenge-hindrance approach to constraints (Pindek and Spector, 2016a) found support for a more complicated view of constraints as having both challenge and hindrance components.Pindek and Spector (2016a) argued, but did not test, that some constraints are perceived as hindrance stressors (e.g.lacking equipment or supplies), while other constraints can be motivating and have a more challenging effect (e.g.inadequate training can lead some workers to invest more effort in developing skills on their own).Indeed while the original work by LePine et al. (2005) considered constraints to be a hindrance, there have been other studies examining challenging aspects of at least some constraints.Dealing specifically with resource scarcity 80 JMP 34,2 (one dimension of constraints), Sonenshein (2014) demonstrated how limited resources can foster creativity.In that sense, constraints can in fact be challenging and enhance motivation to overcome the difficulty.In the same vein, Fritz and Sonnentag (2009) found that constraints predict taking charge at work, presumably because constraints represent a suboptimal situation that can be changed through proactive behavior.It is therefore possible that some constraints are not as harmful to employees' well-being and performance.The current study uses a qualitative approach to gain a deeper understanding of constraints, one that does not assume that all constraints are equal in their effects.
While most organizational constraints research is conducted using quantitative methods (Spector and Pindek, 2016) a more inductive, qualitative approach presents an alternative where participants can be asked various questions regarding their job experiences (Keenan and Newton, 1985).Such qualitative data can provide detailed information about which constraints are more important or prevalent (Liu et al., 2007), and what the effects of those different constraints might be.The use of open-ended questions in an inductive manner can reveal explanations that were not previously hypothesized.This is particularly useful in the case of the relationship between constraints and performance, which was assessed in many previous studies, but for the most part it was left unexplained.
Several researchers have used qualitative approaches to study job experiences.For example, Motowidlo et al. (1986) examined the effects of stressful work events on stress, affect and performance decrements in nurses.Other qualitative studies examined types of stressors, reactions and consequences (e.g.Narayanan et al., 1999;Liu et al., 2007;Parkes, 1985;Keenan and Newton, 1985;Mazzola et al., 2011).In all these studies, the qualitative information obtained provided unique insights into the effects of stressful work experiences.A similar technique is therefore employed in the current research to gain insight on the following questions: RQ1.What are the most constraining aspects of the work environment?RQ2.Are the different constraining aspects of the work environment related differentially to motivation and performance?
Part 2an indirect relationship between constraints and performance The mediating role of motivation In their discussion of the pitfalls of studying the constraints-performance relationship, Bendoly and Hur (2007) argue that researchers may need to go beyond the traditional view of direct effects that represent a mechanistic view of the capabilities of workers and the extent to which their resources enabled them to use those capabilities.When trying to explain the weak association between constraints and performance, one approach is to consider other performance indicators, such as counterproductive behaviors that have stronger relationships with constraints (e.g.Clark and Walsh, 2016).An alternative approach considers the negative effect that constraints have on motivation, and subsequently on task performance.This latter approach is employed in the current research.As mentioned earlier, LePine et al. (2005) mapped constraints to be, overall, a hindrance stressor in their two-dimensional stress model.Hindrance stressors are expected to potentially harm personal growth or gain, and to trigger an array of reactions, including negative emotions and withdrawing from the situation as a means of coping.
When considering a conceptually similar construct to motivation, i.e., work engagement (Gagne and Deci, 2005) and using the job demands-resources model as an organizing framework (Bakker et al., 2008), it is evident that engagement is associated with job performance: engaged individuals are better performers, and job resources that promote performance (e.g.coaching and performance feedback from the supervisor) are some of the best predictors of engagement.Because motivation and engagement are related constructs, constraints, which are conditions at work that make it difficult to complete the job tasks, would therefore be expected to impact performance via lower motivation because they represent not only an increase in demands but also a lack of resources (e.g.constraints include lack of training and poor equipment, which can be considered poor resources).It is this combination of high demands and low resources that is demotivating (Bakker et al., 2007).
Therefore, the current study proposes an indirect effects model of the effects of constraints on performance.As part of this model, there is a hypothesized significant indirect path in the constraints-performance relationship through motivation.Furthermore, an overall negative effect on motivation, as is supported by generalized meta-analytic findings (e.g.Pindek and Spector, 2016a), is expected: There is an indirect path from constraints to performance via motivation.
Specifically, constraints are negatively related to motivation, and motivation is positively related to performance.
The mediating role of workload Workload refers to the volume, pace and difficulty of work (Spector and Jex, 1998).Since organizational constraints interfere with performance by definition, and result in employees having to extend more effort or spend more time in order to get their tasks done, they likely contribute to a higher workload.Some constraints are interruptions, which reduce the time employees can devote to their tasks.Other constraints, such as inadequate resources or help from coworkers, can increase the actual work that needs to be completed.There is support for a positive correlation between constraints and workload (average uncorrected r of 0.39; Pindek and Spector, 2016b), but the nature of the relationship between them has not received much attention.Constraints are expected to lead to higher workload, which serves as a mediator in the constraints-performance relationship.Gilboa et al.'s (2008) meta-analysis examined both the linear relationship between role overload and job performance (r ¼ −0.06), and between constraints and job performance (r ¼ −0.19).One way to conceptualize workload that is too great (termed role overload, and defines as having too many expected responsibilities or activities; Rizzo et al., 1970) is that it occurs when an individual perceives he or she does not have the necessary resources to deal with demands (Kahn et al., 1964).According to this conceptualization, constraints (which include a lack of necessary resources) contribute to an increase in role overload, or workload.As workload increases, employees will struggle to complete tasks, and thus performance will suffer.Thus, workload is expected to mediate the organizational constraints-performance relationship: H2a.There is an indirect path from constraints to performance through workload.
Specifically, constraints are positively related to workload, and workload is negatively related to performance.
The relationship between workload and performance is not a straightforward one, as workload should not impede performance as long as it is not too high.Accordingly, some researchers argue that a curvilinear relationship may be a more appropriate depiction of the effects of workload on performance (e.g.Bendoly and Hur, 2007).For example, Janssen (2001) found a curvilinear (inverted U-shaped) relationship between quantitative demands (similar to workload) and job performance.The inverted U pattern suggests that both workload that is too high (exceeds capacity) and workload that is too low (not challenging enough), can result in decreased performance.This non-linear relationship between workload and job performance can partially explain why there appears to be a lower 82 JMP 34,2 correlation (linear relationship) between the two variables.We therefore expect a curvilinear path between workload and performance: H2b.There is a significant curvilinear path from workload to performance such that when workload is at low and high levels, performance is lower than when workload is at intermediate levels.

Method
Participants and procedure A single online survey was used for both the qualitative and quantitative parts of the study.
Participants in this study, all volunteers, consisted of 660 licensed engineers employed in Florida.Engineers were recruited through e-mails that are available from a public database.Participants were on average 53 years old (SD ¼ 14), and most were men (88 percent).Their average tenure on the current job was 8 years (SD ¼ 11), 91 percent were working full time and 9 percent were working part-time, and job titles and employing organization varied.The questionnaires were distributed through a web-based platform, and were part of a larger investigation.Constraints are particularly relevant to the type of work engineers do, as was demonstrated in the previous research that used open-ended questions and found that "ineffective and wasteful use of time" (similar to the definition of constraints) was a main source of stress (Keenan and Newton, 1985).

Qualitative measures
The qualitative part of the questionnaire contained three open-ended questions concerning the nature of constraints.They were preceded by the following explanation: "This is a study of situational constraints or obstacles, which are conditions at work that interfere with performing the job.These can include lacking or inadequate information, equipment/ budget, services from others, or training.They could also be interruptions from others, lack of cooperation or anything that makes it difficult or impossible to do the job".The questions were: "What aspects of your work did you find constraining?","How did these constraints affect your motivation to do your job?" and "How did these constraints affect your actual performance?"(Bold in original).Two independent subject matter experts (co-authors of this report) read all the responses and independently coded them into categories that were not predetermined.After discussing the overlap between the two coders' categories, a third subject matter expert collapsed similar categories and coded all responses again.There were some cases of disagreements in categorizations between the first two coders (6.2, 2.2 and 4.3 percent disagreement on questions 1, 2 and 3, respectively), and those disagreements were resolved by the third coder.

Quantitative measures
The participants were asked to respond to four previously developed quantitative scales.For each of the scales, the most recent workday was used as the time reference.This time reference was used because performance measures can have range restriction, particularly when they are self-rated, because selection systems attempt to weed out potential poor performers at the hiring stage, and those that are mistakenly hired are likely to then be fired.Another reason for range restriction is biases in self-appraisals.The reasoning is that an employee would be more likely to experience and be willing to report a relatively low-performance day (internal comparison) than low performance in general (external comparison).Unless stated otherwise, all items were measured on a six-point scale ranging Organizational constraints from 1 ("strongly disagree") to 6 ("strongly agree"), and α reliability coefficients are reported in Table V.
Organizational constraints.The Spector and Jex's (1998) OCS was used.The reference for the 11 items was "How difficult was it to do your job during the most recent day because of__."Items were measured on a four-point scale ranging from "not at all difficult" to "extremely difficult."A sample item is "poor equipment or supplies." Performance.The Williams and Anderson's (1991) subscale of in-role performance was adapted for self-assessment.Of the seven original items, one was excluded due to low factor loading in the original work ("engages in activities that will directly affect his/her performance evaluation").A sample item is: "performed tasks that were expected of me." Motivation.Items from two sources were combined to measure motivation.First, we took the seven items that reflected intrinsic motivation from the Utrecht Work Engagement Scale (UWES) (Schaufeli et al., 2006).A sample item is "I felt enthusiastic about my job."The α reliability coefficient for the seven items was 0.89.The engagement scale is an appropriate measure for motivation (Ryan and Deci, 2000;Schaufeli et al., 2002).Second, we wrote three items designed specifically to assess internal motivation: "I felt like doing my job today," "I felt motivated to do my job today" and "Today I felt like I was ready to tackle obstacles on my job."The three new items had an α reliability coefficient of 0.93.Since the resulting patterns were similar when using either the seven UWES items or the three new items, the correlation between them was 0.78, and an exploratory factor analysis with common factor analysis extraction suggested one factor based on the scree plot criterion, all ten items were combined.
Workload.The Spector and Jex's (1998) quantitative workload inventory was used.The scale includes five items.A sample item is "My job required me to work very fast."

Qualitative results
Frequencies of responses to the open-ended questions are reported in Table I, and examples of statements that were coded into each category can be found in Appendix 1.Note that because some responses were coded as belonging to more than one category, percentages are calculated both out of the number of participants who answered the question (100 percent), and out of the total number of endorsements of categories (over 100 percent).As can be seen in Table I, the most frequently mentioned constraints are interruptions or inadequate help from coworkers (mentioned by 23 percent of the participants), organizational rules and procedures (21 percent of the participants), the supervisor or management (12 percent of the participants) and poor equipment (11 percent of the participants).These results provide some answers to RQ1.Furthermore, some participants (27 percent) reported a motivation decrease, some reported no effect (21 percent), and some reported an increase (18 percent).The effects on performance were mainly a decrease (53 percent of the participants) or no effect (34 percent of the participants).Few participants (1 percent) mentioned an increase in performance.
To answer RQ2, three χ 2 tests for categorical data ( χ 2 contingency tables) were conducted using SPSS version 22.0 to establish the relationships between engineers' perceived constraints, their effects on motivation and their effects on actual performance.Prior to conducting the χ 2 tests, every response category that was endorsed by fewer than 10 percent of the participants was excluded so we could conduct a meaningful analysis on the data.One response category was excluded from the effects on motivation question ("Make angry/ frustrate") because it was not a direct answer to that question.We also excluded participants who belonged to more than one category (though this exclusion did not change the pattern of results).Results for the three χ 2 analyses are presented in Tables II-IV.In addition to 84 JMP 34,2 reporting the observed values for each cell in the contingency tables, the percentages for each cell in relation to the row total are also provided for ease of interpretation.Post hoc tests and effect sizes (Ferguson, 2009) were also calculated.
The χ 2 test presented in Table II indicates a significant dependency between the type of constraints and its effect on motivation ( χ 2 (6) ¼ 23.45, po0.01,Cramer's V ¼ 0.23).Post hoc tests using the adjusted standardized residuals with a Bonferroni correction to the p-values revealed that only two cells had a significantly different proportion: supervisors as a constraint was associated with an increase in motivation (Z ¼ 3.00, p ¼ 0.003), and coworkers as the constraint was associated with a decrease in motivation (Z ¼ −2.94, p ¼ 0.003).When examining the estimated proportions, organizational rules and procedures, and the supervisor/management were most often associated with reduced motivation (56 and 66 percent, respectively).For the other two categories, the chances of motivation decreasing were considerably lower.The χ 2 test presented in Table III indicates there is no dependency between the type of constraints and its effect on performance ( χ 2 (3) ¼ 0.97, p ¼ 0.81, Cramer's V ¼ 0.06).

Organizational constraints
The χ 2 test presented in Table IV indicates a significant dependency between the effects of constraints on motivation and on performance ( χ 2 (2) ¼ 14.01, p o0.01, Cramer's V ¼ 0.22).Post hoc tests using the adjusted standardized residuals with a Bonferroni correction to the p-values revealed that only for those who experienced a decrease in motivation there was a significant difference in effects on performance (Z ¼ 3.74, p ¼ 0.0002).Examining the estimated proportions reveals that when the effects on motivation were negative, it was much more likely (71 percent) that the effects on performance were negative as well, while when the motivational effects were positive or unchanged, the likelihood of reduced performance was 49 and 50 percent, respectively.

Quantitative results
Descriptive statistics and correlations between the variables included in the quantitative part of the study are presented in Table V performance are all significantly correlated.Workload is significantly correlated with constraints and performance, but not with motivation.
An indirect effects path analysis model was tested using the maximum likelihood estimator in Mplus (Muthén andMuthén, 1998-2012), using bias-corrected bootstrapped confidence intervals for the indirect effect.We chose to test a path model rather than a full structural equations model because of the formative nature of the constraints scale, and the curvilinear component in our structural model.Furthermore, this approach allows us to easily compare competing models without having the fit of the measurement model mask the fit of the structural model.
To increase our confidence in the proposed model, a series of theoretically plausible competing models were tested.The first alternative model (Alternative model 1) is similar to  (3) ¼ 32.83, p o0.001, RMSEA ¼ 0.13, CFI ¼ 0.90, TLI ¼ 0.80, SRMR ¼ 0.06).Fit statistics for the proposed model as well as the three alternative models are presented in Table VI.

Additional analysis
In addition to testing the hypotheses as part of a model, the quantitative data allowed replicating some of the findings from the qualitative part.One replication was of the finding that certain types of constraints that were deemed more harmful to employees' motivation (organizational rules and procedures, and the supervisor/management) and other constraints were less harmful to motivation (interruption or inadequate cooperation/help from coworkers, and poor equipment or lack of money).The other replication was of the non-significant differences between the constraints types and their effect on performance in the qualitative part.
In order to do that, two subject matter experts examined the 11 items from the OCS scale (Spector and Jex, 1998) and chose the items that appropriately mapped onto one of the four most commonly reported constraints from the qualitative results.This mapping was done in order to allow a direct replication of significant and non-significant differences between constraints in their relationship with motivation and performance.There was 100 percent  agreement between the experts on the categorization of items shown in Table VII.Therefore, Items 2 and 4 (rules and management) were expected to have a stronger relationship with the quantitative measure of motivation than items 1, 3 and 5 (equipment and interruptions or inadequate help).No significant differences between the correlations of the OCS items with performance were expected.As seen in Table VII, all correlations are significantly different from 0 ( p o0.05).The test for differences between specific pairs of the (dependent) correlations using an asymptotic z-test (Steiger, 1980;Hoerger, 2013;Lee and Preacher, 2013) supported our predictions.For correlations with motivation, Item 2 (which had the lowest correlation with motivation among Items 2 and 4, r ¼ −0.28) was compared to Item 3 (which had the highest correlation with motivation among Items 1, 3 and 5, r ¼ −0.18).Since this is testing for differences between the pair of correlations that are closest to each other, if this pair has a significant difference, then all other pairs would be significantly different from each other as well.Items 2 and 3 are significantly correlated with each other (r 2,3 ¼ 0.33, p o0.005), but their relationships with motivation are significantly different (Z ¼ 2.18, po 0.05).Therefore, there is support in the quantitative data to the finding that organizational rules and procedures, and the supervisor/management are more harmful to motivation than interruption or inadequate cooperation and help from coworkers, and poor equipment or lack of money.
For correlations with performance, Item 5 (which had the lowest correlation with performance, p ¼ −0.10) was compared to Item 4 (which had the highest correlation with performance, p ¼ −0.18).Since this is testing for differences between the pair of correlations that are the farthest from each other, if this pair has a non-significant difference, then all other pairs would not be significantly different from each other as well.Items 4 and 5 are significantly correlated with each other (r 4,5 ¼ 0.21), but their relationships with performance are not significantly different (Z ¼ 1.57, p ¼ 0.12).Therefore, there is no support in the quantitative data for differential effects of different types of constraints on performance.This replicates the qualitative results for this relationship.

Discussion
This research utilized a mixed-methods approach to gain a better understanding of the nature of constraints and its indirect relationship with performance.The qualitative part examined constraints in depth, and the quantitative part replicated and expanded on those results.Specifically, we demonstrated how different types of constraints might have both direct and indirect effects on performance.

Differential effects of types of constraints on motivation
This qualitative part of the study focused on examining organizational constraints in depth.First, a qualitative method (open-ended questions) was used to assess which constraints were most frequently reported.The two most prominent constraints were interruption or

89
Organizational constraints inadequate cooperation and help from coworkers, and organizational rules and procedures, each being mentioned by over 20 percent of the participants.Two more constraints were mentioned by over 10 percent of the participants: their supervisor or management, and inadequate material resources (poor equipment or lack of money).Furthermore, the dependency between the type of constraints reported and their perceived effect on the participants' motivation and performance was examined.Of the four types of constraints mentioned most often, organizational rules and procedures, and the supervisor/management were both associated with a greater likelihood of decreased motivation.However, all constraints had a similar likelihood of being associated with lower performance levels.A comparison of the two types of constraints that are more likely to result in decreased motivation with the other constraints shows the constraints that negatively affect motivation (organizational rules and procedures, and the supervisor/management) are more directly representative of the psychosocial aspects of the organization and its culture rather than the more physical aspects of the environment, such as inadequate equipment.
When looking at the differential effect that type of constraints has on motivation and performance, the results from the qualitative part were perfectly replicated using quantitative data.Specifically, two types of constraints (organizational rules and procedures, and the supervisor/management) had a larger correlation with participants' motivation than did the other constraints.Conversely, all types of constraints had a similar relationship with performance levels.The perfect replication of these results across both qualitative and quantitative data increases our confidence that these two constraints truly are more harmful to employees' motivation.This could be because management and organizational rules more directly represent the psychosocial aspects of the organization and its culture than constraints such as poor supplies or interruptions from coworkers, and therefore might be more easily attributed to organizational injustice.Future studies may examine whether these specific constraints have additional harmful effects on motivation through other organizational characteristics, such as lower organizational justice perceptions (justice perceptions have been shown to relate to performance via motivation, e.g.Zapata-Phelan et al., 2009).

The indirect constraints-performance relationship
The mechanisms through which constraints affect performance have not received much empirical attention, most likely because constraints were originally viewed as a direct barrier to performance (Bendoly and Hur, 2007;Peters and O'Connor, 1980).In this research, a more comprehensive model is advocated, whereby constraints indirectly affect performance via motivation and workload.Both the qualitative and quantitative analyses in this research supported these indirect effects.When responding to the open-ended questions, the largest group of respondents (30.5 percent) noted that constraints were a de-motivator for them.However, another substantial group of participants (20.5 percent) indicated constraints actually increased their motivation to find solutions and complete their tasks despite whatever constraints they were encountering.Furthermore, while most participants (55.4 percent) indicated that constraints had a negative effect on their performance, they were much more likely to do so if they had indicated in the previous question that constraints were a de-motivator.This pattern of results is consistent with an indirect effect via motivation.This same path was significant in the model tested using quantitative methods.
The quantitative model provided support for an indirect effect for constraints on performance via workload, in addition to the indirect effect via motivation.Specifically, constraints increase the level of workload, and workload is negatively related with performance, particularly at very high levels of workload, as it is then that employees become overloaded, which hurts their ability to complete their tasks efficiently.We found 90 JMP 34,2 evidence that the workload to performance path was not linear, as the decrease in performance accelerated as workload increased.However, there was no evidence to support the expected decrease in performance associated with the lower levels of workload.
Strengths, limitations and contributions of the study Our use of a mixed-methods approach in examining the constraints-performance relationship is a strength of this research, which allowed us to address the role of motivation using dissimilar methods that converged in their findings.Specifically, both the qualitative and quantitative analyses were consistent in identifying those specific constraints that had the greatest negative effect on motivation, and in indicating an indirect effect of constraints on performance via motivation.Another strength of the current study is the relatively large sample, particularly for a qualitative analysis.This allowed us to get a wide variety of responses to our open-ended questions, including responses with relatively low base rates.On the other hand, as all our participants were sampled from a population of engineers licensed in Florida, the ability to generalize to other populations is limited.
One important limitation of the current study is that its cross-sectional nature limits our ability to draw conclusions regarding time precedence.It is also an important limitation in examining the mediation effects (Stone-Romero and Rosopa, 2008).However, the types of questions used in the qualitative investigation allowed gleaning the process as it is perceived by the participants, although one must recognize that respondents were asked to make attributions that might or might not be completely accurate.Another issue is the concern that common method variance is inflating the observed correlations.This concern, however, is mitigated to a large extent by the small and non-significant correlation (r ¼ −0.11) between two of the variables: workload and motivation, suggesting that there is not a constant inflation effect on these correlations (Lindell and Whitney, 2001).Furthermore, the choice of using self-reports of performance and the most recent day at work as the time reference were an attempt to disattenuate the constraints-performance relationship (Pindek and Spector, 2016a), but risk some information loss.Finally, it is unlikely that we have captured the entire complexity of the process by which constraints affect performance.There are other possible variables that may serve as mediators or moderators in this complex process.One possible variable is conscientiousness (i.e. a personality trait capturing ambitiousness, responsibility and ethical behavior) might have direct and moderating effects.For example, individuals high on conscientious and low on neuroticism tend to be better at dealing with stressors such as constraints and workload (Bowling and Eschleman, 2010), and their performance may not suffer as for those low on conscientiousness.
The main theoretical contribution of the paper is in highlighting that constraints are not simply a direct technical/mechanical barrier limiting performance.Rather, employees who encounter constraints experience a decrease in motivation and an increase in the workload, which help explain the negative impact that constraints have on performance.Moving away from a mechanistic view of constraints and adopting an approach that is more focused on the experiences that employees have when they encounter constraints on the job provides some much-needed insights on how organizations may better manage their employees' work environment and performance.

Practical implications
It is certainly not surprising that constraints due to inadequate resources would adversely affect job performance.It is also likely that in such cases, managers would note that they do not have the resources to give.Perhaps the most important contribution of this study is to show that the sorts of constraints that are most common and harmful in terms of motivation do not arise from inadequate resources, but rather from colleagues, managers, and organizational rules and procedures.These are issues that can be addressed relatively cheaply by focusing attention on sound organizational practices, especially leadership.Top leadership is responsible for assuring that rules and procedures make sense and facilitate rather than inhibit task performance.This requires input from all levels of the organization to be sure that a proposed rule or policy does not have the unintended consequence of adversely affecting motivation and organizational efficiency.Top leaders also set a tone that can filter throughout the organization in how they operate with individuals in subordinate positions, and this means promoting cooperation and teamwork.Individuals in supervisory positions throughout the organization should be trained in how best to facilitate teamwork, and all employees need to be made aware of how important cooperation is among all employees.
Finally, this research has potential implications for the measurement of constraints.For example, by highlighting those constraints that are the most prevalent, and those that are most harmful for motivation and performance, future measures of constraints may focus on those constraints when creating shorter scales, such as needed for diary designs.Furthermore, measures can be developed that consider constraints to be a multidimensional construct, with subscales focusing on specific categories of constraints, such as coworkers, management, rules/procedures, and clients/customers/patients, as these different sources of constraints might have different implications for organizations.

Concluding remarks
This study is one of very few that examined the complex nature of organizational constraints, and their relationship with performance.While there are many reports of the magnitude of the constraints-performance relationship, few have considered the mechanisms by which constraints might affect performance levels.Using a mixed-methods design, this study found support for indirect effects via motivation and workload and is an important step in understanding the underpinnings of this prevalent stressor.
Figure 1.The final indirect effects model Figure 2. The quadratic effect of workload on performance . Note that constraints, motivation and Total n ¼ 232.The percentages shown in parentheses correspond to each type of constraints separately (rows sum to 100 percent but columns do not) and reflect the proportion of individuals who reported each category of effect on motivation.χ 2 (6) ¼ 23.45, po 0.01 Total n ¼ 286.The percentages shown in parentheses correspond to each type of constraints separately (rows sum to 100 percent but columns do not) and reflect the proportion of individuals who reported each category of effect on performance.χ 2) ¼ 0.97, p ¼ 0.81 Total n ¼ 289.The percentages shown in parentheses correspond to each type of motivational effect separately (rows sum to 100 percent but columns do not) and reflect the proportion of individuals who reported each category of effect on performance.χ2(2) ¼ 14.01, po 0.01 Notes: α values are shown in the diagonal.Correlations with an absolute value W0.11 are po 0.01; correlations with an absolute valueW 0.09 are p o0.05.Gender had the values of 1 (men) and 2 (women)