Two studies were conducted to analyze the measure of PsyCap and test the hypotheses.
Study 1 utilized three samples of management students with an age range that can be termed emerging adults (Arnett, 2000).
Study 2 used two separate samples of employees in both service and high-technology manufacturing environments to test the hypotheses in the field. With Study 1, we first review the initial psychometric properties with Sample 1 examining the factor structure, Sample 2 examining the nomological network, and Sample 3 examining the test–retest statistics and additional discrimination from related constructs. With Study 2, we test hypotheses with two independent samples of manufacturing engineers (Sample 1) and insurance service employees (Sample 2). Each sample is generally discussed in this order below, and for more clarity when referring to the two samples in Study 2, we will use the terms “high tech manufacturing” and “services.”
Samples for Study 1
The first sample in Study 1 consisted of 167 management students from a large state university in the Midwest. These participants had an average age of 22.25 years (SD = 1.41) and 67% were men. The second sample of Study 1 was drawn about 5 months later from different management students at the same university and from a second large university from the mideastern United States. These 404 participants in this second sample were similar to the first in terms of demographics (average age 21.10 years, SD = 2.66 and 58% were men). Finally, to investigate the stability of the PsyCap measure, we administered a series of scales at three points in time over the course of 4 weeks to 174 different management students from the same midwestern university noted above.
Samples for Study 1
The first sample in Study 1 consisted of 167 management students from a large state university in the Midwest. These participants had an average age of 22.25 years (SD = 1.41) and 67% were men. The second sample of Study 1 was drawn about 5 months later from different management students at the same university and from a second large university from the mideastern United States. These 404 participants in this second sample were similar to the first in terms of demographics (average age 21.10 years, SD = 2.66 and 58% were men). Finally, to investigate the stability of the PsyCap measure, we administered a series of scales at three points in time over the course of 4 weeks to 174 different management students from the same midwestern university noted above.
Samples for Study 2
The high-tech manufacturing sample for Study 2 consisted of engineers and technicians from a very large (Fortune 100, over 150,000 employees) firm. These 115 participants averaged 44.83 (SD = 7.31) years and 80% were men. The service sample for Study 2 was made up of employees in all functions and levels of a midsized (about 900 employees) insurance services firm (i.e., they service insurance policies from other firms). These 144 subjects averaged 33.79 (SD = 10.85) years and 65% were women.
Procedures for the Studies
In Study 1, management students consenting to participate in an “Organizational Behavior and Leadership” project were provided a Web address to register. They were then sent a unique password via e-mail that allowed them to log in and take a short questionnaire survey. Following the recommendations of Podsakoff and colleagues (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) to reduce same source/common methods bias problems from questionnaire surveys, we administered the study survey at two points in time. The first part of the survey containing the predictor study variables was taken in the first session. Then a week later, they logged back in and completed the survey that included performance and satisfaction study variables.
The high-tech manufacturing sample for Study 2 consisted of engineers and technicians from a very large (Fortune 100, over 150,000 employees) firm. These 115 participants averaged 44.83 (SD = 7.31) years and 80% were men. The service sample for Study 2 was made up of employees in all functions and levels of a midsized (about 900 employees) insurance services firm (i.e., they service insurance policies from other firms). These 144 subjects averaged 33.79 (SD = 10.85) years and 65% were women.
Procedures for the Studies
In Study 1, management students consenting to participate in an “Organizational Behavior and Leadership” project were provided a Web address to register. They were then sent a unique password via e-mail that allowed them to log in and take a short questionnaire survey. Following the recommendations of Podsakoff and colleagues (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) to reduce same source/common methods bias problems from questionnaire surveys, we administered the study survey at two points in time. The first part of the survey containing the predictor study variables was taken in the first session. Then a week later, they logged back in and completed the survey that included performance and satisfaction study variables.
The on line method was also used in Study 2 to gather survey data on the predictor variables from those consenting to participate in the insurance service firm. All participants were advised through informed consent that company performance evaluations would be linked with their survey responses to inform organizational research. The most recent performance data on participants in this sample were gathered from the human resources department a month after the survey was taken. This one month lag was deemed to be appropriate because of our proposed statelike properties of PsyCap. The basis for the appropriateness of a month can be found in our introductory discussion of the state-like nature of PsyCap. Specifically, in contrast to unstable states such as pleasure, positive moods, and happiness, the “state-like” PsyCap is proposed to be relatively more stable and we used one month as a reasonable period of time for conducting a preliminary examination of the performance relationship with PsyCap. For the high-tech manufacturing firm in Study 2, members of the engineering group were sent an “Attendance Optional” meeting notice. During this meeting, the chief engineer (first level executive) announced an opportunity to be a part of an “Organizational Behavior and Leadership” project. Those participating agreed to the informed consent and linking their survey responses to the firm’s performance evaluations. Administered on site by the outside researcher to assure confidentiality, they completed the survey containing the predictor variables. Similar to the service firm, the most recent performance data for these participants in the high-tech manufacturing firm were again gathered a month after the survey from the human resources department. As described above, this one month was deemed an appropriate period of time given the state-like nature of PsyCap. These data were based on both objective and rated performance already being collected by the organization.
PsyCap Measure
The members of the research team for this study, with additional consultation and input from colleagues doing similar research, selected the scales for each of the four positive facets. The selection criteria were not only that the scale had to demonstrate reliability and validity in the published literature and have relevance to the workplace, it also had to either be developed as, or capable of, measuring the state-like constructs making up PsyCap. The four scales that were determined to best meet these criteria were (a) hope (Snyder et al., 1996); (b) resilience (Wagnild & Young, 1993); (c) optimism (Scheier & Carver, 1985); and (d) self-efficacy (Parker, 1998).
PsyCap Measure
The members of the research team for this study, with additional consultation and input from colleagues doing similar research, selected the scales for each of the four positive facets. The selection criteria were not only that the scale had to demonstrate reliability and validity in the published literature and have relevance to the workplace, it also had to either be developed as, or capable of, measuring the state-like constructs making up PsyCap. The four scales that were determined to best meet these criteria were (a) hope (Snyder et al., 1996); (b) resilience (Wagnild & Young, 1993); (c) optimism (Scheier & Carver, 1985); and (d) self-efficacy (Parker, 1998).
Each of these four selected scales have considerable psychometric support across multiple samples in prior research and have also been verified in workplace studies by themselves or in combination (e.g., Jensen & Luthans, 2006; Larson & Luthans, 2006; Luthans et al., 2005; Peterson & Luthans, 2003; Youssef & Luthans, in press). As far as meeting the statelike selection criterion is concerned, the selected hope scale of Snyder et al. (1996) was specifically developed and supported as “State Hope.” Although the Scheier and Carver (1985) scale is associated with dispositional optimism (or life orientation), this instrument has also been demonstrated to be capable of measuring state-like optimism (Shifren & Hooker, 1995). Resiliency and efficacy scales such as those selected are generally associated with state-like measurement, but the Parker (1998) efficacy scale departs from the specific task magnitude and strength measurement suggested by Bandura (1997). Nevertheless, as explained in the previous discussion of efficacy, the Parker scale (1998) is specific to the work domain, and its use of a Likert-type scale rather than traditional magnitude and strength has considerable psychometric support as a measure of efficacy (Maurer & Pierce, 1998). The four selected measures provided the foundation and pool of items from which the research group developed the PsyCap questionnaire (PCQ) measure. Two major criteria were used by the group in constructing the PCQ. First, we proposed that each of the four constructs would have equal weight, so the best six items from each of the four measures would be selected. Second, the selected items should have face and content validity with being statelike and relevant to the workplace or adaptable to wording changes to make them relevant. The group reached agreement on the 24 items and put the response choices into a 6-point Likert-type scale (1 = strongly disagree, 2 = disagree, 3 = somewhat disagree, 4 = somewhat agree, 5 = agree, 6 = strongly agree). To facilitate the state-like framing, the PCQ asks the respondent to describe how you think about yourself right now. The PCQ in its entirety can be found in Luthans, Youssef, and Avolio (2007); here are some sample items: (a) efficacy: “I feel confident in representing my work area in meetings with management” and “I feel confident helping to set targets/goals in my work area”; (b) hope: “Right now I see myself as being pretty successful at work” and “If I should find myself in a jam at work, I could think of many ways to get out of it”; (c) resilience: “When I have a setback at work, I have trouble recovering from it, moving on (R)” and “I usually take stressful things at work in stride”; and (d) optimism: “I always look on the bright side of things regarding my job” and “If something can go wrong for me work-wise, it will (R).”
Performance Measures
Study 1 used a 4-item self-rated performance measure (e.g., How would you rate your performance/effectiveness as compared with your peers?). The scale was framed by asking participants to rate their performance in their current job over the past week. If they were not employed then, they were asked to rate their academic performance over the past week. This measure demonstrated adequate reliability (α >.70) and was only used to examine the nomological network of PsyCap and, because it was a self-measure, was not used to test any hypotheses. Study 2, on the other hand, used actual performance evaluations that were gathered independent of the study. Hence, the performance measures were based on objective data and managerial ratings of participants obtained from the human resources department records of the two organizations studied. For the high-tech manufacturing firm, also as prescribed by its appraisal process, each participant’s performance measure included a sum of ratings based on quality and objective quantity of their work on electrical subsystem designs including error and rejection rates, meeting the schedule, complexity of assignment, and ability to work with peers. This measure was then cross-checked by all managers within a given job family to ensure consistency in performance ratings across work units. Although each engineer may be performing a set of slightly different tasks in this appraisal process, all participants had similar job descriptions, performance evaluation criteria, and were considered peers in that they had similar jobs in term of procedures and deliverables. As consistent with the organization’s policy, managers of job families (up to 15 managers) normalized the ratings to settle on a final performance rating. The insurance services firm provided the most recent performance rating for each participant 1 month after they had taken the PsyCap survey. The ratings were based on the most recent month of performance (i.e., after the survey had been administered). These data consisted of input from both objec
556 Luthans et aL. in Personnel Psychology 60 (2007)
tive performance data (e.g., number of claims processed) and their manager’s overall evaluation as prescribed by the firm’s performance appraisal process in one total composite score.
Job Satisfaction Measure
In addition to performance, this study also examined the relationship of PsyCap with job satisfaction. As commonly used in organizational behavior research (e.g., Judge & Bono, 2001), all of our samples but one used a 3-item scale adapted from Hackman and Oldham (1980) using the same 1–6 rating as for the PsyCap measure. This satisfaction scale had high internal reliability, Cronbach alphas (.89, .87, and .86), for the three samples. To meet its concern with the length of the survey, the high-tech manufacturing sample used a one item overall job satisfaction question (“How satisfied are you with your job?”). We also gathered affective organizational commitment (Allen & Meyer, 1996; 1990) data in Sample 1 of Study 1. The purpose of gathering these additional data was to aid in determining the discriminant validity of the PsyCap instrument and to generate a better understanding of the nomological network of constructs for the proposed PsyCap measure. The affective dimension of organizational commitment has been noted for its unique contribution, given it captures the employee’s affective desire to remain with the organization versus a calculative conclusion (Judge & Bono, 2000) and is often used as a single dimension in organizational research (e.g., Bono & Judge, 2003; Judge & Bono, 2000). On the basis of face and content validity, the research team selected four items from Allen and Meyer’s affective commitment scale for this measure. An example item is “I would be quite pleased to spend the rest of my life working for this organization.” These items demonstrated a reliability coefficient of .89. We gathered the job satisfaction data for all samples and affective organizational commitment for Sample 1 of Study 1 one week later than the predictor variables to minimize potential same-source effects/bias. As noted by Podsakoff et al. (2003, p. 887), this temporal separation procedure “makes it impossible for the mindset of the source or rater to bias the observed relationship between the predictor and criterion variable, thus eliminating the effects of consistency motifs, implicit theories, social desirability tendencies” and other individual attributes that may influence or bias the responses. In addition to the temporal strategy of data collection, to confirm the accuracy of the self-reported demographic data, we randomly cross-checked against actual personnel records and found no inconsistencies.
Psychometric Analyses
Using guidelines offered by Schwab (1980) and Pedhazer and Schmelkin (1991), we determined several requisite conditions for the PsyCap measure. These parallel the conditions and guidelines used for determining the core self-evaluations construct of Judge and colleagues (2003). Specifically, the following needed to be established: (a) content validity such that each facet is equally represented in the overall PsyCap instrument, which we established as discussed above in constructing the 24-item questionnaire; (b) sufficient PsyCap scale reliability; (c) PsyCap must have a unitary factor structure consistent with the proposed latent construct; (d) convergent validity with other theoretically similar constructs; (e) discriminant validity with those constructs with which it is supposed to differ; (f) empirical validity with appropriate outcome constructs such as being significantly related to performance and job satisfaction; and finally, (g) predicts variance in these outcomes (i.e., performance and satisfaction) beyond other similar constructs (in this case Conscientiousness, Extraversion, and core self-evaluation traits).
Study 1 used a 4-item self-rated performance measure (e.g., How would you rate your performance/effectiveness as compared with your peers?). The scale was framed by asking participants to rate their performance in their current job over the past week. If they were not employed then, they were asked to rate their academic performance over the past week. This measure demonstrated adequate reliability (α >.70) and was only used to examine the nomological network of PsyCap and, because it was a self-measure, was not used to test any hypotheses. Study 2, on the other hand, used actual performance evaluations that were gathered independent of the study. Hence, the performance measures were based on objective data and managerial ratings of participants obtained from the human resources department records of the two organizations studied. For the high-tech manufacturing firm, also as prescribed by its appraisal process, each participant’s performance measure included a sum of ratings based on quality and objective quantity of their work on electrical subsystem designs including error and rejection rates, meeting the schedule, complexity of assignment, and ability to work with peers. This measure was then cross-checked by all managers within a given job family to ensure consistency in performance ratings across work units. Although each engineer may be performing a set of slightly different tasks in this appraisal process, all participants had similar job descriptions, performance evaluation criteria, and were considered peers in that they had similar jobs in term of procedures and deliverables. As consistent with the organization’s policy, managers of job families (up to 15 managers) normalized the ratings to settle on a final performance rating. The insurance services firm provided the most recent performance rating for each participant 1 month after they had taken the PsyCap survey. The ratings were based on the most recent month of performance (i.e., after the survey had been administered). These data consisted of input from both objec
556 Luthans et aL. in Personnel Psychology 60 (2007)
tive performance data (e.g., number of claims processed) and their manager’s overall evaluation as prescribed by the firm’s performance appraisal process in one total composite score.
Job Satisfaction Measure
In addition to performance, this study also examined the relationship of PsyCap with job satisfaction. As commonly used in organizational behavior research (e.g., Judge & Bono, 2001), all of our samples but one used a 3-item scale adapted from Hackman and Oldham (1980) using the same 1–6 rating as for the PsyCap measure. This satisfaction scale had high internal reliability, Cronbach alphas (.89, .87, and .86), for the three samples. To meet its concern with the length of the survey, the high-tech manufacturing sample used a one item overall job satisfaction question (“How satisfied are you with your job?”). We also gathered affective organizational commitment (Allen & Meyer, 1996; 1990) data in Sample 1 of Study 1. The purpose of gathering these additional data was to aid in determining the discriminant validity of the PsyCap instrument and to generate a better understanding of the nomological network of constructs for the proposed PsyCap measure. The affective dimension of organizational commitment has been noted for its unique contribution, given it captures the employee’s affective desire to remain with the organization versus a calculative conclusion (Judge & Bono, 2000) and is often used as a single dimension in organizational research (e.g., Bono & Judge, 2003; Judge & Bono, 2000). On the basis of face and content validity, the research team selected four items from Allen and Meyer’s affective commitment scale for this measure. An example item is “I would be quite pleased to spend the rest of my life working for this organization.” These items demonstrated a reliability coefficient of .89. We gathered the job satisfaction data for all samples and affective organizational commitment for Sample 1 of Study 1 one week later than the predictor variables to minimize potential same-source effects/bias. As noted by Podsakoff et al. (2003, p. 887), this temporal separation procedure “makes it impossible for the mindset of the source or rater to bias the observed relationship between the predictor and criterion variable, thus eliminating the effects of consistency motifs, implicit theories, social desirability tendencies” and other individual attributes that may influence or bias the responses. In addition to the temporal strategy of data collection, to confirm the accuracy of the self-reported demographic data, we randomly cross-checked against actual personnel records and found no inconsistencies.
Psychometric Analyses
Using guidelines offered by Schwab (1980) and Pedhazer and Schmelkin (1991), we determined several requisite conditions for the PsyCap measure. These parallel the conditions and guidelines used for determining the core self-evaluations construct of Judge and colleagues (2003). Specifically, the following needed to be established: (a) content validity such that each facet is equally represented in the overall PsyCap instrument, which we established as discussed above in constructing the 24-item questionnaire; (b) sufficient PsyCap scale reliability; (c) PsyCap must have a unitary factor structure consistent with the proposed latent construct; (d) convergent validity with other theoretically similar constructs; (e) discriminant validity with those constructs with which it is supposed to differ; (f) empirical validity with appropriate outcome constructs such as being significantly related to performance and job satisfaction; and finally, (g) predicts variance in these outcomes (i.e., performance and satisfaction) beyond other similar constructs (in this case Conscientiousness, Extraversion, and core self-evaluation traits).
No comments:
Post a Comment
Thank you for your comment. Have a nice day!
Note: Only a member of this blog may post a comment.