Article in Group & Organization Management · January 014 citations reads 13,031 authors: Some of the authors of this publication are also working on these related projects

Ho and Nesbit 
397
insurance industries. Both “supervisor” and “subordinate” questionnaire 
packages included a cover letter clearly explaining the purpose of the 
research and stating that participation was voluntary. All completed ques-
tionnaires were returned directly to the researchers through a self-addressed, 
stamped envelope.
In total, 560 matched surveys were distributed to employees and supervi-
sors. We received 412 completed and usable matching pairs, which repre-
sented an overall response rate of 73%. Deletion of responses with missing 
data (five forms were incomplete) reduced the final usable sample to 407. 
Our final sample, therefore, included employees from a broad cross-section 
of jobs, including technicians, engineers, production foremen, sales and mar-
keting personnel, assembly line supervisors, quality control inspectors, ser-
vice representatives, human resource personnel, middle to senior managers, 
accountants, secretaries, estate officers, clerical personnel, track workers, and 
mechanical operators. For the whole sample, 62% of the employees were 
male, the average age range of employee respondents was 26 to 35 years and 
the average job tenure was 4.7 years (SD = 6.09).
Measures
To use pre-validated measures, the questionnaire items of the following mea-
sures, other than objective work performance, were translated into Chinese 
using Brislin’s (1980) translation/back-translation procedure. Survey items 
were translated into Chinese by the first author, who is bilingual in Chinese 
and English. Next, we obtained a back-translation from another bilingual 
academic. Finally, another academic examined the original version in English 
and the back-translated English version and found no back-translation 
discrepancies.
Self-leadership. In this study, self-leadership was assessed using the modified 
Self-leadership Questionnaire (MSLQ) developed by Ho and Nesbit (2009). 
The MSLQ consists of 38 items describing various behaviors associated with 
self-leadership and participants use a 5-point Likert-type scale (1 = not all 
accurate; 2 = somewhat accurate; 3 = a little accurate; 4 = mostly accurate; 
5 = completely accurate) to indicate how accurately each behavior describes 
them. The dimension of behavior-focused strategy consists of five subscales 
which include Self-Goal Setting (4 items, e.g., “I consciously have goals in 
mind for my work efforts”; α = .79), Task and Relation-Based Self-Observation 
(4 items, e.g., “I usually examine how well I’m doing at work”; α = .70), Self-
Reward (3 items, e.g., “When I have successfully completed a task, I often 
reward myself with something I like”; α = .89), Self-Punishment (4 items, e.g., 
 by guest on July 31, 2014
gom.sagepub.com
Downloaded from 

---

398 
Group & Organization Management 39(4)
“I feel guilty when I perform a task poorly”; α = .80), and Self-Cueing (2 items, 
e.g., “I use written notes to remind myself of what I need to accomplish”; α = 
.81). Natural reward strategy involves two subscales: the first relates to Task-
Based Natural Reward (4 items, e.g., “I think that the enjoyment gained from 
work is more important than external rewards”; α = .76) and the second sub-
scale relates to Relation-Based Natural Reward (3 items, e.g., “I pay attention 
to the enjoyment I gain from working in harmony with my colleagues/team 
members”; α = .68). Constructive thought strategy includes four subscales 
which are as follows: Self-Talk (3 items, e.g., “When I’m in difficult situations 
I will sometimes talk to myself [out loud or in my head] to help me get through 
it”; α = .84), Individual-Oriented Evaluation of Beliefs and Assumptions (5 
items, e.g., “I try to evaluate the consequences of my negative thinking”; α = 
.79), Social-Oriented Evaluation of Beliefs and Assumptions (3 items, e.g., “I 
examine whether my thinking can fit in with the opinions of my colleagues and 
team members”; α = .50), and Visualizing Successful Performance (3 items, 
e.g., “I visualize myself successfully performing a task before I do it”; α = .70). 
Due to poor reliability, we excluded the social-oriented evaluation of beliefs 
and assumptions scale from data analysis.
Prior studies (Houghton, Bonham, Neck, & Singh, 2004; Houghton & 
Neck, 2002) have provided empirical support for the second-order hierarchi-
cal factor structure of self-leadership. In this model, the first-order factors 
consisted of three self-leadership dimensions (behavior-focused, natural 
reward, and constructive thought), which were indicated by the correspond-
ing parcels (composite indicators). Following the item parceling method rec-
ommended by Bogozzi and Edwards (1998), items within each subscale were 
summed and averaged to create five composite indicators for the behavior-
focused dimension, four composite indicators for the constructive thought 
dimension, and two composite indicators for natural reward dimension. The 
second-order factor was the global self-leadership factor, which was indi-
cated by three first-order factors (self-leadership dimensions). Results of our 
confirmatory factor analysis (CFA) supported the second-order hierarchical 
structure of self-leadership (χ
2
= 107, df = 32, p < .01, goodness-of-fit index 
[GFI] = 0.96, incremental fit index [IFI] = 0.94; comparative fit index [CFI] = 
0.94, standardized root mean square residual [SRMR] = 0.05, root mean 
square error of approximation [RMSEA] = 0.07). We concluded that the 
items used in our study measured a single, global construct. We proceeded in 
our analysis using a single scale to represent self-leadership behaviors.