A Development of Ethical Leadership, and the Cause-and-Effect Factors of Ethical Leadership Scales for Students at Thailand National Sports University

บทคัดย่อ

This secondary research developed student ethical leadership and its causes and effects
scales. There were six objectives: 1) To examine the construct validity of the ethical leadership;
and its cause and effect; 2) To examine the discriminatory power of the developed scales; 3 )
To create T-normal criteria for measuring the variables; 4) To examine the slope parameters of
each item and the Threshold values of variables; 5) To examine the different functions of items
on each variable; and 6) To compare the number of validated items using Classical Testing
Theory (CTT) and Item Response Theory (IRT) methods. The secondary data set comprises 1,048
students from the 16 campuses of Thailand National Sports University (TNSU). The research
procedures were as follows: 1 ) Reviewing research documents regarding the meaning
components of ethical leadership, its causes and effects in primary research; 2) Carrying out data
analysis to answer the research objectives, including analyzing the difference in mean scores
using the independent t-test; 3) Checking the structural validity of the variables by confirmatory
factor analysis; 4) Creating measurement criteria by finding normalize T-score; 5) Analyzing
according to the IRT concept by examining the joint slope parameters of the items and its
Threshold values of each answer item in each variable with Multilog program; 6) Checking the
different functions of each item with the SIBTEST program; and 7) Comparing the number of
items that pass the criteria of CTT and IRT using t-test dependent.
The main research results can be summarized as follows: (1 ) All 62 variables are
validated: the index values of all models were not statistically significant. (2) The developed
scales have discriminatory power. The results of testing the difference in the mean scores of the
group of people with high, and low score for each variable show that t-values for each variable
and its items are statistically significant (p value <.01). (3) The normalized T score for all 62
variable scales in the raw score ranged from 2to 20 points. The normalized T score ranged from
2.25 to 65.24. Sample characteristics that be measured can be classified into three groups,
namely, those with low, medium, and high measurement score. (4 ) The common slope
parameters of each item assessment item and the Threshold value of each answer item for all
211 items of 62 variables have β values distributed over a range of 
2
. The slope parameters of
142
the item ranged from 2.17 to 6.86, and the Threshold values of each answer item were
β1<β2<β3<β4. (5) Items in each variable classified according to the background of the sample
(e.g., gender, faculty, year of study, region, etc.) reveal that there were 13 items biased toward
gender, 19 items biased toward faculty, 16 items biased toward year of study, and 7 questions
biased toward region. (6) Questions examined by analysis using CTT and IRT methods indicate
that 27 of 62 variables had the same number of questions. There were 33 variables with no
statistically-significantly difference, and there were only two variables with a distinct difference
in their average number of scales.