Development and Validation of a Scale for Christian Character Assessment of University Students

After the Character Education Act was enacted, providing character classes became a prerequisite in some Korean colleges starting in spring 2017. Keeping in step with these changes, experts researching character education cited the need for the development of character scales. The purpose of the present study is to explore the validity and reliability of the Christian character scale for university students. The participants were 994 students attending a Christian university in Seoul. The original 75 questions were developed from 13 factors, within three domains, from the three Biblical references and Christian universities’ virtuous character traits. These questions were reduced to 39, three per each factor, through the examination of content validity by a validation panel. Twenty-nine questions were finalized through exploratory factor analysis and confirmatory factor analysis within five factors: relationships with others I—loving and caring, relationship with God, relationships with others II—loving and peacemaking, personal responsibility, and social responsibility.


Introduction
The tragedy of the Sewol Ferry Incident, with the loss of 304 lives-mostly high school students on a school trip-symbolizes the spirit of the age.It bears witness to the erosion of traditional values and a crisis of character.Faced with societal problems such as school violence, teenage suicide, and violent crime along with the Sewol Ferry Incident, Korea has sought a remedy to these difficulties through character development by enacting the Character Education Promotion Act in July 2015 (Lee 2015).According to this law, the objective of character education includes the attitude of both mind and personality, such as good manners, honesty, responsibility, respect, consideration, communication, cooperation, and others.State or local governments should establish and execute long-term and systematic policies to nurture citizens skilled in the best expressions of character (Korea Ministry of Government Legislation 2017).At the same time, the Korean Ministry of Education designated character courses as prerequisite classes at universities and colleges of education starting in 2017.These actions reflect the idea that before beginning subject-specific course work, there should be foundational development of one's humane character-the basis of all education (Nam 2015).Beyond these factors, students have become aware that university programs lack character education, and have expressed the need for such coursework (Cho 2016).In this regard, it is important for those responsible for curriculum development to identify which character traits should be nurtured through education to insure, as much as possible, that this process enjoys universal support and cooperation (Yoo et al. 2012).
(Calvin College 2016) and Wheaton College (Wheaton College Center for Vocation and Career 2015), and the Christian Character Index (Assess-Yourself.org2016).The character virtues disclosed from these are summarized as follows in Table 1.The character traits described above have been categorized into intrapersonal values, interpersonal values, and social values, as proposed by Park and Hue (Park and Huh 2012).Where all of the intrapersonal, interpersonal, and social values are applicable, such as diligence and responsibility, they were placed in the most appropriate categories as discussed by the panel of experts.From the character traits listed above, the category of intrapersonal values includes patience/hope, self-control/purity, faith/loyalty, faithfulness/honesty, diligence/responsibility, and joy/gratitude.The category of interpersonal values lists compassion, love, forgiveness/gentleness/peace, kindness/consideration/manners, and generosity/benevolence.The category of social values consists of justice and community spirit.

Item Development and Content Validity
Based on a Biblical and theoretical foundation, an item pool of 75 questions was developed.Seven professors who major in Christian ethics, Christian philosophy, or Christian education, or who were instrument development specialists, were invited to serve as a panel of experts to evaluate the developed item pool.They ranked each question by a 5-point-Likert scale for relevance and clarity (Gable 1993).They were also asked to check whether the questions fit into individual, interpersonal, and social values properly, and to add supplementary questions that the researcher failed to include.Based on the relevance and clarity scores, 39 questions were selected of 75.

Pilot Study
Following Patten's recommendation that 25 or more respondents who represent the target population should be invited to a pilot study (Pattern 2001), this study was conducted with 37 students.Since subjects showed difficulty understanding Item 17, "I am honest, even when there is an advantage to lie," was modified to "I do not tell a lie, even when there is an advantage to lie."The questions were formatted with a 5-point Likert scale.Internal consistency using the Cronbach's alpha coefficients of the instrument was r = 0.79.

Research Subjects
The data were collected from 994 Christian university students who attended C Christian university and self-evaluated their character.This university, which advocates Reformed theology, has 1500 students with nine majors including theology and Christian education.The sample was 331 male (33.3%) and 663 female students (66.7%); by school year, there were 290 freshmen (29.2%), 254 sophomores (25.6%), 256 juniors (25.8%), and 194 seniors (19.5%).

Data Analysis
In order to perform this study, the SPSS 22 program was conducted to determine reliability and validity through exploratory factor analysis.The Amos 20 was used to confirm the model fit through confirmatory factor analysis.

Data Screening
To assess normality, the distribution of the data was examined with respect to skewness and kurtosis.The values between −2 and +2 for skewness and kurtosis were considered acceptable for proving normal univariate distribution (Field 2009).Skewness ranged from −0.91 to 0.03 and kurtosis ranged from −0.70 to 0.17.Therefore, no indicators that appeared to have a problematic distribution in statistical analyses were found.

Exploratory Factor Analysis
Exploratory factor analysis was used to simplify the set of data and explore the possible structure of the scale.A principal component analysis using varimax rotation was performed.In this study, the Kaiser-Meyer-Olkin Measure of Sampling (KMO) value (0.968) and Bartlett's test value (χ 2 (741) = 18,447.27,p = 0.000) indicate that factor analysis is appropriate (Pallant 2010).
The five factors that had eigenvalues greater than ones that were extracted, and 52.99% variance was explained.Rotated factor loadings, variances, and eigenvalues are provided in Table 2.The first factor, entitled "relationships with others I-loving and caring," included 11 items, representing 12.58%.The second factor, entitled "relationship with God," included 9 items, and accounted for 10.94%.The third factor, labeled "relationships with others II-loving and peacemaking," included 6 items, accounting for 10.94%.The fourth factor entitled "personal responsibility," included 7 items, representing 10.02%.The fifth factor, named "social responsibility," included 6 items, and explained 9.32%.

Model Fit
An initial confirmatory factor analysis was conducted to assess 39 items, within five factors, based on exploratory factor analysis results.However, the original model did not provide an acceptable fit to the data (χ 2 (692) = 3020.48,p = 0.000; SRMR = 0.056; CFI = 0.870; TLI = 0.861; RMSEA = 0.058 (low 0.056; high 0.060)).For modification, the items in which the loadings of each factor are less than 0.50 were eliminated.Modification indices were examined, and some of the reasonable modifications were incorporated through correlated residuals and cross-loadings (Hooper et al. 2008).With these changes, the model produced a reasonably good fit.The first modified model was comprised of 29 items within five factors.This model was found to fit the data marginally (χ 2 (367) = 1663.95,p = 0.000; SRMR = 0.042; CFI = 0.906; TLI = 0.896; RMSEA = 0.059 (low 0.057; high 0.063)).With the exception of Item 6, "I give thanks in all circumstances," all construct measures loaded significantly onto their respective factors.The second modified model included 28 items excluding Item 6 pertaining to five factors.This revised model provided a highly satisfactory fit to the data (χ 2 (340) = 1529.01,p = 0.000; SRMR = 0.042; CFI = 0.911; TLI = 0.901; RMSEA = 0.059 (low 0.056; high 0.062)).Table 3 presents the model fit indices for the original and revised models.Although the third model showed a better model fit, this researcher prefers the second model that includes Item 6, "I give thanks in all circumstances," which showed slightly higher loading scores across four factors.The reason is that "thankfulness" was one of the factors originally included, and without thankfulness, Christian character would not be measured properly.As a final point, it must be noted that, while fit indices are a useful guide, a structural model should also be examined with respect to substantive theory (Hooper et al. 2008).Therefore, the scale was finalized to 29 items including Item 6, even this model is less satisfactory.

Convergent & Discriminant Validity
Confirmatory factory analysis was used to assess the measurement model with respect to convergent and discriminant validity, as well as fit.Large standardized factor loadings, which indicated large inter-correlations among items associated with the same latent variable, support convergent validity (Kline 2011).The averaged variance extracted (AVE) and construct reliability
The correlations between factors through CFA are in Table 5.In the case of discriminant validity, correlations are less than 0.76 and square roots of AVE are greater than inter-construct correlations (Cho 2015).Therefore, discriminant validity was proved.

Discussion and Conclusions
In the recently enacted Character Education Promotion Act, character education is defined as "education with an objective of nurturing human characters and abilities that are necessary for raising one's inner self in an upright and sound manner and living in harmony with others, community, and nature" (Korea Ministry of Government Legislation 2017).However, Christian character education is to educate people to maintain, in a consistent and constant manner, lifestyles and behaviors that are in accordance with the inherent personal moral character traits, given by God, and the characteristics of a person created in the image of God (Han 2014).Not only expressed behaviors but also hidden motives need to be counted.For ancient philosophers such as Plato and Aristotle, virtues were connected to purposes.Just as a knife for cutting and a horse for transportation, so goodness is based on the goal Christians wish to achieve (Gill 2000).Therefore, lifestyles, behaviors, and motivations are important to character formation.In this sense, Christian character education should not merely be ethicalized character education.A relationship with God should take precedence as the basis of Christian character education (Kang 2015).
Therefore, this study developed a scale that measures the university student's character traits based on a Biblical standard, which are beyond the morality, conscience, respect for others, civic awareness, self-regard, and taking the initiative that are emphasized in general character education.Calvin College sets an example of the biblical criterion for character education, explaining its core virtues: "As God's chosen people, we are enjoined by St. Paul to clothe ourselves with 'compassion, kindness, humility, meekness, and patience,' but above all to clothe ourselves 'with love, which binds everything together in perfect harmony' (Colossians 3:12, 14)" (Calvin College 2016).The Biblical references cited in this study are related to the fruits of character that are evident in the lives of people who have undergone a second birth.This fruit is recognized as the core of Christian character education (Kang 2015).This study commenced by identifying 13 virtues of Christian character commonly emphasized in Bible texts and at Christian universities-patience/hope, self-control/purity, faith/loyalty, faithfulness/honesty, diligence/responsibility, joy/gratitude, compassion, love, forgiveness/gentleness/peace, kindness/consideration/manners, generosity/benevolence, justice, and community spirit-and then categorizing them by intrapersonal values, interpersonal values, and social values.From these, 75 questions were formed and further reduced to 39 questions after undergoing the content validity test by a panel of experts consisting of seven people.Ultimately, they were finalized into 29 questions related to five factors through an exploratory factor analysis and a confirmatory factor analysis.It can be seen that the five factors named in this study include the original three values of intrapersonal, interpersonal, and social values.These also share factors with the four areas of reconciliation explained by Harbermas and Issler-communion, character, community, and commission (Habermas and Issler 1992).
It is meaningful in that it has developed a valid and reliable scale for measuring the Christian character of university students through a statistically necessary procedure.Taken together, the present study provides the first stage for measuring the Biblical character of emerging adults.This study, however, has a limitation in that the 994 research subjects were collected from one Christian university only.The fact that this scale measures the student perception of character, not their actual character practice, might be another limitation.The Narnian Virtues curriculum was performed to test the understanding and application of virtues.One hundred sixty students, aged 9-14 years, participated in this experiment.A statistically significant increase in their ability to correctly identify 12 virtues (knowledge) was found.No statistically significant change was found in assessing their application of 12 virtues or in assessing personality and attitude characteristics (Pike et al. 2015).These results show that knowledge is different from the application of knowledge or the attitude reflecting the knowledge.This scale also measures the students' perceptions of their character and does not actually measure their character.One's actual character might be different from their perception of it.
To maximize the utility of the Christian character scale for university students, additional studies of its use are needed.First, an analysis of the factors affecting character development is needed.If the factors that influence emerging adults to be Christ-like and mature are known, educators can equip them more effectively.Second, as the Character Education Promotion Act suggests that Christian character programs need to be developed to evaluate their effect on Christian character formation.Experimental studies regarding Christian character programs might contribute to Christian education by developing new programs and proving their effectiveness.The author hopes that this instrument might be a useful tool to measure biblical character for other Christian university students or Christian students attending university.It could be used in other countries after culturally inappropriate items are revised.

Figure 1 .
Figure 1.The path diagram of the final five-factor model.

Figure 1 .
Figure 1.The path diagram of the final five-factor model.

Table 1 .
Bases of Christian character virtues.
(Lewis 2015; Pike et al. 2015; The Church of England Education Office 2015) were used as reference resources, but were not used in factor selection.

Table 3 .
Goodness-of-fit indices from confirmatory factor analysis.

Table 4 .
Construct reliability and averaged variance extracted results.