The Indonesian Values Scale: an Empirical Assessment of the Short-Form Scale

Culture is an important variable that influences people's behavior. Culture involves several elements, such as language, myth, ritual, custom, artifact, law, and values. However, values are known as the most important elements in describing culture. In 2011, a survey was conducted in Jakarta, Bandung, Semarang, and Surabaya with 2,000 respondents to identify current Indonesian values. The results showed that Indonesian values could be operationalized with 35 items and seven dimensions of mutual assistance: democracy, religion, harmony, hospitality, religious fanaticism, and individualism. On the other hand, the extensive number of Indonesian values items (i.e., 35 items) may have several practical problems, such as longer questionnaires and sample requirements. Therefore, a short-form scale of Indonesian values is needed to enhance the understanding of Indonesian culture through its values. This research aimed to provide a short-form instrument for understanding Indonesian values. Specifically, this research explores psychometric assessments, including the dimensionality, reliability, and validity of the original and short-form scales of Indonesian values. In 2013, a survey with more than 1,000 questionnaires was distributed in Jakarta, Tangerang, Bandung, Semarang, and Surabaya. This research produced a short-form scale of Indonesian values that involves 13 items and the three dimensions of religion, democracy, and harmony. This paper provides an analysis of the data, a discussion of the findings, research limitations, and directions for future research.


Introduction
The topic of culture has interested researchers in diverse disciplines, such as psychology, marketing, consumer behavior, and so on (Horvat et al., 2003). Indeed, the variable of culture is important in understanding human behavior (Durmaz et al., 2011;Matsumoto, 2007;Kacen & Lee, 2002). Specifically, culture influences what we think, feel, and do (Kiyoshi, 2010).
Culture contains seven important elements: language, myth, custom, ritual, material artifact, law, and value (Solomon, 2013). However, the most fundamental determinant of culture is value (Lamb et al., 2009;Schwart, 2006;Yuan & Dong, 2006). Value is defined as a type of belief that guides the selection or evaluation of behavior (Schwartz, 1999). This type of belief is based on mental images that influence and drive people's behaviors (Hemingway, 2005) and acts as guiding mechanism for everyday life (Hitlin & Piliavin, 2004). All people have specific values (Fraj & Martinez, 2006). However, different cultures reflect values differently (Evans et al., 2009). For example, western values are characterized by independence, while eastern values reflect collectivity. In other words, western values emphasize separateness and individuality, whereas eastern values focus on similarity and the need to blend into a society (Blackwell et al., 2007).
Indonesian values reflect the beliefs that affect the way Indonesian people behave in their society. These beliefs are internal references that are used to select and justify culturally appropriate behaviors. Indonesian values are widely accepted and implemented by every member of Indonesian society. Even though there are many tribal cultures in Indonesia, all Indonesians hold dominant core values, such as harmony, tolerance, mutual assistance (gotong royong), and religion (Wirawan & Irawanto, 2007;Sarwono, 1998;Weatherbee, 1966). Values are stable (Miller & Yu, 2003;Meglino & Raylin, 1998), but they may evolve and change continuously (Blackwell et al., 2007;Arnett, 2002) because of globalization (Hawkins & Mothersbaugh, 2010), political change , economic development (Tibbs, 2011), and so on. However, the research on current Indonesian values is sparse. To fill this gap, Sihombing andPongtuluran (2011a, 2011b) developed the Indonesian values scale (INDVALS) to identify current Indonesian values. The construction of the INDVALS is shown in Figure 1. Figure 1 can be explained as follows. The initial stage in the construction of the INDVALS was to identify Indonesian values. The objective of this step was to generate scale items, which were generated by a combined deductive and inductive approach. Exploratory research was conducted, in which respondents were asked to use an open-ended format to describe Indonesian values. In 2011, about 2100 questionnaires were distributed to respondents in Jakarta, Bandung, Semarang, and Surabaya.
The results of the open-ended survey included a list of frequently mentioned values. Two main criteria were applied to select specific personal core values (Schiffman et al., 2010). First, the value must be pervasive. In other words, a significant number of people in a society must accept that value and use it as a guide for their behavior. Second, the value must be enduring. It can be stated that the specific value must have influenced the actions of people in that society for an extended period of time.
This list of frequently mentioned values was checked by experts in the in-depth interviews. The experts were chosen based on their expertise in cultural subjects. They were asked whether frequently mentioned values are part of Indonesian values. Multi-item scales then were developed based on those frequently mentioned values. This generation of items is the most important part of developing the scale (Worthington & Whittaker, 2006). This initial stage produced 162 statements. The content validity of the items was assessed by two experts. Content validity refers to the degree to which an item represents the content or how well the content material was sampled in the measure (Rubio et al., 2003: 94). Furthermore, content validity is also viewed as the minimum psychometric requirement for measurement adequacy in the construct validation of a new measure (Schriesheim et al., 1993, cited in Hinkin, 1995. The items resulting from content validity then were retained and small revisions were made to improve their clarity. The second stage was a further examination of the 162 statements that were distributed to 2,000 respondents. Specifically, from November 2011 to January 2012, the surveys were distributed to respondents in Jakarta, Bandung, Semarang, and Surabaya. The data were then subjected to a scale-reduction analysis consistent with procedures recommended by several researchers (e.g., Verbeke, 2007;Parasuraman et al., 2005;Adcock & Collier, 2001;Churchill, 1979) and reported as suggested by Cabrera-Nguyen (2010 However, the extensive number of INDVALS items (i.e., 35 items) might hinder their application in further research because in a long scale, survey administration and cost are always concerns (Nenkov et al., 2008;Smits & Vorst, 2007;Hibbard et al., 2005).
The original scales contain long items that provide much information, but short scales should enhance several aspects of research. First, a short scale makes survey administration much less burdensome and costly (Smits & Vorst, 2008;Ruvio et al., 2008). Second, when "time is money", the responses to a short questionnaire can be made in a few minutes. Third, a short scale requires less space, which allows researchers to include other variables in the questionnaire (Ruvio et al.,

Mutual Assistance
The importance of cooperation The important thing in a community is mutual assistance

Religion
Religion guides people's behavior. Every person has to have a religion Religion is my faith that God exists. Religion is a foundation of people's behavior.

Democracy
Democracy helps Indonesia become a better nation. Every citizen has the right to choose his or her life Elections are a form of democracy in Indonesia.

Hospitality
I greet people first. I am easy to get close to others. I usually start conversations with others.

Harmony
Living in harmony Harmony prevents disputes. Everyone should maintain harmony. Harmony prevents separation. Mutual assistance is a foundation of social life. The importance of cooperation in a community Every one should help others in a community program. Helping each other is a human social characteristic.  Richins, 2004). Fourth, reducing the length of scales may improve the survey results (Pather & Uys, 2008). Thus, there is a need to develop a short-form, parsimonious INDVALS scale.

Religious Fanaticism
This research aims to provide a short-form instrument to enhance the understanding of Indonesian values. Specifi-cally, this research explores psychometric assessment, including dimensionality, reliability, and validity of the original and short-form scales of Indonesian values. Thus, the research hypotheses regarding psychometrics assessment can be specified (Kohli et al., 1993;Netemeyer et al., 1991) in order to guide the research design of the present study. The research hypotheses

Methods
Sample and sampling design. The survey sample used in this research included students in Jakarta, Tangerang, Bandung, Semarang, and Surabaya. A purposive sampling method was applied, and two major criteria guided the selection of the respondents: 1) the respondent must be a university student; 2) the respondent lives in one of the five research areas (Jakarta, Tangerang, Bandung, Semarang, or Surabaya).
There are five justifications for using students in the sample. First, students ranging from 18-22 years have a solid formation of personal values. In other words, at the age of 10 years, most children have acquired their basic value systems (Hofstede, 1994), which are shaped by families, neighborhoods, and schools (Shuar & Khuntia, 2010;Karahanna et al., 2005). Second, values are stable (de Mooij, 2004;Lombaert, 2003;Meglino & Ravlin, 1998;Hofstede, 1994). Therefore, the values possessed by students will remain with them until they become old. Third, university students have similar major demographic characteristics (Lee, 2000). Fourth, the student sample is a generally accepted method in crosscultural research (Lee, 2000) and scale development research (e.g., Schwartz et al., 2012;Yoo et al., 2011;Kim et al., 1999;Netemeyer et al., 1991). Furthermore, student samples are widely applied as a surrogate for future managers (Hughes & Gibson, 1991), consumers (Klaus, 2001), and others. Finally, student samples are easy to access (Bond, 1988).
Instrument. This research applied the original INDVALS (35 items), which was developed by Sihombing (2012). All items in the questionnaire were measured on a five-point Likert scale anchored from "strongly disagree" (1) to "strongly agree" (5).

Data analysis and scale reduction.
The data were subjected to scale-reduction analyses consistent with the procedures used in developing scales (e.g., Verbeke, 2007;Parasuraman et al., 2005;Adcock & Collier, 2001;Churchill, 1979). Following Bearden and Netemeyer (1999), this research applied the fundamental criteria of good measures (i.e. reliability, dimensionality, and validity). The reliability analysis was conducted first. The examination of coefficient alpha and item-to-total correlation was conducted to delete items in order to improve the reliability coefficient alpha.
The next step was exploratory factor analysis (EFA). Factor analysis is used as a data reduction technique  (Hair et al., 2006). Specifically, EFA was applied to determine the number of underlying dimensions in an item of data (Hair et al., 2006). Following to Hair et al. (2006), a confirmatory factor analysis (CFA) was conducted on all items.

Realibility.
A measurement is said to be reliable when the measure is error free and offers similar results over time (Sekaran & Bougie, 2010). Realibility is a necessary pre-condition for achieving valid measurements (Nunally, 1978). This research applied the Cronbach alpha, composite reliability, and average variance extracted (Baumgartner & Homburg, 1996;Hulland et al., 1996).
Dimensionality. Unidimensionality exists when each item reflects only one underlying construct (Anderson et al., 1987). Dimensionality was assessed through confirmatory factor analysis (Anderson et al., 1987;Gerbing & Anderson, 1988) by applying maximum likelihood and varimax rotation. Assessing reliability can also serve as indicator of unidimensionality (Gerbing & Anderson, 1998).
Validity. Construct validity shows that a measured construct is significantly related to another construct to which it should theoretically be related (Sekaran & Bougie, 2009). The key aspect of validity that guides this research is construct validity, which is the extent to which a set of measured items actually reflects the theoretical latent construct those items are designed to measure (Hair et al., 2006: 776). Furthermore, construct validity is a necessary condition for theory development and testing (MacKenzie & Podsakoff, 2003). Four measures are used to assess construct validity (Hair et al., 2006: 1) standardized loading estimates should be 0.5 or higher; 2) average variance extracted should be 0.5 or greater; 3) average variance for two factors should be greater than the square of the correlation between the two factors; and 4) construct reliability should be 0.7 or higher. The assessment of construct validity in this study is done through measuring convergent, discriminant, and nomological validity.
Convergent validity is established in the confirmatory factor analysis when the factor loadings are significant (i.e. Critical Ratios>1.96, p<0.05), which shows the overall fit of the model is acceptable (Steenkamp & Baumgartner, 1992). On the other hand, discriminant validity is assessed in two ways. First, discriminant validity was achieved when the correlation between constructs was significantly less than 0.75 (Zikmund et al., 2010). Second, the value of average variance extracted more than 0.5 indicates that discriminant validity was achieved (Fornell & Larcker, 1981). Finally, nomological validity can be assessed by determining which predictions from key constructs are consistent with the theory (Bagozzi, 1980).

Results and Discussion
Sample profile and response rate. Of the 1,000 questionnaires distributed, a total of 778 questionnaires were collected. After excluding 32 questionnaires that were incomplete (missing data), 746 usable questionnaires were retained for the data analysis, yielding a usable response rate of 74.6%. Table 2 shows the sample profile. The majority of the respondents were females (52.4%) and the respondents were mainly between the ages of 18 and 20 years (70%). Almost half of the respondents (44.2%) were in the sample surveyed in 2011 (Table. 2).

Scale-reduction.
The 35 items on the Indonesian value scale were subjected to scale reductions. Reliability analysis was first conducted by grouping the items according to the a priori dimensions from which they were derived. The next step was to assess the corrected item-to-total correlations. Items below 0.3 were then deleted to improve the coefficient alpha. Reliability is a necessary contributor to validity, but it is not a sufficient condition for validity (Cooper & Schindler, 2011: 283).
Thus, the next step was the assessment of validity, which was done by conducting EFA and CFA. Specifically, EFA was conducted to purify the scales, which were then assessed to CFA (Cabrera-Nguyen, 2010; Amyx et al., 2008). Table 3 shows the results of the EFA, which provides evidence of the load of personal value items on each factor and their respective factor loading. The factor analysis generated seven factors, which explained a total variance of 63.349%. Furthermore, the overall Kaiser-Meyer-Olkin (KMO) measure of sample adequacy had a highly acceptable value of 0.896 and a Bartlett Test of Sphericity Value of 3974.565 (p=0.000). Descriptive statistics. The mean and the standard deviations are the most common descriptive statistics for interval data (Sekaran & Bougie, 2010). The mean reflects the central tendency of the data, whereas standard deviation shows the spread of the data distribution (Table 4).

Dimensionality.
The results showed that the INDVALS has a unidimensional factor. CFA was applied to assess the unidimensionality of the INDVALS. Table 5 shows several indices of the goodness of fit of the CFA model.

Reliability.
The internal consistency of the scales was assessed using Cronbach Alpha, composite reliability, and average variance extracted to examine the reliability of the INDVALS. Table 5 shows that the internal consistency of the INDVALS for both original and short versions was quite high, which is evidence of good levels of internal consistency.

Validity.
A confirmatory factor analysis model was performed to examine convergent validity. Table 6 shows that all factor loadings were significant (Critical Ratios>1.96, p<0.05), indicating the presence of convergent validity. All fit indices indicated a good fit for both original and short versions of the INDVALS.
This research applied the Pearson correlation and average variance extracted to assess discriminant validity. In other words, discriminant validity was examine by two ways: (1) the correlation between constructs should significantly less than 0.75 (Zikmund et al., 2010), and (2) average variance extracted should be greater than 0.5 (Butler et al., 2012;Fornell & Larker, 1981). Table 7 shows that there no correlation between constructs more than 0.667. Furthermore, except for democracy, average variance extracted for each construct was higher than 0.5 indicating discriminant validity. Regarding nomological validity, the results showed that the INDVALS was positively correlated with attitude toward knowledge sharing, subjective norms, perceived behavioral control, intention, and knowledge sharing (Table 7).
In short, the correlations were in the predicted directions, and most were significant. Therefore, the findings provided evidence for the nomological validity of the INDVALS.
The results of this research will contribute to the scant literature on current Indonesian values. Specifically, the results showed that the short-form of INDVALS, which includes democracy, religion, and harmony, represents core Indonesian values. Understanding values is important because they play an important role in determining people's behaviors (McCarthy & Shrum, 2000).   (Dewi & Yulika, 2014). People in Indonesia have practiced tolerance towards people with different religious backgrounds for many decades. However, Indonesia has also experienced religious intolerance, such as conflicts in Ambon and Poso and other parts of the country. Therefore, religious values that emphasize "unity in diversity" can be one way to reduce intolerance towards other religions (Ma'arif, 2006).
In addition to religious tolerance, harmony has also been a core value for many decades. Harmony refers to maintaining relationships with others. Avoiding conflicts with others and being oriented to others are values that most Indonesian people practise in their daily social interactions. Furthermore, Indonesians practise the principles of musyawarah mufakat (consensus and compromise) in decision making and gotong royong (mutual assistance) in helping others in their daily lives.
This crisis further caused unstability in political conditions in the country. Student and public demonstrations were aimed to force President Suharto's resignation. Since then, democracy has become a popular issue discussed in society through the freedom of speech. The ultimate result of the Indonesian crisis in 1998 was the stepping down of President Suharto from the position that he upheld for thirty years. Thus, it can be stated that democracy has now become an Indonesian value.