Skip to main content
Open AccessMultistudy Report

Development and Validation of a Work Values Scale for Assessing High School Students

A Mixed Methods Approach

Published Online:https://doi.org/10.1027/1015-5759/a000408

Abstract

Abstract. Assessing work values with high school students is a critical component of career counseling practice, however it remains a relatively understudied area of research. The purpose of this study was to develop and provide psychometric evaluation of a Work Values Assembly (WVA) scale for assessing high school students. This study employed a mixed methodology to gather research data and conduct data analyses. In the first study, 30 participants were involved in focus-group interviews about their work values. The interview data were analyzed through a grounded theory approach and a framework of seven-dimension work values was derived. In the second study, the WVA scale was constructed based on the descriptions and dimensions of the first study. Seven hundred fifty three high school students participated in the pilot study. The revised scale was then administered to 896 high school students in a formal test. The exploratory- and confirmatory factor analyses re-verified the quality of the items and the construct validity of the WVA scale. The scale also demonstrated good test-retest reliability and criterion-related validity. Finally, 896 and 592 participants from high schools and colleges, respectively, participated in a test of measurement invariance between the two groups. Implications for counseling as well as suggestions for future research were discussed.

The Meaning and Importance of Work Values

The concept of work values is derived from the concept of values; values are all of a person’s long-term preferences and enduring convictions that are used to gauge the importance and agreeableness of an experience, or those that are used to incite a person to action or decision (Rokeach, 1973). Thus work values are those lasting values that act as standards in the field of work, and are therefore used to judge work-related phenomenon, behaviors, and goals, as well as provide the basis for the assessment of an individual’s career choices (Super, 1970).

The implications of work values are quite extensive with many scholars considering them able to explain, predict, and even become the basis or standard for career decisions. For example, researchers (Rokeach, 1973; Sortheix, Chow, & Salmela-Aro, 2015) consider work values capable of providing for a relatively stable psychological system, and, as a result, able to explain and even predict an individual’s internalized traits and behaviors in a variety of work environments. Schwartz (1992) considers work values to be a barometer for objective appraisal and behavioral guidance. Researchers (Judge & Bretz, 1992; Maynard & Parfyonova, 2013) also view work values and work-related objectives as the internalized standards by which an individual selects a vocation or evaluates a job. Work values, as seen in this light, do not merely aid in the selection of an occupation or the appraisal of a job, but also reveal a worker’s expectations of the ideal career, becoming the basis for the explanation as to why we work (Dose, 1997; Jambrak, Deane, & Williams, 2014; Wöhrmann, Fasbender, & Deller, 2016).

Super (1970) and Gerpott (1988) both asserted that work values tests also have implications for applicable aspects of career guidance. Their importance, like similar aptitude and interest tests, lies in their usefulness in the understanding of internalized criteria for on-the-job expectations of satisfaction and cognitive preferences toward the content of the work, becoming the referential basis for career planning. In Asia and Europe, streaming is a very important education system (Sung, Chao, & Tseng, 2016; Sung, Huang, Tseng, & Chang, 2014; Sung, Cheng, & Hsueh, 2017; Trautwein, Lüdtke, Marsh, Köller, & Baumert, 2006). High school especially is an important time for the designation of a career path (Sung & Chao, 2015; Sung, Cheng, & Wu, 2016; Super, 1980); it is the time when students, according to their abilities, are separated into academic or vocational tracks (Betts, 2011; Sung et al., 2014), and as a result, their future university majors or other vocational directions are determined. Students, at this time, find themselves in the midst of the formative period for work values, where they must, based on the ebb and flow of societal experiences, come up with their own preferences, capabilities, and values, thereby committing themselves to a future career (Super, 1980). Therefore, if information on work values and interests, prior to career designation, can be amalgamated and applied to the field of career counseling, then besides providing high school students with more opportunities to explore suitable career options via links between various assessments and relevant occupations, this counseling will also be able to enhance student self-awareness, avoiding the development of negative attitudes toward education due to the selection of an inappropriate academic discipline (Dawis, 1991; Lent & Brown, 2006).

The Dimensions and Measurement of Work Values

Dimensions of Work Values

Ginzberg, Ginsburg, Axelrad, and Herma (1951) formulated one of the earlier classifications of work values, proposing three types based on what kind of reward they are associated with: intrinsic, extrinsic, and concomitant (as cited in Ros, Schwartz, & Surkiss, 1999). Intrinsic rewards can be achieved by doing the job itself, which in itself can motivate an individual to work, such as autonomy (Super, 1970). Extrinsic rewards are the things of value brought about as a result of doing the work, such as financial rewards and convenience (Super, 1970). Concomitant rewards are largely related to social values, examples of which are social relationships with coworkers or contributions to society (Zytowski, 1994). Rokeach (1973) classified values into two major types: terminal and instrumental. Terminal values emphasize the goals the individual hopes to achieve in his or her lifetime, such as personal values and social values, while instrumental values focus on the beliefs, practical actions, and modes of behavior for achieving those goals, such as moral and behavioral values.

Work Values Assessments and Their Limitations

Recognizing the importance of work values, various tools to assess work values have been developed. Table 1 provides brief descriptions of several of the measurement tools that are used most frequently by researchers.

Table 1 Summary of work values scales

A commonly used work values assessment is the Work Values Inventory (WVI), developed in 1957 by Super and Crites. WVI included 15 types of work values and 45 items, including altruism, esthetics, creativity, intellectual stimulation, achievement, independence, prestige, management, economic returns, security, surroundings, supervisory relations, associates, way of life, and variety. Later, WVI measurements were expanded through extensive research, such as the Work Importance Study (WIS; Super & Sverko, 1995) and Super’s Work Values Inventory-Revised (SWVI-R; Zytowski, 2006); both have modified the dimensions and content of WVI.

Rounds and Armstrong (2005) recommended adopting the experience gained from tools used for job adjustments as a means to supplement career counseling so as to extend it beyond the campus and into the work site. Based on the Theory of Work Adjustment, the purpose of the Minnesota Importance Questionnaire (MIQ; Paired Comparison Version; Rounds, Henly, Dawis, Lofquist, & Weiss 1981) was to estimate the impact of the work environment on the worker, focusing on the needs and values that the job role brings to the worker and whether or not the work environment meets the worker’s expectations and satisfaction. The MIQ comprises 20 separate work needs items related to work requirement descriptions differentiated through factor analysis into six work values: achievement, comfort, status, altruism, safety, and autonomy. Based on the MIQ, McCloy et al. (1999a) developed the Work Importance Profile (WIP) for the American career consulting agency Occupational Information Network (O*NET). WIP included similar 6 value dimensions and 21 needs.

Based on Table 1, we can also find some limitations of those existent tools. Firstly, in the examination of reliability and validity, although the majority of scales conform with Nunnally’s (1978) criterion that reliability coefficients be greater than .70, most studies did not demonstrate appropriate validity. Scholars (Cabrera, Nora, & Castaneda, 1993; Long, 1983) have proposed that exploratory factor analysis (EFA) alone cannot reveal the complete structure of the factors, and so suggested that confirmatory factor analysis (CFA) should be used to confirm the existence of factor structure and goodness-of-fit indices. However, CFA has been applied to only a few work values scales to analyze their construct validity, and even if researchers employed CFA, the studies suffered from the limitations of small samples and poor model fit. For example, the WIP (McCloy et al., 1999a) included CFA in its analyses, the value for root-mean-square error of approximation (RMSEA) was .11, the comparative fit index (CFI) was .8, and TLI was .76, thus most of the indicators failed to meet the criteria proposed by Hu and Bentler (1999).

Secondly, cultural issues need to be emphasized more in the construction of work value scales. Many researchers have asserted that work values scales are influenced by cultural factors. For example, Robinson and Betz (2008) using SWVI-R, and Hartung, Fouad, Leong, and Hardin (2010) using the WVI scale both found that African Americans placed more importance on intrinsic work values than did Americans of European descent. Fuller, Edwards, Sermsri, and Vorakitphokatorn (1993) used theoretical constructs and measurement tools developed by American academia in their cross-cultural research that surveyed local Thai sociological phenomenon. The research results indicated that, despite back-translation, many of the statements in the English questionnaire were too abstract and confusing and needed to be reworded.

Furthermore, comparisons of foreign versus domestic research also revealed that differences in environments and samples could lead to the extraction of different factor constructs between various studies even though the scales adopted were the same. Using a WVI scale as an example, White (2006) extracted four factors from a sampling of college students in an English-speaking country, namely, comfort-independent, stimulation, affiliation, and achievement. Dolan, Díez-Piñol, Fernández-Alles, Martín-Prius, and Martínez-Fierro (2004), on the other hand, extracted four factors of extrinsic work values, social, economic work status, intrinsic work values, and self-realization values based on their Spanish participants. Moreover, Lee, Hung, and Ling (2012) in a sampling of Malaysian teachers classified three different factors, that is, environment, intrinsic, and security. The above research findings reveal that using the same WVI scale for different cultures will result in different factor structures; different ethnic and cultural backgrounds must therefore be considered when constructing a values scale, but in most cases the work value dimensions of respondents from different cultural backgrounds are not taken into account when values scales are constructed.

Thirdly, although research indicates an individual’s work values change greatly during the transition from adolescence to adulthood (Johnson, 2002), current theory suggests that the structure of work values in a culture does not change due to age (Harpaz & Fu, 2002; Jin & Rounds, 2012). However, when formulating a work values scale using one age group, this theory must still be validated for that particular scale if researchers intend to apply it to a different age group. Unfortunately, most current work values scales were developed surveying working adults, and never rigorously validated with a high school sample. For example, the widely used work values scale of Super (1970) was constructed on the basis of the researcher’s work with middle school students; whether or not the scale is suitable for use with high school students still awaits validation.

Because of different backgrounds or mental sets, different groups of participants may conceptualize constructs within the same measurement tool differently. Therefore, directly comparing the testing results of two groups with the same measurement tools may not be appropriate (Raju, Laffitte, & Byrne, 2002). Previous research usually focused on the evaluation of the reliability and validity of a measurement tool through a single group of participants with little consideration for the appropriateness of using the same tool for another group.

Thankfully, there exist statistical tools for testing the generalizability of a measurement method. Measurement invariance has been proposed as a way to evaluate the plausibility of applying testing tools to different groups while measuring the same construct (Cheung & Rensvold, 2002). The testing of measurement invariance is a kind of multigroup confirmatory factor analysis (MGCFA), which can evaluate if there exists invariance or equivalence in the estimated parameters of factor models in different groups (Cheung & Rensvold, 2002). By testing the configural invariance (model 1; same factor structures across groups), metric invariance (model 2; same factor loadings across groups), and scalar invariance (model 3; same item intercepts across groups), researchers can get more objective information about the appropriateness of comparing the test results of different groups with the same measurement tool (e.g., Marsh, Nagengast, & Morin, 2013). Based on the importance of evaluating the measurement invariance, this study will test the measurement invariance of a Work Values Assembly (WVA) scale among high school and college students.

Aim of the Current Research

In Taiwan the most frequently used work value scales were compiled some years ago, with most of the scales having been based on scales developed in the Western societies, such as Wu, Li, Liu, and Ou (1996) and Chen, Wang, Liu, Ou, and Li (1987). Those scales suffered from the three limitations mentioned above: They have a lack of rigorous evaluation for construct validity, they directly adopted the structure/dimension of values from other countries with little consideration for indigenous culture, and they need more evaluation for the appropriateness of application to cross-age groups of participants (Hung & Liu, 2003). To help compensate for these limitations, our research has four aims:

  • (a)
    Determine the work values dimensions of Taiwanese students and compare those with Western cultural values by means of qualitative research.
  • (b)
    Develop a WVA scale suitable for use with high school students based on the work values dimensions obtained by qualitative research.
  • (c)
    Empirically examine the psychometric properties of the WVA scale, including its reliability and validity with representative samples and abundant participants.
  • (d)
    Evaluate the measurement invariance in the WVA scale’s structure between groups of high school and college students.

We employed mixed methodology (Tashakkori & Teddlie, 2003) with an exploratory design of the instrument development model (Creswell, Plano Clark, Gutmann, & Hanson, 2003) to fulfill the above purposes. The first study employed a grounded theory approach to collect the qualitative data on the work value dimensions. The second study evaluated various facets of measurement theory, including reliability of internal consistency, validity from EFA, CFA, and criterion-relatedness, and measurement invariance for the WVA scale across high school and university students. During the construction and evaluation process of the WVA scale, the principles of the Standards for Educational and Psychological (AERA/APA/NCME, 2014) were referred to as a major guide. For example, the content and relative importance of aspects of the content in the WVA scale were carefully derived from the interview results of focus groups, to ensure the representativeness of the work value content and dimensions in Taiwanese culture; the validity evidence based on internal structure was gathered and evaluated by EFA and CFA; the validity evidence based on external criteria was evaluated by comparing the WVA scores with the scores of another validated work value scale, the Work Importance Locator (WIL; McCloy et al., 1999b); and for the evidence of generalizability for the WVA scale across different groups, a measurement-invariance analysis was conducted for both data sets from high school and college students.

Study 1: Qualitative Research on the Content of Taiwanese Work Values

In this study, a qualitative dual-approach (inductive and deductive) research methodology was implemented. The focus-group interview is often used for the process of identifying dimensions and compiling statements in a scale or questionnaire (Krueger & Casey, 2000). The unique characteristic of focus-group interviews is their ability, through group dynamics, to allow members to express, in a relaxed setting, a diversity of opinions on all types of experiences, perspectives, and views related to the research topic, and, thereby, assist the researcher to collect information-rich qualitative data and to examine the topic from a totally new point of view (Morgan & Krueger, 1993).

Method

Participants

The main purposes of this qualitative study were to elucidate the dimensions of work values in contemporary Taiwanese culture and to establish a structural framework for the WVA scale. According to previous researchers, while the emphasis people place on different work values dimensions may change substantially from adolescence to adulthood (Johnson, 2002), the actual structure of those dimensions should remain stable over time (Harpaz & Fu, 2002; Jin & Rounds, 2012). Therefore, we believe that the structure/dimensions derived from Taiwanese adults may also represent the work values structure of Taiwanese senior high school students. Furthermore, because each participant of this study needed to attend the focus group at least twice, and because Taiwanese high school students are so intensely busy studying for entrance examinations, which are very high stakes for their future career development (Sung & Chao, 2015), recruiting senior high school students for investigating work value structure/dimensions would be impractical for this study. Based on the two reasons above, this study recruited college-aged adults as our focus-group sample while using senior high school students as the sample for evaluating the reliability and validity of the WVA scale developed based on the dimensions provided by the college students. However, conscious of the risks of generalizing across age groups, we additionally tested the scale measurement invariance between these two groups in Study 2.

The study followed Krueger’s (1988) suggestions on the use of purposive sampling for focus-group research. First, a pool of candidates was selected that met the objectives (identifying work value dimensions) and conditions (college students or graduates aged from 20 to 30 years with part- or full-time work experience) of the study, and then actual participants were randomly chosen from this pool; so that, in this way, sampling errors could be reduced to a minimum. The study utilized online advertisements with web-based registration, and personal contact via campus bulletins to solicit participants. The participants were recruited based on their gender and work experiences; participants with different work fields were given priority. In the end we recruited 30 Taiwanese nationals split evenly between men and women and divided into five separate groups in such a fashion as to maximize the diversity of each group.

The participants included four students from technology-related university departments, seven students from liberal arts and business departments, seven individuals with more than 2 years of working experience in technology-related fields, four individuals with liberal arts or business backgrounds who were working in related areas, and eight individuals working in fields unrelated to their university majors. Each interview lasted 1.5 hr and was recorded with the respondents’ permission. Due to the length of each session and the desire to obtain high-quality data, each group met 2 or 3 times to ensure each participant could fully express themselves without time constraints.

Procedure

Data Collection

Following Krueger and Casey (2000), each discussion group consisted of six members, as too many in a group would hinder opportunities for experiential and observational exchanges between group members. At the beginning of the first focus-group session, the researcher first gave a brief self-introduction and explained the purpose of the research. If members were still willing to participate in the focus group they then filled out a consent form that included basic data. The researcher also verbally notified the participants that the entire session would be taped and transcribed word-for-word, and notes would be taken, so as to collect data from the group’s discussions and record the opinions of participants. The researcher further explained that privacy would be guaranteed and that, should any participant feel uncomfortable, they could leave at any time during the session. Only after all members expressed agreement did a focus-group session begin.

The duration for each focus-group session lasted between 90 and 120 min with a semi-structured interview format, and each group met 2 or 3 times. Prior to a session, developmental questions, covering the scope of the study, were drafted as a means to keep the focus on research objectives and to allow time for participants to reply. Once a session was over, analysis of the data proceeded only after researchers had completed the verbatim transcription of source material and participants had been given the opportunity to review transcripts for accuracy (Krueger & Casey, 2000).

Adhering to the guidelines on question types applicable to focus-groups interviews proposed by Krueger (1988), the interviewers asked the following questions:

  1. 1.
    “Please state your name and briefly describe a day at work.”
  2. 2.
    “What do you think of when you hear ‘work values’?”
  3. 3.
    “What do you get out of your current job?”
  4. 4.
    “What concerns you most when looking for a job?” and
  5. 5.
    “What do you feel you can do to obtain the most from your job?”

Researchers followed the interview questionnaire outline and modified the questions according to the participants’ background, such that individuals who were employed were asked to share their on-the-job experiences and to describe what they hoped to gain from their jobs, while students shared information about what they desired in their future work.

Data Analysis

As a means to avoid overlap between work value dimensions, the researchers considered the need to increase the clarity of the boundaries between the conceptual meanings for each representative category. Therefore Corbin and Strauss’ (2007) grounded theory approach, which is based on methods of coding and separation for real-time data via hierarchical classifications and analysis, was selected for further analysis of the textual data. In formulating the basis of a statement, each derivative dimension was explained and named. Not only were the concepts and properties of each dimension clarified in this manner, but so was the meaning of the statement it represented. This process was divided into the following steps:

  1. (1)
    Reading of the transcripts. Transcripts derived from the recordings were read from beginning to end to gain a comprehensive understanding of the material. While reviewing each group, thorough coding was done that showed the date the session took place, number of sessions, participant code names, and positions of important junctures. Following this, the materials were stored, using specific rule-based filing so as to be easily indexed for future reference.
  2. (2)
    Open coding. Three research assistants with psychology master degrees and a background of qualitative research reviewed and coded the transcripts. The coding process was as follows: first, each researcher read and open-coded the transcripts while searching for descriptive content related to work values. The descriptive statements were collated. Similar descriptive statements were grouped together into a code until all were coded. Researchers compiled a total of 231 descriptive sentences. A codebook was created, which included a comprehensive list of all codes, the properties of each code, and descriptive statements.
  3. (3)
    Axis coding. During this process, researchers interpreted and analyzed the open-coded data, summarizing concretely interviewee discourse, assigning label names with close associations to content meaning, and assembling a hierarchical classification structure. The three researchers named the 231 descriptive sentences appropriately according to the implications they carried. When there was disagreement between the three researchers regarding how a particular statement should be classified, discussions were held to obtain a consensus on the construct classification.
  4. (4)
    Selecting coding. Researchers referenced open-coded and axial coded categories as units. Over the course of extensive classification, shared characteristics were captured until core categories, comprising the full meaning of the research, could be selected, linking each category and forming a preliminary explanatory structure. Then the data was reinspected for structural completeness and superfluous categories were eliminated while insufficient ones were augmented, constituting a complete explanatory structure (Corbin & Strauss, 2007). See Table 2 for additional examples. After repeated cycles of selective coding, the researchers constructed an initial framework of 21 categories.
Table 2 Example of optional translated code table

Following several reclassifications and in-depth discussion, six categories, which were biased toward personality traits, were eliminated, for example, sense of responsibility, devotion, patience, conscientiousness, avoidance of competition, and emotional maturity. Afterward similarity of meanings was compared and categories were merged with a total of five being combined into two, for example, “upward mobility,” “office amenities,” and “stable living,” being combined into the category of “compensation and benefits,” and “interaction with colleagues” was merged with “social interaction.” There were 231 descriptive sentences belonging to the following 12 core categories: (1) compensation and benefits (57 sentences, 24.7%), (2) professional development (37 sentences, 16.0%), (3) social relationships (26 sentences, 11.3%), (4) autonomy (20 sentences, 8.7%), (5) variety (17 sentences, 7.4%), (6) self-actualization (16 sentences, 6.9%), (7) social approval (14 sentences, 6.0%), (8) prestige (12 sentences, 5.2%), (9) happiness (11 sentences, 4.8%), (10) prosocial motivation (9 sentences, 3.9%), (11) work pressures (8 sentences, 3.5%), and (12) importance of family (4 sentences, 1.7%). In order to ensure that scale measurements were clear and unambiguous, researchers reviewed the 12 categories in an attempt to identify any overlap between the categories. For example, social approval and prestige were merged due to similarity. Additionally, the work-pressures category has a negative connotation and is not something that people would hope to obtain from their jobs (Super, 1980), and was thus discarded. Happiness was also discarded because the concept was too vague; in addition to it being the result of the satisfaction of needs, it is also affected by satisfaction with other work values (Dawis & Lofquist, 1984). Therefore, the 12 core categories were reduced to seven dimensions, and they are presented in the next section. We utilized qualitative methods and constructed a conceptual work values framework from empirical data, as shown in Table 3.

Table 3 Work values structural framework

Results and Discussion

The seven dimensions that we derived in this study corresponded to the following four main typical dimensions that scholars have previously obtained through factor analysis: intrinsic values, extrinsic values, social values, and prestige values (Duffy & Sedlacek, 2007; Elizur, 1984; Ros, Schwartz, & Surkiss, 1999).

When comparing the work values framework revealed by this study with commonly used work values measures of the past, that is, WVI and MIQ, the self-actualization category of our framework belongs to the intrinsic values dimension; its implications include the importance of achieving goals and actualizing their competence; this dimension is similar to achievement in the WVI (Super, 1970) and the MIQ. The professional growth dimension also belongs to the intrinsic values, and it includes the need for job challenges, the application of knowledge, and the importance of developing professional ability. This dimension is similar to the creativity, variety, and intellectual stimulation categories in the WVI; the implications of the autonomy and prosocial motivation dimensions correspond to those found in both the WVI and MIQ.

As for extrinsic values, comfort emphasizes the stability, rewards, and benefits obtained from work; this dimension is similar to that of economic rewards and surroundings from the WVI and to that of comfort and safety from MIQ.

In terms of social values, the descriptive sentences for the interpersonal factor indicate that Taiwanese respondents place importance on interacting with others and having good relationships with colleagues; when compared with the WVI, supervisory relationships are mentioned less, and hence a dimension could not be constructed for this value. Prestige emphasizes power and status; this dimension is similar to that of prestige from the WVI and to that of status from MIQ.

A comparison of Taiwanese work values and Western work values reveals some similarities in the structural implications. The findings of this study indicate that Taiwanese work values are characterized by multiple work value dimensions. “Comfort,” which included receiving compensation and benefits associated with work, was the most widely mentioned dimension. The focus-group interviewees also suggested that “professional growth” was important to them at a job, and it was the second most widely discussed value category.

Taiwanese also emphasized work setting features that are interesting and challenging, autonomy in decision-making, and the opportunity to achieve ideals with colleagues. They placed more emphasis on meeting self-actualization, self-expression, and self-development needs. More and more are seeing work as a way to accomplish intrinsic needs, which stands in contrast to previous generations (Lu & Lin, 2002; Wong & Yuen, 2012). “Social relationships” was the third most commonly mentioned category. Unlike the WVI, supervisory relationships were rarely mentioned in our sample, and hence a dimension could not be constructed for this value. In Western interpersonal values, good relationships require the satisfaction of an individual’s emotional needs and his or her desire for achieving security through interaction with others, whereas in East Asian cultures, in addition to satisfying individual emotional needs, good relationships include maintaining a harmonious atmosphere and cooperating in the sharing of resources (Huang, Eveleth, & Huo, 1998; Wong & Yuen, 2012). Also, comparing our transcripts with the literature, past methods of classifying values dimensions often categorized power, status, economic reward, and work environment as extrinsic values. However, in the present study these four dimensions were classified into two separate categories. By comparing our transcripts with the relevant literature on the subject, it can be seen that the needs for power and status (i.e., an individual’s desire to obtain the approval and respect of others) belong to the satisfaction of intrinsic psychological needs. Needs expressed in surroundings and economic rewards (i.e., an individual’s desire for a job that provides a good living environment, salary, and benefits) belong to the satisfaction of extrinsic material needs. Therefore, power and status were renamed as prestige, and work environment and rewards were subsumed into comfort.

Our work values structural framework was established with the goal to develop a work value scale designed to be used in Taiwan for high school students. Although it was found that the value dimensions that we obtained from the qualitative study reflect universal values, distinct cultural differences were also found. However, to see existing cultural differences will require future studies to acquire cross-national samples for qualitative studies. The development of the work values framework in this study demonstrates a potential research instrument that measures a number of intrinsic and extrinsic values for studying cultures and values in future studies.

Study 2: Quantitative Evaluation of the WVA Scale

Pilot Study

Item Development and Review for the WVA Scale

After confirming the seven work values dimensions described in Study 1, the content of the 231 qualitative descriptive statements was reexamined, and the items were rewritten and revised according to the implications of their dimensions, so that the rewritten items corresponded more closely to the category meanings. Finally, researchers selected the items they considered to be the most representative of the dimensions. In total, 60 items were selected.

In order to assure the suitability of scale content, items underwent a two-stage review process conducted by experts in the field. In the first stage, 13 Taiwanese experts with practical experience in the field of education and counseling reviewed each of the 60 descriptive sentences, checking and scoring items for readability and content; in total, 27 items were revised. In the second stage, nine Taiwanese experts including three licensed psychologists, three high school and university counselors, and three professors holding doctorates conducted dimension classification of the 60 items and revised the item dimension relevance. Items were revised to ensure consistency and to discard items that had relatively weak or ambiguous relationships with the dimensions. After the review, 58 items derived from the item development remained in the WVA scale. The seven dimensional subscales were as follows: prosocial motivation (10 items), interpersonal connections (9 items), prestige (9 items), comfort (9 items), professional growth (8 items), self-actualization (7 items), and autonomy (6 items). The 5-point Likert rating scale was utilized. Respondents answered according to the degree of importance they placed on each work values item, choosing one of the following five options: “not important,” “somewhat important,” “important,” “very important,” and “extremely important.”

Participants

For the pilot test, the sample was comprised of 753 senior-high-school students, aged 16–19 years, from Northern Taiwan. Among them were 306 males (40.64%) and 444 females (58.96%); three students did not specify their gender on the questionnaire.

Procedure

The pilot-test data were collected between late March and mid-April 2012. The purpose of the scale and the method of completing it were explained to the participating students before the scale was administered to them. After they completed the scale, the students received a gift in appreciation of their participation. The administration of the WVA scale lasted for around 15–20 min.

Results

The exploratory factor analysis (EFA) with principal axis oblimin rotation was conducted. Oblimin (Promax) rotation was used to assist interpretation of the components (Fabrigar, Wegener, MacCallum, & Strahan, 1999; Thompson, 2004). The Bartlett’s test value was significant at 15,780.647 (p < .001) and the Kaiser-Meyer-Olkin measure of sampling adequacy was .928. Therefore, exploratory factor analysis was appropriate for these data. According to the criterion of eigenvalue-greater-than-one (Kaiser, 1958), seven factors were extracted, which accounted for 56.77% of the variance. Eigenvalues ranged from 1.24 to 11.73. According to the item deletion threshold proposed by Tabachnick and Fidell (2007), a factor loading smaller than 0.45 indicates that the factors explain less than 20% of the observed variance, which is less than ideal for the purposes of the present study (Leuty & Hansen, 2011).

Among the 58 items, there were 14 items (PS08, IP07, IP08, IP09, CS07, PR09, PR08, PS10, SA07, GR07, GR08, AU04, AU05, and AU06) with factor loadings smaller than 0.45, therefore they were deleted from the scale. Furthermore, five items (PS07, PR07, PS09, SA03, and SA04) showed cross-loading with other items, which means that those items were correlated with more than one factor, which may increase bias and reduce the interpretability of the scale, therefore those cross-loaded items were deleted (Hair et al., 2006, pp. 149–151). After the selection, an EFA with principal axis oblimin rotation was conducted again for the remaining 39 items. In Table 4, we can see that the factor loadings of the dimensions were .72–.87 for prosocial motivation, .66–.88 for interpersonal connections, .64–.94 for prestige, .57–.74 for comfort, .48–.81 for growth, .55–.70 for self-actualization, and .61–.72 for autonomy. The communalities ranged from .41 to .81. Since only three item loadings in the autonomy dimension met the criterion, only those three items were selected. The correlation matrix of the seven factors is presented in Table 5. Overall, 19 items were removed and the remaining 39 items were retained. There were six items in each of the six dimensions except for the autonomy dimension. The factor loadings and the internal consistency reliability coefficient (Cronbach’s α) of the seven dimensions are listed in Table 4. Since only three items remained in the autonomy dimension after the pilot test, seven new items were developed using the same method as those items produced by the pilot study and were added to this dimension for the formal test.

Table 4 Rotated factor loadings and internal consistency reliability for the seven-factor model in the pilot test
Table 5 The correlation matrix of factors in pilot study

Formal Study

Participants

Three samples of students were recruited in this study. The first sample comprised 896 high school students from Taipei City and New Taipei City. There were 445 males (49.67%) and 451 females (50.33%), aged 15–19 years. They were sampled for the item analysis, reliability of internal consistency, and construct validity. The second sample comprised 491 students (46.8% males and 53.2% females) from public and private high schools and vocational high schools in Taipei City. They were recruited for the test-retest reliability and the criterion-related validity of the WVA scale. The third sample included 896 students of the first sample and 592 students (41.72% males and 58.28% females), aged 19–23 years, from public and private colleges in Taipei City. The two groups of students were sampled for the testing of measurement invariance.

Measurement Tools

This study employs two measurement tools. The first one is the WVA scale developed in the pilot study. Including the seven new items for the autonomy dimension, there was a total of 46 items in the WVA scale. The second tool is the Work Importance Locator (WIL), which was used as the criterion of the criterion-related validity. The WIL was produced and funded by the O*NET project of the US Department of Labor (McCloy et al., 1999b), for the measurement of job seekers’ work values, which included 20 needs/items for respondents’ sorting through 5-point Likert scales ranging from “5 = very important” to “1 = least important.” The test interface and statements were changed to Chinese whereby the test prompts and translated statements underwent rigorous back-translation to ensure that the translated item meanings completely matched those of the WIL English version. WIL had several dimensions closely related with the dimensions of the WVA scale, such as altruism (WIL) versus prosocial motivation (WVA) and self-actualization (WVA) versus utilization of ability (WIL). The Pearson’s Moment-Product correlation coefficients between the related dimensions of the WVA scale and WIL were used as the indicators of criterion-related validity.

Procedure

The formal-test data were collected between September, 2012 and March, 2013. For the first sample, the procedure of administration was identical to the pilot study. For the second sample, the retest of the WVA scale was administered two weeks after the first test, and the procedure of both tests was identical to the pilot study. The administration of WIL was done immediately after the retest of the WVA scale, the administration of the two tests lasted for around 30 min. For the third sample, the recruitment of college students and the administration of the WVA scale were conducted in January and March, 2013. The procedure was identical to the pilot test of the WVA scale.

Results

Item Analysis and Exploratory Factor Analysis

Because seven new items of the autonomy dimension were added and tested along with the 39 old items, we re-conducted the item analysis and EFA as in the pilot test to reconfirm the quality of items. The EFA was subsequently conducted with the principal axis factor with promax rotation. The Kaiser-Meyer-Olkin coefficient was .944, and Bartlett’s test value was significant at 19,243.192 (p < .001). Therefore, exploratory factor analysis was appropriate for these data. The EFA of the 42-item version identified the most plausible models based on a scree plot. The eigenvalue ≥ 1 was used as the criterion for determining the number of factors to retain (Thompson, 2004). As illustrated in Figure 1, these models were composed of seven factors, the eigenvalues ranged from 1.28 to 11.94. A total of 62.6% of the variance was accounted for by the seven components. The factor loadings (Table 6) of the dimensions were .77–.85 for prosocial motivation, .60–.86 for interpersonal connections, 56–.89 for prestige, .50–.77 for comfort, .46–.83 for growth, .36–.68 for self-actualization, and .61–.75 for autonomy. The communalities ranged from .47 to .78. The correlation matrix of factor loading is given in Table 7. Three items of the newly added seven items in the autonomy dimensions were selected based on their factor loadings and item-total correlations and were integrated into the formal version of the WVA scale, for a total of 42 items, 6 items per dimension. The results reconfirmed the quality of the items and the robustness of the seven dimensions (factors) of the WVA scale.

Figure 1 Exploratory Factor Analysis scree plot of the 42 items.
Table 6 Rotated factor loadings and reliability indices for the 7-factor model in the formal test
Table 7 The correlation matrix of factors in formal study

Reliability Analysis

The internal consistency reliability coefficient ranged from .81 to .92. The overall scale reliability coefficient was .95. The test-retest reliability coefficients ranged from .83 to .91 (see Table 6).

Construct Validity Analysis

LISREL 8.70 (Jöreskog & Sörbom, 2004) statistical software was used to conduct CFA to validate the 42 formal items obtained by EFA and to determine whether a reliable 7-factor work values model had been established. The maximum likelihood estimates from the sample covariance matrix were used in the CFA.

For CFA, the chi-square (χ2) values are regularly used as a criterion for model fit. However, because an excessive sample size or a stronger correlation between variables may have been the reason for the increased chi-square values (Kline, 2010), approximate fit indices, such as the Comparative Fit Index (CFI), which is less sensitive to large samples, the root-mean-square error of approximation (RMSEA), and the Standardized Root-Mean-Square Residual (SRMR), were also assessed (Byrne, 2001). According to the criteria of Hu and Bentler (1999), the CFI values acceptable for model fit are .90 or greater (Hu & Bentler, 1999); RMSEA values less than .05 indicate close model fit, values between .05 and .08 indicate reasonable fit, those between .08 and .10 indicate mediocre fit, and values greater than .10 indicate unacceptable fit; the SRMR values acceptable for model fit are .08 or greater.

For the model evaluation as a whole, χ2 (798, N = 896) = 2,199.86 (p < .001) did not support the fit of the model. The other values (RMSEA = .004, CFI = .98, and SRMR = .05) met the criteria recommended by Hu and Bentler (1999) as being indicative of a good model fit. The CFA results are also shown in Figure 2. These analyses show that there was a good fit between the framework model proposed by this study and the research data collected in the study.

Figure 2 Confirmatory factor analysis of the seven work values dimensions.

Criterion-Related Validity Analysis

Table 8 indicates that the coefficient (r = .52) of the prosocial motivation dimension from the WVA scale and that of the social service dimension from the WIL were significant (p < .05). The autonomy dimension from the WVA scale and the creativity, autonomy, and responsibility dimensions from WIL were all significantly and positively correlated (r = .29, .36, and .39, respectively; p < .01 for all). The comfort dimension from the WVA scale and the work conditions, company policies, security, and rewards dimensions from WIL were also all significantly positively correlated (r = .19, .21, .27, and .29, respectively; p < .01). The professional growth dimension from the WVA scale and the ability utilization and achievement dimensions from WIL were significantly and positively correlated (r = .25 and .27, respectively; p < .05). The coefficient of the interpersonal connections dimension from the WVA scale and the coworkers dimension from WIL was .21, (p < .01). The prestige from the WVA scale and advancement and compensation from WIL were significantly and positively correlated (r = .22 and .37, respectively; p < .01). The self-actualization dimension from the WVA scale and the ability utilization and achievement dimensions from WIL were significantly and positively correlated (r = .33 and .35, respectively; p < .05). The results that scores of the WVA dimensions were significantly correlated with the relevant dimensions of WIL provided supporting evidence for the criterion-related validity of the WVA scale.

Table 8 Criterion-related validity for the WVA scale and WIL scale

Measurement-Invariance Analysis

Previous research proposed the tests of configural invariance, metric invariance, and scalar invariance for measurement invariance (Schmitt & Kuljanin, 2008). Three steps, each step corresponding to a test model, were conducted to test the measurement of invariance of the WVA scale between the groups of college and high school students. The first was testing the model of configural invariance, which implies that there should be the same number of factors in each group and the same pattern of fixed and free parameters. The second step is testing the metric invariance model, which implies equal factor loadings across groups. The third step is testing the scalar invariance model, which indicates the invariance of the item intercepts linking the indicators (items) to the latent variable (Steinmetz, Schmidt, Tina-Booh, Wieczorek, & Schwartz, 2009).

The most commonly used model estimation parameter is Δχ2, but Δχ2 is very sensitive to the slight difference between two comparative models for large sample size data sets, so four other criteria of invariance analysis proposed by researchers (Chen, 2007; Cheung & Rensvold, 2002; Little, 1997; Meade, Johnson, & Braddy, 2008), including changes in the |ΔCFI|, |ΔRMSEA|, |ΔSRMR|, and the Tucker Lewis Index (ΔTLI), were also used for evaluation. When |ΔCFI| ≤ .01 (Cheung & Rensvold, 2002), |ΔTLI| ≤ .05 (Little, 1997), |ΔRMSEA| ≤ .02 (Meade et al., 2008), and |ΔSRMR| ≤ .03 (Chen, 2007), it would suggest that the two comparative models were not substantially different from each other. However, according to previous researchers, there is not a single criterion that is powerful enough to support the claim of measurement invariance; therefore, a compromised way was proposed: the results that the majority of criteria are within the suggested thresholds can be interpreted as evidence for measurement invariance (Vandenberg & Lance, 2000; Wu & Hughes, 2015). University and high school student WVI scale samples and their nested assays on invariance have been conducted (see Table 9). With respect to configural invariance (model 1), except for χ2 (798, N = 1,488) = 2,945.168 (p < .001), the model’s other indices of goodness-of-fit (e.g., CFI = .947; TLI = .943; RMSEA = .043; and SRMR = .041) are acceptable, which indicated that configural invariance has been established and the two samples had the same number of factors in the WVA scale.

Table 9 Measurement-invariance analyses of WVA for students of different education levels

The metric invariance test revealed that Δχ2 (833, N = 1,488) = 2,001.785, p < .001, |ΔSRMR| = .017, |ΔRMSEA| = .009, |ΔCFI| = .047, and |ΔTLI| = −.05. Except for Δχ2 and |ΔCFI|, all other indicators met the criteria proposed by researchers (Chen, 2007; Meade et al., 2008). According to Vandenberg and Lance (2000) and Wu and Hughes (2015), the majority of indicators supported the model fit of metric invariance. The intercept invariance test (model 3) showed that the Δχ2 (833, N = 1,488) = 2,001.785 (p < .001), |ΔSRMR| = .009, |ΔRMSEA| = .005, |ΔTLI| = .022, and |ΔCFI| = .02. Except for Δχ2, all other indicators met the recommended threshold values for measurement invariance (Chen, 2007; Meade et al., 2008). Accordingly, the seven-dimension model of the WVA scale demonstrated appropriate cross-group invariance and can be used to measure the work values for both university and high school students in Taiwan.

General Discussion and Conclusions

There are several features in the current research. In Study 1, a qualitative research methodology was employed to obtain 231 descriptive statements related to work values, and after repeated coding and classification, seven work values dimensions were derived. In Study 2, quantitative research methods were implemented to conduct scale evaluations in order to verify that the seven derived work values dimensions are appropriate for, and capable of, explaining the framework of a Taiwan work values scale. The dimension content is clear and, with no content overlap, is an improvement on existing work values scales, the dimensions of which are often vague or redundant. Moreover, items were revised or deleted, and those with low factor loadings or a conceptually vague content were discarded, thus ensuring that items adhere to the core concepts of the dimensions.

In terms of reliability, the scale derived herein compared favorably with other work values scales in which high school students were used as samples. The internal consistency reliability coefficients for the WVA scale developed in the present study ranged from .84 to .92, with an overall scale reliability coefficient of .95 and a test-retest reliability coefficient of .94. All of the above values indicate good reliability that is generally superior to similar scales (see Zytowski, 1970, or Pryor, 1979; for some comparisons). The values indicate that when reapplied to samples of high school students, the WVA scale should obtain consistent and stable results.

In terms of validity, previous studies have rarely employed CFA to explore construct validity, such as in the Career Value Scale developed by Macnab, Bakker, and Fitzsimmons (2005). The O*NET WIP utilized CFA, but the model fit index was poor. Nevertheless, appropriate sample sizes of participants enabled the WVA scale to use two types of statistical analyses to yield even more rigorous results, and the model fit was within the reasonable range suggested by Hu and Bentler (1999). Furthermore, the significant correlations between dimensions of the WVA scale and WIL provided solid evidence that the WVA scale has the capability of measuring relevant work value constructs which had been measured in other popularly used Western tools. The measurement-invariance analysis revealed that, whether for preemployment high school of university students, the WVA scale developed through this research had identical factor structure and measurement capability, the test items did not demonstrate inconsistent meanings for students with different education levels, and can be trusted with considerable validity between the different groups. This finding also supported the claims of Harpaz and Fu (2002) that the structure of meanings of work remains consistent during the transition of adolescence to adulthood.

In terms of practical application, the WVA scale will provide students with a better understanding of their own work values, which, for high school students especially (Ginzberg et al., 1951; Super, 1980), may still be developing. School counselors may provide adolescents with the opportunity to engage in self-exploration related with their emphasized values based on the results of the WVA scale, which may greatly benefit their later career development. Hence, helping young people to understand their own work values is an important function of the WVA scale.

Qualitative research methods were used in this study to derive work values dimensions that reflect local culture. Although the WVA scale was constructed based on work values dimensions that reflect local culture, this study also found that the work value dimensions in the WVA scale are also highly similar to those for Western society. Despite the similarity of the framework and dimensions, future research may investigate whether there exist different emphases for different dimensions between different cultures. In addition, multi-sample structural equation modeling can be used to conduct intercultural measurement invariance to test whether the extension validity of the structural model of the present study is stable in cross-cultural settings.

References

  • American Educational Research Association, American Psychological Association, National Council on Measurement in Education. (2014). The standards for educational and psychological testing. Washington, DC: AERA/APA/NCME. First citation in articleGoogle Scholar

  • Betts, J. (2011). The economics of tracking in education. In E. HanushekS. MachinL. WoessmannEds., Handbook of the economics of education (Vol. 3, pp. 341–381). Amsterdam, North Holland. First citation in articleGoogle Scholar

  • Byrne, B. M. (2001). Structural equation modeling with AMOS, EQS, and LISREL: Comparative approaches to testing for the factorial validity of a measuring instrument. International Journal of Testing, 1, 55–86. doi: 10.1207/S15327574IJT0101_4 First citation in articleCrossrefGoogle Scholar

  • Cabrera, A. F., Nora, A. & Castaneda, M. B. (1993). College persistence: Structural equations modeling test of an integrated model of student retention. Journal of Higher Education, 64, 123–139. doi: 10.2307/2960026 First citation in articleCrossrefGoogle Scholar

  • Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14, 464–504. doi: 10.1080/10705510701301834 First citation in articleCrossrefGoogle Scholar

  • Chen, Y. H., Wang, J. T., Liu, Y. S., Ou, T. H. & Li, K. J. (1987). Work values scale revised report. Journal of National University of Tainan, 20, 1–33. First citation in articleGoogle Scholar

  • Cheung, G. W. & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255. doi: 10.1207/S15328007SEM0902_5 First citation in articleCrossrefGoogle Scholar

  • Corbin, J. & Strauss, A. (2007). Basics of qualitative research: Techniques and procedures for developing grounded theory (3rd ed.). Thousand Oaks, CA: Sage. First citation in articleGoogle Scholar

  • Creswell, J. W., Plano Clark, V. L., Gutmann, M. L. & Hanson, W. E. (2003). Advanced mixed methods research designs. In A. TashakkoriC. TeddlieEds., Handbook of mixed methods in social and behaviour research (pp. 209–240). Thousand Oaks, CA: Sage. First citation in articleGoogle Scholar

  • Dawis, R. V. (1991). Vocational interests, values, and preferences. In M. D. DunnetteL. M. HoughEds., Handbook of industrial organizational psychology (pp. 833–871). Palo Alto, CA: Consulting Psychologist Press. First citation in articleGoogle Scholar

  • Dawis, R. V. & Lofquist, L. H. (1984). A psychological theory of work adjustment. Minneapolis, MN: University of Minnesota Press. First citation in articleGoogle Scholar

  • Dolan, S. L., Díez-Piñol, M., Fernández-Alles, M., Martín-Prius, A. & Martínez-Fierro, S. (2004). Exploratory study of within-country differences in work and life values: The case of Spanish business students. International Journal of Cross Cultural Management, 4, 157–180. doi: 10.1177/1470595804044747 First citation in articleCrossrefGoogle Scholar

  • Dose, J. J. (1997). Work values: An integrative framework and illustrative application to organizational socialization. Journal of Occupational and Organizational Psychology, 70, 219–240. doi: 10.1111/j.2044-8325.1997.tb00645.x First citation in articleCrossrefGoogle Scholar

  • Duffy, R. D. & Sedlacek, W. E. (2007). The work values of first-year college students: Exploring group differences. The Career Development Quarterly, 55, 359–364. doi: 10.1002/j.2161-0045.2007.tb00090.x First citation in articleCrossrefGoogle Scholar

  • Elizur, D. (1984). Facets of work values: A structural analysis of work outcomes. The Journal of Applied Psychology, 69, 379–389. doi: 10.1037/0021-9010.69.3.379 First citation in articleCrossrefGoogle Scholar

  • Fabrigar, L. R., Wegener, D. T., MacCallum, R. C. & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4, 272–299. doi: 10.1037/1082-989X.4.3.272 First citation in articleCrossrefGoogle Scholar

  • Fuller, T. D., Edwards, J. N., Sermsri, S. & Vorakitphokatorn, S. (1993). Housing, stress, and physical well-being: Evidence from Thailand. Social Science and Medicine, 36, 1417–1428. doi: 10.4135/9781483349008.n6 First citation in articleCrossrefGoogle Scholar

  • Gay, E. G., Weiss, D. J., Hendel, D. D., Dawis, R. V. & Lofquist, L. H. (1971). Minnesota studies in vocational rehabilitation: Vol. XXVIII. Manual for the Minnesota Importance Questionnaire. Minneapolis, MN: University of Minnesota, Industrial Relations Center. First citation in articleGoogle Scholar

  • Gerpott, T. J. (1988). Career orientations in different countries and companies. Journal of Management Studies, 25, 439–462. doi: 10.1111/j.1467-6486.1988.tb00709.x First citation in articleCrossrefGoogle Scholar

  • Ginzberg, E., Ginsburg, S. W., Axelrad, S. & Herma, J. L. (1951). Toward a theory of occupational choice. New York, NY: Columbia University Press. First citation in articleGoogle Scholar

  • Hair, J. F.Jr., Black, W. C., Babin, B. J., Anderson, R. E. & Tatham, R. L. (2006). Multivariate data analysis (6th ed.). Upper Saddle River, NJ: Pearson-Prentice Hall. First citation in articleGoogle Scholar

  • Harpaz, I. & Fu, X. (2002). The structure of the meaning of work: A relative stability amidst change. Human Relations, 55, 639–667. doi: 10.1177/0018726702556002 First citation in articleCrossrefGoogle Scholar

  • Hartung, P. J., Fouad, N. A., Leong, F. T. & Hardin, E. E. (2010). Individualism-collectivism links to occupational plans and work values. Journal of Career Assessment, 18, 34–45. doi: 10.1177/1069072709340526 First citation in articleCrossrefGoogle Scholar

  • Hu, L. & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. doi: 10.1080/10705519909540118 First citation in articleCrossrefGoogle Scholar

  • Huang, H. J., Eveleth, D. M. & Huo, Y. P. (1998, December). Chinese work-related value system: Developing a “GCF-LEACH” framework for comparative studies among Chinese societies. Paper presented at the Inaugural Conference of the Asia Academy of Management, Hong Kong. First citation in articleGoogle Scholar

  • Hung, J. P. & Liu, C. M. (2003). Studies on work values in Taiwan: A critical review. Research in Applied Psychology, 19, 219–250. First citation in articleGoogle Scholar

  • Jambrak, J., Deane, F. P. & Williams, V. (2014). Value motivations predict burnout and intentions to leave among mental health professionals. Journal of Mental Health, 23, 120–124. doi: 10.3109/09638237.2013.869576 First citation in articleCrossrefGoogle Scholar

  • Jin, J. & Rounds, J. (2012). Stability and change in work values: A meta-analysis of longitudinal studies. Journal of Vocational Behavior, 80, 326–339. doi: 10.1016/j.jvb.2011.10.007 First citation in articleCrossrefGoogle Scholar

  • Johnson, M. K. (2002). Social origins, adolescent experiences, and work value trajectories during the transition to adulthood. Social Forces, 80, 1307–1340. doi: 10.1353/sof.2002.0028 First citation in articleCrossrefGoogle Scholar

  • Jöreskog, K. G. & Sörbom, D. (2004). LISREL 8.7 for Windows [Computer Software]. Lincolnwood, IL: Scientific Software International. First citation in articleGoogle Scholar

  • Judge, T. A. & Bretz, R. D. (1992). Effects of work values on job choice decisions. The Journal of Applied Psychology, 77, 261–271. doi: 10.1037/0021-9010.77.3.261 First citation in articleCrossrefGoogle Scholar

  • Kaiser, H. F. (1958). The varimax criterion for analytic rotation in factor analysis. Psychometrika, 23, 187–200. doi: 10.1007/BF02289233 First citation in articleCrossrefGoogle Scholar

  • Kline, R. B. (2010). Principles and practice of structural equation modeling (2nd ed.). New York, NY: Guilford Press. First citation in articleGoogle Scholar

  • Krueger, R. A. (1988). Focus groups: A practical guide for applied research. London, UK: Sage. First citation in articleGoogle Scholar

  • Krueger, R. A. & Casey, M. A. (2000). Focus groups: A practical guide for applied research (3rd ed.). Thousand Oaks, CA: Sage. First citation in articleCrossrefGoogle Scholar

  • Lee, C. S., Hung, D. K. M. & Ling, T. C. (2012). Work values of Generation Y preservice teachers in Malaysia. Procedia-Social and Behavioral Sciences, 65, 704–710. doi: 10.1016/j.sbspro.2012.11.187 First citation in articleCrossrefGoogle Scholar

  • Lent, R. W. & Brown, S. D. (2006). On conceptualizing and assessing social cognitive constructs in career research: A measurement guide. Journal of Career Assessment, 14, 12–35. doi: 10.1177/1069072705281364 First citation in articleCrossrefGoogle Scholar

  • Leuty, M. E. & Hansen, J. I. C. (2011). Evidence of construct validity for work values. Journal of Vocational Behavior, 79, 379–390. doi: 10.1016/j.jvb.2011.04.008 First citation in articleCrossrefGoogle Scholar

  • Little, T. D. (1997). Mean and covariance structures (MACS) analyses of cross-cultural data: Practical and theoretical issues. Multivariate Behavioral Research, 32, 53–76. doi: 10.1207/s15327906mbr3201_3 First citation in articleCrossrefGoogle Scholar

  • Long, J. S. (1983). Confirmatory factor analysis: A preface to LISREL. Beverly Hills, CA: Sage. First citation in articleCrossrefGoogle Scholar

  • Lu, L. & Lin, G. C. (2002). Work values and job adjustment of Taiwanese workers. Research and Practice in Human Resource Management, 10, 70–76. First citation in articleGoogle Scholar

  • Macnab, D., Bakker, S. & Fitzsimmons, G. (2005). Career values scale manual and user’s guide. Alberta, CA: Psychometrics Canada. First citation in articleGoogle Scholar

  • Marsh, H. W., Nagengast, B. & Morin, A. J. S. (2013). Measurement invariance of big-five factors over the life span: ESEM tests of gender, age, plasticity, maturity, and La Dolce Vita effects. Developmental Psychology, 49, 1194–1218. doi: 10.1037/a0026913 First citation in articleCrossrefGoogle Scholar

  • Maynard, D. C. & Parfyonova, N. M. (2013). Perceived overqualification and withdrawal behaviours: Examining the roles of job attitudes and work values. Journal of Occupational and Organizational Psychology, 86, 435–455. doi: 10.1111/joop.12006 First citation in articleCrossrefGoogle Scholar

  • McCloy, R., Waugh, G., Medsker, G., Wall, J., Rivkin, D. & Lewis, P. (1999a). Development of the O*NET computerized work importance profiler. Raleigh, NC: National Center for O*NET Development. First citation in articleGoogle Scholar

  • McCloy, R., Waugh, G., Medsker, G., Wall, J., Rivkin, D. & Lewis, P. (1999b). Development of the O*NET paper-and-pencil work importance locator. Raleigh, NC: National Center for O*NET Development. First citation in articleGoogle Scholar

  • Meade, A. W., Johnson, E. C. & Braddy, P. W. (2008). Power and sensitivity of alternative fit indices in tests of measurement invariance. The Journal of Applied Psychology, 93, 568–592. doi: 10.1037/0021-9010.93.3.568 First citation in articleCrossrefGoogle Scholar

  • Morgan, D. L. & Krueger, R. A. (1993). When to use focus groups and why. In D. L. MorganEd., Successful focus groups: Advancing the state of the art (pp. 3–19). Thousand Oaks, CA: Sage. First citation in articleGoogle Scholar

  • Nunnally, J. C. (1978). Psychometric theory. New York, NY: McGraw Hill, Inc. First citation in articleGoogle Scholar

  • Pryor, R. (1979). In search of a concept: Work values. Vocational Guidance Quarterly, 27, 250–258. doi: 10.1002/j.2164-585X.1979.tb00993.x First citation in articleCrossrefGoogle Scholar

  • Raju, N. S., Laffitte, L. J. & Byrne, B. M. (2002). Measurement equivalence: A comparison of methods based on confirmatory factor analysis and item response theory. The Journal of Applied Psychology, 87, 517–529. doi: 10.1037/0021-9010.87.3.517 First citation in articleCrossrefGoogle Scholar

  • Robinson, C. H. & Betz, N. E. (2008). A psychometric evaluation of Super’s Work Values Inventory-Revised. Journal of Career Assessment, 16, 456–473. doi: 10.1177/1069072708318903 First citation in articleCrossrefGoogle Scholar

  • Rokeach, M. (1973). The nature of human values. New York, NY: Free Press. First citation in articleGoogle Scholar

  • Ros, M., Schwartz, S. H. & Surkiss, S. (1999). Basic individual values, work values, and the meaning of work. Applied Psychology, 48, 49–71. doi: 10.1111/j.1464-0597.1999.tb00048.x First citation in articleCrossrefGoogle Scholar

  • Rounds, J. B.Jr. & Armstrong, P. I. (2005). Assessment of needs and values. In D. BrownR. LentEds., Career development and counseling: Putting theory and research to work (pp. 305–329). New York, NY: Wiley. First citation in articleGoogle Scholar

  • Rounds, J. B.Jr., Henly, G. A., Dawis, R. V., Lofquist, L. H. & Weiss, D. J. (1981). Manual for the Minnesota importance questionnaire. Minnesota, MA: Vocational Psychology Research Center. University of Minnesota. First citation in articleGoogle Scholar

  • Schmitt, N. & Kuljanin, G. (2008). Measurement invariance: Review of practice and limitations. Human Resource Management Review, 18, 210–222. doi: 10.1016/j.hrmr.2008.03.003 First citation in articleCrossrefGoogle Scholar

  • Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 1–65. doi: 10.1016/S0065-2601(08)60281-6 First citation in articleCrossrefGoogle Scholar

  • Sortheix, F. M., Chow, A. & Salmela-Aro, K. (2015). Work values and the transition to work life: A longitudinal study. Journal of Vocational Behavior, 89, 162–177. doi: 10.1016/j.jvb.2015.06.001 First citation in articleCrossrefGoogle Scholar

  • Steinmetz, H., Schmidt, P., Tina-Booh, A., Wieczorek, S. & Schwartz, S. H. (2009). Testing measurement invariance using multigroup CFA: Differences between educational groups in human values measurement. Qualitative and Quantitative, 43, 599–616. doi: 10.1007/s11135-007-9143-x First citation in articleCrossrefGoogle Scholar

  • Sung, Y. T. & Chao, T. Y. (2015). Construction of the examination stress scale for adolescent students. Measurement and Evaluation in Counseling and Development, 48, 44–58. doi: 10.1177/0748175614538062 First citation in articleCrossrefGoogle Scholar

  • Sung, Y. T., Cheng, Y. W. & Hsueh, J. H. (2017). Identifying the career-interest profiles of junior-high-school students through latent profile analysis. Journal of Psychology, 151(3), 229–246. doi: 10.1080/00223980.2016.1261076 First citation in articleCrossrefGoogle Scholar

  • Sung, Y. T., Chao, T. Y. & Tseng, F. L. (2016). Reexamining the relationship between test anxiety and learning achievement: An individual-differences perspective. Contemporary Educational Psychology, 46, 241–252. doi: 10.1016/j.cedpsych.2016.07.001 First citation in articleCrossrefGoogle Scholar

  • Sung, Y. T., Cheng, Y. W. & Wu, J. S. (2016). Constructing a situation-based career interest assessment for junior high school students and examining their interest structure. Journal of Career Assessment, 24, 347–365. doi: 10.1177/1069072715580419 First citation in articleCrossrefGoogle Scholar

  • Sung, Y. T., Huang, L. Y., Tseng, F. L. & Chang, K. E. (2014). The aspects and ability groups in which little fish perform worse than big fish: Examining the big-fish-little-pond effect in the context of school tracking. Contemporary Educational Psychology, 39, 220–232. doi: 10.1016/j.cedpsych.2014.05.002 First citation in articleCrossrefGoogle Scholar

  • Super, D. E. (1970). Work values inventory. Boston, MA: Houghton Mifflin. First citation in articleGoogle Scholar

  • Super, D. E. (1980). A life-span, life-space approach to career development. Journal of Vocational Behavior, 16, 282–298. doi: 10.1016/0001-8791(80)90056-1 First citation in articleCrossrefGoogle Scholar

  • Super, D. E. & Crites, J. O. (1957). Vocational development: A framework for research. New York, NY: Teachers College Press, Columbia University. First citation in articleGoogle Scholar

  • Super, D. E. & Sverko, B. (1995). Life roles, values and careers: International findings of the Work Importance Study. San Francisco, CA: Jossey Bass. First citation in articleGoogle Scholar

  • Tabachnick, B. G. & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Needham Heights, MA: Allyn and Bacon. First citation in articleGoogle Scholar

  • Tashakkori, A. & Teddlie, C. (2003). Handbook of mixed methods in social and behavior research. Thousand Oaks, CA: Sage. First citation in articleGoogle Scholar

  • Thompson, B. (2004). Exploratory and confirmatory factor analysis: Understanding concepts and applications. Washington, DC: American Psychological Association. First citation in articleCrossrefGoogle Scholar

  • Trautwein, U., Lüdtke, O., Marsh, H. W., Köller, O. & Baumert, J. (2006). Tracking, grading, and student motivation: Using group composition and status to predict self-concept and interest in ninth-grade mathematics. Journal of Educational Psychology, 98, 788–806. doi: 10.1037/0022-0663.98.4.788 First citation in articleCrossrefGoogle Scholar

  • Vandenberg, R. J. & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4–70. doi: 10.1177/109442810031002 First citation in articleCrossrefGoogle Scholar

  • White, C. (2006). Towards an understanding of the relationship between work values and cultural orientations. International Journal of Hospitality Management, 25, 699–715. doi: 10.1016/j.ijhm.2005.07.002 First citation in articleCrossrefGoogle Scholar

  • Wöhrmann, A. M., Fasbender, U. & Deller, J. (2016). Using work values to predict post-retirement work intentions. Career Development Quarterly, 64, 98–113. doi: 10.1002/cdq.12044 First citation in articleCrossrefGoogle Scholar

  • Wong, S. W. & Yuen, M. (2012). Work values of university students in Chinese Mainland, Taiwan, and Hong Kong. International Journal for the Advancement of Counseling, 34, 269–285. doi: 10.1007/s10447-012-9155-7 First citation in articleCrossrefGoogle Scholar

  • Wu, J. Y. & Hughes, J. N. (2015). Teacher network of relationships inventory: Measurement invariance of academically at-risk students across ages 6 to 15. School Psychology Quarterly, 30, 23–36. doi: 10.1037/spq0000063 First citation in articleCrossrefGoogle Scholar

  • Wu, T. S., Li, K. C., Liu, Y. S. & Ou, H. M. (1996). Work values scale development. Taipei, Taiwan: Youth Development Administration, Ministry of Education. First citation in articleGoogle Scholar

  • Zytowski, D. G. (1970). The concept of work values. Vocational Guidance Quarterly, 18, 176–186. doi: 10.1002/j.2164-585X.1970.tb00231.x First citation in articleCrossrefGoogle Scholar

  • Zytowski, D. G. (1994). A super contribution to vocational theory: Work values. The Career Development Quarterly, 43, 25–31. First citation in articleCrossrefGoogle Scholar

  • Zytowski, D. G. (2006). Super Work Values Inventory-Revised: Technical manual (version 1.0), Retrieved from http://www.Kuder.com/PublicWeb/swv_manual.aspx First citation in articleGoogle Scholar

Yun-Tim Yvonne Chang, Department of Educational Psychology and Counseling, National Taiwan Normal University, 162, Sec. 1, Ho-Ping E. Rd., 10610 Taipei, Taiwan, E-mail