مدارک و شواهد از اعتبار سازه به ارزش کار
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|21682||2011||12 صفحه PDF||سفارش دهید||9257 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Vocational Behavior, Volume 79, Issue 2, October 2011, Pages 379–390
Despite the importance of work values in the process of career adjustment (Dawis, 2002), little empirical research has focused on articulating the domains represented within the construct of work values and the examination of evidence of validity for the construct has been limited. Furthermore, the larger number of work values measures has made it difficult to determine the key domains that constitute the construct. The current study sought to examine multiple measures of work values to understand domains represented within the construct of work values and to establish evidence of validity for these domains. Principal Components Analysis utilizing scores on the Minnesota Importance Questionnaire (Rounds, Henley, Dawis, Lofquist, & Weiss, 1981), the revised edition of Super's Work Values Inventory (SWVI-R; Zytowski, 2006), and Manhardt's Work Values Inventory (Manhardt, 1972) found that six components best explained the data. These components reflected the importance of: the working environment, having challenging work, opportunities for status and income, autonomy, organizational support, and relationships. Implications for research and practice are discussed.
Research on work values suggests that values are an important predictor of job satisfaction and tenure (Dawis, 2002), but little research on work values has been focused on understanding the domain of work values. Moreover, the research to further understand the construct of work values has been stagnant in the past few decades. This deficit in the literature, combined with the ending of research programs on values that were lead by Donald Super during the 60's and 70's, and Rene Dawis and Lloyd Lofquist in the 60's through 90's, has resulted in little 21st century research on the construct of work values. Furthermore, few commercially available measures of work values with adequate evidence of validity and reliability are readily available, leading to various measures of values being created for the purpose of individual studies. This trend, if anything, has further contributed to the lack of research on the validity of work value domains. The purpose of the current study was to add to evidence of validity for the construct of work values by determining the domains represented across multiple measures of work values. 1.1. Conceptualizing work values In the vocational literature, Allport (1961) provided the preliminary definition of values as beliefs that cause individuals to act on their preferences. Donald Super (1980) later defined work values as “an objective, either a psychological state, a relationship, or material condition, that one seeks to attain” (p.130). Furthermore, Dawis and Lofquist (1984) suggested that work values were central to understanding job satisfaction, as posited in their Theory of Work Adjustment (TWA), which assumes that individuals develop job satisfaction when their values are fulfilled by aspects of their job. The development of different instruments to measure work values also provides some insight into how the construct of work values has been conceptualized. 1.2. Minnesota Importance Questionnaire As Berings, de Fruyt, and Bouwen (2004), as well as Rounds and Armstrong (2005) note, many of the existing measures of work values are very similar despite varying conceptualizations of work values that drove the construction of the instruments. Popular measures of work values – the Minnesota Importance Questionnaire (MIQ; Rounds et al., 1981) and the revised edition of Super's Work Values Inventory, (SWVI-R; Zytowski, 2006) – assess the most widely accepted domains within the construct of work values. The MIQ was developed to understand individual factors that contribute to job satisfaction, as an element of decades of research on TWA by Dawis and Lofquist (1984). The MIQ is hierarchically structured and is comprised of 20 separate work needs which are grouped into six work values: Achievement (the importance of accomplishment), Comfort (freedom from stress), Status (the importance of recognition and prestige), Altruism (the importance of helping others), Safety (the importance of stability and structure), and Autonomy (the importance of control over one's work). The six values scales were developed empirically by factor analysis of the 20 work needs represented on the MIQ (Lofquist & Dawis, 1978). Although the lower-order scales on the MIQ are referred to as needs (Lofquist & Dawis, 1978), they are analogous to work values on other values measures (Macnab & Fitzsimmons, 1987). Strong evidence of validity and reliability exists for MIQ scores (see Hendel and Weiss, 1970 and Weiss et al., 1966). Of the work values measures available, Rounds (1990) noted that the MIQ appears to be the most comprehensive measure of the construct. Moreover, a strength of the MIQ is that it is presented as a comparison task, with individuals asked to indicate their preference among pairs of need statements, allowing for more differentiation of the importance of the 20 work needs, versus ranking or Likert scaling methods (Thurstone, 1954). 1.3. Super's Work Values Inventory Donald Super offered another popular conceptualization of work values. Super's Work Values Inventory (Super, 1970) was created to operationalize values in Super's historic Career Pattern Study that followed the career development of a group of 9th grade boys into adulthood and eventually led to the development of Super's Life-Span, Life-Space theory (Super, 1985). Super (1970) mentions that items for the Work Values Scale were selected from Spranger (1928), the Allport, Vernon, and Lindzey (1960), and research on job satisfaction and morale by Hoppock, 1935 and Centers, 1948. Additional items were selected based on other theories, leading to the inclusion of additional work values proposed by Darley and Hagenah, 1955 and Fryer, 1931, Ginzberg and colleagues (Ginzberg, Ginsburg, Axelrad, & Herma, 1951) and Super (1957). The current version of this measure is Super's Work Values Inventory-Revised (SWVI-R; Zytowski, 2006). Updates to the measure included dropping three scales (Altruism, Esthetics, and Management) that demonstrated poor evidence of discriminate validity given high correlations with measures of interests (Zytowski, 2006). The 72 item SWVI-R includes 12 work values scales—Achievement, Co-Workers, Creativity, Income, Independence, Lifestyle, Challenge, Prestige, Security, Supervision, Variety, and Workplace. 1.4. Manhardt's Work Values Inventory Along with the work values included in the MIQ and SWVI-R, additional work values have been suggested. Manhardt (1972) developed a measure of job characteristics called the Work Values Inventory (MWVI). However, information about the process used to select items for this measure is limited. Consisting of 25 items, Manhardt (1972) found that three factors emerged from factor analysis of items. Principal components analyses by Meyer, Irving, and Allen (1998), on a sample of university students, confirmed three components fit the data and yielded interpretable components. Meyer and colleagues labeled the three extracted factors Comfort and Security (α = 0.72), Competence and Growth (α = 0.65), and Status and Independence (α = 0.68). The first factor (Comfort and Security) describes characteristics of a comfortable working environment including having a routine schedule, leisure time, and good relationships with coworkers. The second factor (Competence and Growth) included items that were characteristics of successful workers such as the importance of responsibility, advancement, and supervision of others. The final factor (Status and Independence) included items that were intrinsic characteristics related to the nature of work such as, independence, continued development of skills, and intellectual stimulation. At the item level, Manhardt's measure includes questions that are not included in other measures such as “satisfies your cultural and esthetic interests” which loaded on the Competence and Growth factor, and “permits a regular routine in time and place of work” which loaded on Comfort and Security. 1.5. Additional measures of work values Beyond the values measured by the MIQ, SWVI-R, and MWVI, other domains have been hypothesized to comprise the work values construct. For instance, Pryor (1981) developed a values measure named the Work Aspect Preference Scale, which included 13 values scales—Security, Self Development, Surroundings, Altruism, Life Style, Physical Activity, Detachment, Independence, Prestige, Management, Co-workers, Creativity, and Money. A unique aspect of Pryor's conceptualization is the inclusion of the importance of being physically active (Physical Activity) and separation of life and work roles (Detachment). Berings (2002) also offers another conceptualization of work values. Much like other measures, some of Berings' values cover areas such as relationships, autonomy, creativity, earnings, security, and achievement. However, Berings' conceptualization also incorporates some different domains such as the importance of innovation, rationality, structure, and stress avoidance. Additional conceptualizations of work values have appeared in different studies as well (see for example, Elizur et al., 1991, Furnham et al., 1999, Kalleberg, 1977, Kraut and Ronen, 1975 and Lyons et al., 2010). Many authors have noted that the plethora of work values, measured by different instruments, makes comparisons across research studies on work values difficult and limits understanding of the construct of work values (e.g., Furnham et al., 2005 and Roe and Ester, 1999). Furthermore, the work values literature lacks consensus on the domains comprising the construct of work values as no two measures perfectly overlap. To further the confusion, many investigators develop their own measures of work values to use in their research. 1.6. Research to clarify the construct of work values Little work has been completed to summarize and organize different conceptualizations of work values. Making comparisons between the MIQ (Rounds et al., 1981), Ronen's classification of Hofstede's (1980) work values (Ronen, 1994), and Super's work orientations from the Work Importance Study (Super & Sverko, 1995), Rounds and Armstrong (2005; see Table 13.4) illustrate that there were likely five work values being captured across these three classifications. These values include the importance of achievement/self-actualization, autonomy, power or status, social relationships, and the work environment (e.g. job security, benefits, physical workspace). Berings and colleagues (see Table 1 in Berings et al., 2004) also discuss the similarities across different work values instruments. Their examination of several work values assessments also suggests that the domains within the construct of work values may be limited to a few domains. In their summary of the similarities between Berings (2002) 12 work values, the MIQ (Rounds et al., 1981), the Super's Work Values Inventory (Super, 1973), the Work Importance Study Values Scale (Nevill & Super, 1986), the Customer Service Questionnaire (Saville & Holdsworth, 1992), and the Values Survey (Schwartz, 1992), Berings et al. concluded that essentially six values were captured across nearly all instruments. Similar to Rounds and Armstrong's (2005) conclusions, these values included the importance of independence, creativity, coworker relationships, achievement, earnings, and security. While the efforts of Rounds and Armstrong, 2005 and Berings et al., 2004 demonstrate some progress to consolidate the various conceptualizations of work values, these attempts encompass only anecdotal information about the domains represented within the construct of work values. Only one study has attempted empirical comparison of work values across measures to provide evidence of validity for measures of work values (Macnab & Fitzsimmons, 1987). Macnab and Fitzsimmons compared similarly named scales on four different measures of work values including the MIQ (Rounds et al., 1981), the Work Values Inventory (Super, 1970), the Values Scale (Super & Nevill, 1986), and the Work Aspect Preference Scale (Pryor, 1981), finding support for eight separate work values shared across instruments—authority, co-workers, creativity, independence, security, altruism, work conditions, and prestige. Examining evidence of convergent validity using both correlational and confirmatory factor analytic methods, they found that more variance was explained by differences in traits (i.e., different values) than by methods (i.e. different instruments), suggesting that the measures of work values were all measuring the same traits. Results from Macnab and Fitzsimmons (1987) research added valuable information to the literature by illustrating consensus exists on the domains represented within the construct of work values given the overlap in conceptualizations of work values across measures. Despite being a much needed addition to the research on work values, Macnab and Fitzsimmons' research was not without limitations. While the MIQ and the Work Values Inventory are still widely used, the use of the Values Scale and the Work Aspect Preference Scale has diminished making the usefulness of Macnab and Fitzsimmons research less relevant to the understanding of work values today. Moreover, their sample of undergraduate students in Canada was nearly 80% female. Finally, Macnab and Fitzsimmons' work did not contribute further understanding of the larger domains represented within the construct of work values. With no additional research conducted on the content validity for the construct of work values using multiple instruments, the generalizability of Macnab and Fitzsimmons conclusions are limited. 1.7. Establishing validity Cronbach and Meehl (1955) illustrated the steps in establishing evidence of validity for a construct, which includes establishing the content of the construct and understanding the different domains that comprise the construct. Both theoretical and empirical methods can be used to determine the domains of the construct. A theoretical method of construct definition is employed when expert judges determine what domains are represented in a construct (AERA, APA, & NCME, 1999). However, empirical methods also can be used to determine what domains relate to a construct using statistical procedures. Factor analyses and correlation matrices often are employed for this purpose and help to develop the nomological net for a construct. While methods, such as a multi-trait, multi-method matrix design (MTMM), can be employed to examine construct validity, the use of a MTMM design requires some a priori assumptions as to the traits to be examined across methods (Campbell & Fiske, 1959). Additionally, a MTMM procedure provides limited understanding of how differently named domains (e.g. traits) on different measures may be related. In contrast, using intercorrelations, factor analytic methods using multiple traits and methods between multiple measures can determine the common traits being measured, without making assumptions as to what traits are related prior to analyses. During the process of factor analysis, the factors extracted from the original measures might explain the essential traits of a construct across measures. Thus, using factor analytic methods on multiple measures of work values may be able to provide evidence of construct validity for work values in the hopes of advancing the quality and quantity of research on work values. 1.8. The present study Despite the importance of work values in vocational psychology, research on further understanding of the construct has been limited by the conceptual difficulty in defining the concept of work values (Rounds and Armstrong, 2005 and Zytowski, 1994). The purpose of the current study was to examine the operationalization of the construct of work values across different work value measures and to assess the shared domains of work values across instruments. Principal Component Analysis (PCA) of the items of a variety of work values measures was used to identify the domains of work values that were shared among measures and those that were captured in only one or two measures. Measures for the study – the MIQ, SWVI-R, and MWVI (1972) – were selected based on their acceptance in the scientific literature and their psychometric properties. Although other measures of work values exist, such as Pryor's (1982) Work Aspects Preference Scale, Ronen's Taxonomy of Needs (Ronen, Kraut, Lingoes, & Aranya, 1979), or Berings (2002) 12 work values, they were not selected because of their low frequency of use, inconsistent support of psychometric soundness, and/or lack of use with samples from the United States.
نتیجه گیری انگلیسی
Examination of the correlations between work values measures (MIQ, MWVI, and SWVI-R) offered support for evidence of convergent and discriminant validity for scale scores on each measure based on higher correlations between similarly themed scale scores (e.g. MIQ Ability Utilization, MWVI Intrinsic Rewards, and SWVI-R Challenge) versus lower correlations between differently themed scale scores. The pattern of correlations provides evidence of construct validity, as described by Campbell and Fiske's (1959), for these three measures. Further exploration of correlations between similarly named scales suggests some error due to use of different instruments to assess similar values (e.g., MIQ versus SWVI-R). For instance, the correlation between the MIQ scale of Creativity and the MIQ scale of Responsibility (r = 0.69) was higher than the correlation between MIQ Creativity and SWVI-R Creativity (r = 0.52). However, MIQ Creativity does appear to be most related to the scale of SWVI-R Creativity among SWVI-R scales. In sum, while similarly themed values scales across measures are more highly related than dissimilar scales, most scales have some higher correlations with dissimilar scales on the same instrument, suggesting measurement bias for scales to be more highly related when measured with the same instrument regardless of the content of the scale. Results from the PCA provided further evidence of validity for the construct of work values. The six value components comprised of scales from the MIQ, MWVI, and SWVI-R indicated that some similar values were being measured across instruments, namely values related to the importance to the components of Competence and Status. Other values related to domains such as Organizational Culture, the Environment, Autonomy, and Relationships were contained on one measure, suggestive of unique information to these measures. Research by Elizur (1984) produced similar components within the construct of work values, providing support for the current results. Using a sample of Israeli adults collected in the late 1970's, Elizur found that 21 work values, based on a measure developed using items from Jurgensen (1978), formed five components established from multidimensional scaling. The first of these related to instrumental values, such as working conditions. Instrumental values refer to desirable modes of behavior, such as being honest or polite. To some extent, this reflects items in the Environment component. Secondly, Elizur found another domain represented within work values that pertained to affective values. These values dealt with emotional experiences, such as relationships with others and self-esteem, and are similar to the values within the Relationships component which included the importance of friendly coworkers and helping others. The Status component comprised of items related to income, prestige, and status, is akin to Elizur's third component that captured the importance of rewards, like pay and advancement opportunities. The Organizational Culture component also resembles a component reflecting values related to resources (e.g. supervisors, security) found by Elizur. Finally, Elizur found a component related to cognitive values, such as opportunities for independence, using one's abilities, and having autonomy, which is similar to both the Competence and Autonomy components in the present study. Comparatively, results of current research confirm the components found by Elizur. Furthermore, research has confirmed Elizur's organization of work values domains (Borg, 1986 and Elizur and Sagie, 1999) and provided evidence of this classification across cultures, including the United States (Elizur et al., 1991). Present results were also similar to results from Macnab and Fitzsimmons (1987), who found support for the values of authority, co-workers, creativity, independence, security, altruism, work conditions, and prestige. However, given the use of PCA, the current results extended this information finding that the values supported by Macnab and Fitsimmons could be further consolidated into six underlying value domains. Prior research may provide some justification of the work values components identified. For example, Self-Determination Theory (Deci & Ryan, 2000) posits that individuals have three basic needs – autonomy, competence, and belongingness – that motivate behavior. Self-Determination Theory has been applied to work context findings that meeting these needs is related to job satisfaction (Ilardi, Leone, Kasser, & Ryan, 1993), which supports TWA's (Dawis, 2005) assumption that work values, when reinforced, are the main source of job satisfaction and may explain the findings of work values related to autonomy, competence, and relationships in the current study. Moreover there is some evidence that these needs are related (Deci et al., 2001 and Vansteenkiste et al., 2007) which may explain why Autonomy and Competence values were found to be significantly related. Research on factors related to job satisfaction provides support for the remaining components of Status, Organizational Culture, and Environment. Regarding the component of Status, Jones (2005) summarizes the trend between occupational prestige and job satisfaction, concluding that prestige is related to job satisfaction within occupations. Further evidence by Weaver (1977) found that occupational prestige partially explains differences in job satisfaction between different occupations. Moreover, having advancement opportunities is also related to job satisfaction (Kalleberg, 1977) and supports the inclusion of a domain pertaining to power or status as an essential work value. Numerous other studies have found that perceptions of organizational culture are also related to job satisfaction (Friedlander and Margulies, 1969, Johnson and McIntye, 1998, Kalleberg, 1977, Ostroff, 1993, Pritchard and Karasick, 1973 and Schneider and Snyder, 1975), which provides some insight as to why the component of Organizational Culture emerged from the data. Finally, the importance of the work environment has been shown to be related to increased organizational commitment (Meyer et al., 1998), which could partially explain why the reinforcement of values has been shown to predict job tenure (Hesketh, McLachlan, & Gardner, 1992). 4.1. Implications of the results A theme running through discussions on work values for the past 40 years has been the concern that research, on the relationship between work values and other vocational constructs, has been hindered by the inability of the field to define the construct of work values (Kinnane and Suziedelis, 1962 and Rounds and Armstrong, 2005). Few empirical endeavors have been made to examine how various conceptualizations, or instruments, of work values compare as evidenced by only one prior study approaching this issue (Macnab & Fitzsimmons, 1987). The results established that six values components were being captured across work values instruments. Based on this finding, further revision and development of work values assessment should incorporate these components in the hopes of bringing continuity to this area of scholarship. 4.1.1. Assessment and applications of work values These results also provide some guidance in evaluating the more widely used measures of work values, namely the MIQ and SWVI-R. Findings suggest that neither of these measures fully captured the domains within the construct. Comparatively, the MIQ was found to provide a more comprehensive assessment of work values, as MIQ scales loaded on every component with the exception of the importance of the environment, which was comprised solely with scales on the SWVI-R. Based on these findings, for practitioners and researchers looking for guidance in selecting a measure of work values, the MIQ appears to be the most inclusive assessment, which has been also suggested by others (Rounds, 1990). However, revision of current work values assessments or development of new assessments, in light of the current results, may produce instruments with greater evidence of construct validity. 4.1.2. Importance for individuals and organizations For individuals, clarifying the construct of work values can promote assessment of work values to assist in career development. For instance, while interests have been shown to be predictive of career choice (Borgen, 1972 and Hansen and Dik, 2005), work values have been also associated with career choice (Kalleberg and Stark, 1993, Young, 1984 and Zytowski, 1994). Thus, it may be that work values provide additional information, beyond interests, to assist individuals to further narrow career options. Furthermore, as Rounds (1990) discovered, work value correspondence (e.g. the match between an individual's values and the values reinforced by the job) is a stronger predictor of future job satisfaction than interest congruence. If one of the ultimate goals of career counseling is to assist individuals with identifying occupations to maximize satisfaction (Dawis and Lofquist, 1984 and Super, 1970), then both career professionals and clients could benefit from incorporating assessment of work values into career interventions. A more comprehensive understanding of the construct of work values is also important to organizations. Prior research suggests that work values are related to work performance (Judge et al., 2001, Swenson and Herche, 1994 and Vora, 1983), likely because work values are related to job satisfaction (Dawis, 2002), which, in turn is related to work performance. Therefore, organizations may derive benefit from understanding employee values in the service of creating more efficacious interventions to boost employee satisfaction. Moreover, both voluntary and involuntary job turnover cost organizations an estimated $11 billion a year (Ivancevich, 1998). Abbasi and Hollman (2000) note that a common cause for turnover is a poor fit between the employee's needs and the work environment. Organizations can use information on employee work values to develop effective programs to assist in retaining workers. Additionally, given increases in the rates of unemployment, organizations have more applicants to choose from, making selection a daunting task. Examining work values of job applicants and the values reinforced in the work environment can provide data to assist selecting the best fit for the organization, and reduce costs associated with turnover. Furthermore, current data on work values of younger workers provide organizations some insight into the possibly different needs of this cohort of workers. Again, due to the connection between values, job satisfaction, and tenure, organizations may benefit from trying to understand the values of younger workers. 4.2. Limitations Some limitations of the present study may have had an impact on the results. Primarily, the current sample consisted of young adults, which may limit the generalizability of the data to middle age or older adult populations. Although it is important to understand the work values of young adults, as this typically is a critical period in one's life span for career decision-making, existing literature suggests that work values may be influenced by age (Cherrington et al., 1979 and Singer and Stefflre, 1954), meaning that the current results may not extend to adult populations. The sample also was predominately of European descent, and mostly female. Furthermore, while the majority of the sample was currently employed and nearly all participants had been employed at some point, values have been shown to be influenced by work experience (Pinfield, 1984). Thus, having fewer years of work experience could have potentially influenced the values endorsed by the current sample. While these sample characteristics further limit the generalizability of the findings, the value of pursing this research outweighs current sample limitations. Future research on adult working samples, with more diversity in participant characteristics, will be able to provide replication and validation to the current findings. 4.3. Future research The current study provides much needed consolidation and validation of the construct of work values. In spite of this, more work is needed to fully support evidence of validity for the construct of work values. For example, some research suggests that the construct of work values is closely related to vocational interests (Kinnane and Suziedelis, 1962 and Smith and Campbell, 2009; Super, 1962) and personality (Berings et al., 2004, Furnham et al., 1999 and Furnham et al., 2002). However, other research suggests that work values have little overlap with interests (Breme and Cockriel, 1975 and Rottinghaus and Zytowski, 2006) or personality (Furnham et al., 2005). More consistent results on the relationship between work values and personal values suggest that these constructs are highly related (Kinnane and Gaubinger, 1963, Roe and Ester, 1999 and Ros et al., 1999). Closer examination of these studies shows that inconsistent findings may relate to the use of different measures of work values, further suggestive of little agreement on the domains within the construct of work values. Future research can build on the information from the present study that identified the domains represented within the construct of work values to examine additional evidence of construct validity given relationships between work values and other vocational constructs, such as interests and personality. Despite understanding the similarity of work values covered by various instruments, both Brown, 1996 and Rounds and Armstrong, 2005 question if current measures of work values capture the breadth of work values and hypothesize that there may be other work values that have yet to be identified. Likewise, Nord and colleagues contend that most conceptualizations of work values are deficient because they lack inclusion of other values that are likely important in a work context, such as customer relationships or spirituality (Nord, Brief, Atieh, & Doherty, 1990). Furthermore, values not previously assumed to be that important to work, may now be more relevant to the workplace for younger workers, such as work-family accommodations, availability of leisure time, or causal work environments, as suggested in popular media (Armour, 2005; Kelly Services, 2005 and Lancaster and Stillman, 2002) and empirical studies (Catalyst, 2001 and Twenge et al., 2010). This raises questions for studying the adequacy of the item pools to cover the entire domain of work values and examining whether these items cluster by existing values or represent additional values. Adding to the argument for examination of additional domains in the construct of work values is the reality that the majority of work values assessments were constructed decades ago. More contemporary research concludes that more recent generations of workers may have different values than older generations of workers (Jurkiewicz, 2000, Lyons et al., 2005, Smola and Sutton, 2002 and Twenge et al., 2010) which also may suggest that other values may no longer be related to work. Overall, current discussions of work values imply that work values instruments may not have changed whereas the domain of work values may have changed. Thus, future research may be able to explore if additional domains, beyond those identified in the present study are relevant to the modern construct of work values. In sum, this study was an empirical investigation of the content validity for the construct of work values. Results support current conceptualizations of work values across the instruments examined. Based on the measures included in the present study, the construct of work values includes six domains including the importance of the work environment, opportunities to develop competence, earnings and opportunities for status, autonomy, organizational culture and policies, and work relationships, which were similar to domains identified by others (Borg, 1986, Elizur, 1984, Elizur et al., 1991 and Elizur and Sagie, 1999). While the current study provided much needed examination of work values, future research can continue to develop our understanding of the construct.