Measurement tools for measuring commitment

The following section is going to provide an explanation of the measurement scale that is going to be adopted for the dependent and independent variables and the rational behind for choosing the particular approach.

Measurement of Commitment to the Organization

Employee commitment to an organization has been defined in a variety of ways which can be composed into three themes: commitment reflecting an affective orientation toward the organization, recognition of costs associated with leaving the organization and moral obligation to remain with an organization (Meyer and Allen, 1997; Fields, 2002). Fields (2002) describes 11 measures of organizational commitment that are widely adopted by researchers in different disciplines (e.g. Bhatnagar, 2007; Jacobs, 2005; Lok and Crawford, 2001; Hopper, 2009; Huang and Hsiao, 2007; Morrison, 2004; Pretorius and Roodt, 2004; Wasti, 2003). They are the Organizational Commitment Questionnaire (OCQ) (Mowday, 1979; Roodt, 1997); Affective, Normative and Continuance Commitment; Psychological Attachment Instrument; Organizational Commitment (Cook and Wall, 1980; Marsden et al. 1993; Jaros et al. 1993); Organizational Commitment Scale; Career Commitment; Commitment to a Parent Company Versus Local Operation and Supervisor-Related Commitment.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

According to Fields (2002), there are three primary issues need to be addressed in measuring organizational commitment: the basis for the commitment, the manifestation of the commitment and the focus of the commitment. Therefore, most of the measures will classify commitment into multiple dimensions and use separate scales for measurement.

In this research, the “Affective, Normative and Continuance Commitment Measures” developed by Allen and Mayer (1990) will be used to assess the level of commitment amongst the research participants to their working organization. There are eight items for measuring each type of commitment. The items using a 7-point Likert-type response scale ranging from “Strongly Disagree” to “Strongly Agree” to indicate participants’ level of commitment. The affective commitment scale is designed to measure the employee’s comfortable feeling in their relationship with the work role and the organization. The normative commitment scale is designed to measure feelings of obligation by employees to remain with an organization. The continuance commitment scale used to measure employee’s desire to stay because of the cost to the employee. As mentioned in Fields (2002), these measures have also been applied to describe commitment to an occupation or profession (e.g. Coleman, 1999; Pare et al. 2007).

The Meyer and Allen’s measurement is chosen because of its apparent advantages. First, numerous studies have reported acceptable internal consistency reliabilities for the three scales. Fields (2002) found that the Cronbach’s alpha coefficients ranged from 0.77 to 0.88 for affective commitment, 0.65 to 0.86 for normative commitment and 0.69 to 0.84 for continuance commitment (e.g. Meyer et al. 1997; 2002; Wasti, 2003. Second, the distinguishable relations between the three commitments have been supported by confirmatory factor analyses (e.g. Allen and Meyer, 1990; Cohen, 1999; Hackett et al., 1994; Wasti, 2003). Third, the generalizability of this instrument is applicable to a wide diversity of occupations and in different countries (e.g. Bhatanger, 2007; Cho and Kwon, 2005; Hsu, 2009; Pare et al. 2007; Wasti, 2003). Fourth, this instrument has shown that organizational commitment is empirically distinct from other work-related constructs / variables such as job satisfaction, intention to leave and job involvement (e.g. Cohen, 1996; Meyer et al. 2002).

Measurement of Job Satisfaction

For decades, organizational researchers have developed numerous measurement scales for measuring job satisfaction. Fields (2002) has provided a brief description about those measurements that are widely adopted by researchers in different disciplines and have obtained acceptable construct validity and reliability. They are the Overall Job Satisfaction, Job Satisfaction Relative to Expectation, Minnesota Satisfaction Questionnaire (MSQ), Job in General Scale, Global Job Satisfaction, Job Satisfaction Survey (JSS), Job Satisfaction Index, Job Perception Scale, Job Diagnostic Scale (JDS), Job Descriptive Index (JDI), Satisfaction with Job Facets, Global Job Satisfaction, Pay Satisfaction Questionnaire (PSQ) and Satisfaction With My Supervisor. These measurements are also used popularly by researchers in different countries and in different disciplines (e.g. Leung et al. 2008; Lok and Crawford, 2001; Morrison, 2004; Huang and Hsiao, 2007; Sang, Ison and Dainty, 2008; Wu, 2009; Griffin et al. 2010)

Among them, it is noted that the measures of job satisfaction may be classified into global measure or facet measure. According to Cranny et al. (1992), a global measure of satisfaction focuses on the overall level of satisfaction with a particular job, whereas a facet measure focus on satisfaction with specific aspects of the job such as pay, supervision, promotion co-worker and so forth. Wright and Bonett (1992) pointed out that facet measures are sometimes averaged together for as an overall measure level of satisfaction. It is also noted that some studies (e.g. Watson and Slack, 1993; Taber and Alliger, 1995) have used measures of both global and specific job facet satisfaction because specific facet satisfaction measures may better reflect changes in relevant situational factors, whereas a global measure may more likely reflect individual differences than responses to specific items (Witt and Nye, 1992).

Apart from this, some researchers have suggested that job satisfaction measures may differ in the extent to which they tap affective satisfaction or cognitive satisfaction (Moorman, 1993). Affective satisfaction is based on an overall positive emotional appraisal of the job and focuses on whether the job evokes a good mood and positive feelings (Moorman, 1993; Fields, 2002). Cognitive satisfaction is based on logical and rational evaluation of the job, such as conditions, opportunities or outcomes (Moorman, 1993). Therefore, it is noted that the job satisfaction measures appear to differ in the degree they reflect with the mix depending on the nature of the items used in the measure.

The Minnesota Satisfaction Questionnaire (MSQ) developed by Weiss et al. (1967) will be used in this research to assess the level of satisfaction amongst the research participants to their working organization. The MSQ consists in two forms, a 100-item long version and a 20-item short form. The long version covers 20 facets and each consists five items, many of which are more specific than most other satisfaction scales (Spector, 1997). On the other hand, the short form still measures 20 facets, but only 20 different items from the long-form MSQ that best represents the 20 facets is adopted i.e. only 1 item is used to measure per facet. In this research, the short from of MSQ, namely the MSQ20 will be used to measure job satisfaction. It is because job satisfaction is not the only variable that the researcher is going to measure in this research; the long version will be too lengthy to be used in the questionnaire survey, which may also affect the response rate.

The factor analysis of the items measure also results in two factors, namely intrinsic and extrinsic satisfaction. Intrinsic satisfaction refers to the nature of job tasks themselves and how people feel about the work they do (Weiss et al. 1967). In this instrument, those of intrinsic nature are – activity, independence, variety, social status, moral values, security, social service, authority, ability utilization, responsibility, creativity and achievement. On the other side, extrinsic satisfaction concerns aspects of work that have little to do with the job tasks or work. Those of extrinsic nature are – supervision-human relations, supervision -technical, company policies and practices, compensation, advancement, working conditions, co-workers and recognition. The questions using a 5-point Likert-type response scale ranging from “Very Dissatisfied” to “Very Satisfied” to indicate participants’ level of satisfaction. Thus, the MSQ20 can determine the degree of job satisfaction in characteristics associated with the task (intrinsic satisfaction), in non-task characteristics of the job (extrinsic satisfaction) and in overall job satisfaction (Weiss et al 1967). Spector (1997) also commented that many researchers prefer the use of short form and combine all the items into a single total score, or compute extrinsic and intrinsic satisfaction subscales from the subsets of items.

The MSQ20 is chosen because of its apparent advantages. First, numerous studies have reported acceptable internal consistency reliabilities for the extrinsic, intrinsic and total scores (Fields, 2002; Spector, 1997). The Coefficient alpha values range from 0.85 to 0.91 (e.g. Hart, 1999; Jacobs, 2005; Mathieu, 1991; Sempane et al. 2002; Wong et al. 1998). The values for intrinsic satisfaction subscale ranged from 0.82 to 0.86 (e.g. Davy et al. 1997; Wong et al. 1998). For extrinsic satisfaction, the value ranged from 0.71 to 0.82 (e.g. Davy et al. 1997; Wong et al. 1998). Second, the MSQ20 offers a global measure of job satisfaction that is applicable to a wide diversity of occupations and in different countries (e.g. Igalens et al. 1999; Jacobs, 2005; Sempane et al. 2002). Third, the MSQ20 can measure both the affective and cognitive nature of job satisfaction. It is noted that most satisfaction measures ask respondents to compare facet of their job to some referent (a cognitive process) but did not ask for judgments and emotions (Fields, 2002). Though MSQ20 are also predominantly cognitive, but some questions are also centered on the emotional reaction to the work. These facets are considered to be especially suitable in this research because the research is about the relationship of micro and emotional issues and individual’s behaviour to the working organization.

Measurement of Length of Employment within the Organization

Length of employment within the organization refers to the length of time that an individual has been associated with the organization. The length of employment within the organization will be measured directly by using a single item question to ask the respondent the number of years that he / she has worked for his / her organization.

Measurement of Propensity to Trust

From literature review, it is noted that the measurement of trust and general attitudes towards others has shown progress over the past few decades. As propensity to trust refers to the general tendency of a person to trust others, therefore not all trust measures are suitable for measuring this component. Colquitt et al. (2007) named several measures that are commonly used to measure the propensity to trust, including the Interpersonal Trust Scale (Rotter, 1967); Faith in People Scale (Rosenberg, 1957); Philosophies of Human Nature Scale (Wrightsman, 1964); Trust Inventory (Harnett and Cummings, 1980), NEO PI-R Agreeableness Scale (Costa and McCrae, 1992) and Trust Propensity Scale (Mayer and Davis, 1999).

In this research, the Trust Propensity Scale developed by Mayer and Davis (1999) is selected for measuring the general tendency of the respondents to trust other parties working in the organization. This scale is an eight-item measure derived from Rotter (1967) trust scale. The items using a 7-point Likert-type response scale ranging from “Strongly Disagree” to “Strongly Agree” to indicate participants’ level of commitment.

The Mayer and Davis (1999) measurement scale is chosen because of its apparent advantages. First, the measurement scale emphasized the expectancies the person has developed that a given behaviour will lead to a specific positive or negative outcome. As Stack (1978) noted, “Each individual has different expectancies for reinforcement in interactions involving trust… After many experiences with different agents in varying situations, as individual builds up generalized expectancies.”. The purpose of this scale is to measure this expectancies. Second, the generalizability of this instrument is applicable to a wide diversity of studies and in different countries (e.g. Chiaburu and Baker, 2005; Gill et al. 2005; Ovaice, 2001). Third, this instrument consists only 8 items, which is relatively short when compared with other measures (such as the Interpersonal Trust Scale and Philosophies of Human Nature Scale). It is modified from the Interpersonal Trust Scale and widely acceptable in the last decades. Choosing a shorter measurement scale is more appropriate because propensity to trust is not the only one independent variable to affect job satisfaction and commitment in this research. Too lengthy of the scale may lead the respondents / readers too heavily focus on this issue. In addition, this can also reduce the time for the respondents to fill the whole questionnaire (since people may have low intention to fill in the questionnaire if it is too lengthy).

On the other hand, when compared with other measurement scales, the internal reliability of this scale is not very satisfactorily. In various studies (e.g. Chiaburu and Baker, 2005; Gill, 2003), the internal reliability of this scale (Cronbach alpha) is between 0.60 and 0.70, which is relatively lower than the normal acceptable standard (Cronbach alpha > 0.70). However, this estimated reliability is still considered as moderately acceptable because it is noted that even if trust can be measured in a perfectly reliable way, trust may still grow and decline over time as fluctuation occurs in time (Mayer and Davis, 1999). Together with this, the second and third trial results of the test-retest reliability in Mayer and Davis (1999) study are rather close to the first trial (range around 0.60), which also indicates that this measurement scale provides a fairly stable measure of propensity to trust over time (Mayer and Davis, 1999).

Measurement of Ingroup Identification

The social identification concept is regarded as a parsimonious and powerful tool that has contributed significantly to the study of group identity issues. However, in looking at some psychological studies (e.g. Hogg et al., 2000), it has found that the direct measurement of ingroup identification is a bit problematic because of the methodological complexities associated with the operationalization of the identification concept (Phua, 2002). This may also be due to the disagreement about the nature of group identity and the number of basic components included (Jackson and Smith, 1999). This view is also supported by other researchers (e.g. Brewer and Miller, 1996; Deaux, 1995) who found that as the pool of knowledge in the area increases, the findings derived from the social identity theory will be inconsistent as there is a mixed support for different key hypotheses developed. In relation to this, researchers have tried to exploit the underlying components and processes of ingroup identification that can be used as measurement such as the inter-group differentiation (Brewer, 1979), the differentiation index (Brown et al. 1986), in-group favourism (Karasawa, 1991). However, they are not appropriate to be used because of several reasons.

First, there is still debate amongst researchers to which scale is better than others, as each has its own strengths and weaknesses. A closer examination of the various scales found that the basis of the measure of identification lies primarily on the specific dimensions of social identity that researchers perceive as theoretically relevant to their particular study (Phua, 2002). For example, Hinkle et al (1989) empirical investigation to establish the identification-differentiation hypothesis has noted that the investigation “does not find a consistent positive relationship between identification and differentiation: rather the relationship varied considerably across intergroup contexts” (Hinkle et al. 1989, pp. 306). Following this reasoning, the question of inconsistency, strength and weakness in the literature about social identification can be related to the different components that are incorporated into each study, which may also be regarded as methodological inconsistency.

Second, there is one important aspect about the consideration of the use of these scales that has not been mentioned in previous studies – the issue of scale validity (Jackson and Smith, 1999). It is agreed that research results that show evidence of construct reliability, convergent validity and discriminant validity can provide a considerable support to the usefulness of a particular scale. However, it has been found that there is only little work done to establish the validity of various scales that measure group identification (Jackson and Smith, 1999; Phua, 2002). While the construct reliability that used to test the internal consistency of items on the scale are being demonstrated in some studies, the test for convergent and discriminant validity are not incorporated.

In contrast to the above mentioned group identification measurement scale, Tropp and Wright (2001) introduced a single item scale that used to measure the ingroup identification in terms of the degree to which the ingroup is included in the self. This expression of ingroup identification is firstly inspired by the work of Aron and his colleagues on close relationships (Aron et al., 1991; 1992). Aron et al (1991) use the term “inclusion of other in the self” to refer to the interconnectedness between self and other in close relationships. Rather than being perceived as separate beings, self and others are regarded as “overlapping selves” (Aaron, 1991, pp. 597) of which the characteristics of the close other are considered as part of one’s self. This concept also ties in with the affective component of social identification which is referred to as a sense of emotional involvement with the group (Tajfel, 1981).

Tropp and Wright (2001) took the view that identification is related to the approaches of assessing the degree to which individuals include in a specific ingroup, which used to recognize the interconnectedness between self and ingroup. This view also underlines all contemporary and emerging views on ingroup identification (Phua, 2004). In applying this approach to the assessment of ingroup identification, the IIS measure is represented diagrammatically by using 7 pairs of circle (Venn-like diagrams) with varying in degree of overlapping in demoting an individual’s identification with a given ingroup (Tropp and Wright, 2001). Respondents are asked to score on the scale that ranges from the first pair of circles (no overlapping – no sense of identification) to the seventh pair of circles (highest overlapping – strongest sense of identification) to represent their level of identification with a given ingroup.

It is also acknowledged that the use of the IIS in this research may be preferable for several reasons. First, the visual representation of the measure’s overlapping circles captures the essence of the inclusion of ingroup in the self rather than relying on many differently worded items in the concept (Tropp and Wright, 2001, pp. 587). This offers an advantage of minimizing the ambiguity and confusion of what constitutes the social identity. Second, in contrast to previous scales (e.g. Brown et al., 1986) of identification where the nature of the group hinges the utility of different dimensions of social identity, the conceptualization of ingroup identification in Tropp and Wright (2001) provides a “more basic and concise means of studying in-group identification across members of a variety of groups”. This conceptualization can also minimize the ambiguity and confusion of what constitutes social identity. Third, this measurement emphasizes the degree to which a specific ingroup is part of the person’s self-representation, not the tendency to define oneself in terms of group membership on a more general level (Tropp and Wright, 2001). This is because some measurement scales (e.g. Cheek et al., 1994) have included items for researchers to assess individuals’ feelings towards their group memberships. At the same time, these measures are also constructed for examining the collective identity and collective self-esteem in the scale rather than just focusing on the personal significance granted to special group membership (Tropp and Wright, 2001). Fourth, as a single-item measure, the IIS can be adapted for a variety of group membership easily and also administrated more quickly than multi-item measures.

Fifth, the statistical criteria for testing the validity and reliability of the scale are also included in Tropp and Wright (2001) study. In contrast with other single-item measurement scales which may be susceptible to validity and reliability problems (Hinkle, 1995), this measurement scale has demonstrated the discriminant and convergent validity, plus stability using test-retest procedures (Tropp and Wright, 2001). The findings in the study have shown that the IIS measure is strongly correlated with other measures of ingroup identity and strongly related to many cognitive and relational variables associated with ingroup identification (Tropp and Wright, 2001, pp. 598). In addition, the test for scale reliability (in form of test-retest reliability) which used to show the correlations between the IIS score for two periods are also highly significant, indicating that the degree to which individuals feel identified to their ingroup varies with time. Last but not least, the utilization of the scale has also shown successful results in current studies, particularly when the scale is intended to be used for measuring ingroup identification in real groups (e.g. Crisp et al. 2009; Sani et al. 2009; Turner et al. 2008).

Demographic Information

Since respondents’ demographic characteristics may affect the questionnaire responses and may be used as alternative explanations for certain results, a demographic section will be included in the questionnaire survey. The demographic section will contain both structured and open-ended items for obtaining background information from the respondents, including age, gender, professional affiliation, level of education, length of time working in the construction industry, organization types and size. Moreover, these demographic data may help the researcher to understand the job satisfaction and commitment level of the respondents, which may also help for data classification and analyses at later stages. In addition, personal information will also be collected at the end of this section for further contact (if necessary) and providing the survey outcomes to the respondents.