This chapter provides information on the methodology used to conduct this research. It will start by presenting the framework and the hypotheses tested. Furthermore, it provides information about the sample, data collection methods and the questionnaires utilized to develop the research.
Research Framework
The following framework illustrates the relationships tested throughout this study.
Figure 3.1. Research Framework
This framework shows the relationship between perceived employee voice in labor unions and satisfaction with collective bargaining results and employee compliance.
Moreover, it shows the relationship between perceived employee voice and a positive union-management environment. Moreover, it demonstrates the effect of on satisfaction of collective bargaining results and employee compliance. Furthermore, it shows the
Perceived Employee Voice
Positive Union-Management
Environment
Satisfaction with Collective Bargaining
Results
Employee Compliance
26
mediating effect a positive union-management environment has on the relationship between perceived employee voice in labor unions and satisfaction with collective bargaining results and the mediating role of a positive union-management environment between perceived employee voice in labor unions and employee compliance.
Hypotheses
The following hypotheses are proposed:
Hypothesis 1: Perceived employee voice in labor unions will positively affect satisfaction with collective bargaining results.
Hypothesis 2: Perceived employee voice in labor unions will positively affect employee compliance.
Hypothesis 3: Perceived employee voice in labor unions will positively affect a positive union-management environment.
Hypothesis 4: A positive union-management environment will positively affect satisfaction with collective bargaining results.
Hypothesis 5: A positive union-management environment will positively affect employee compliance.
Hypothesis 6: A positive union-management environment will mediate the relationship between perceived employee voice in labor unions and satisfaction with collective bargaining results.
Hypothesis 7: A positive union-management environment will mediate the relationship between perceived employee voice in labor unions and employee compliance.
27
Research Design
Given the importance of labor unions in Honduras, an empirical study was designed to test perceptions of union workers’ of union-management environment in their organization. Moreover, it tested employees’ perceptions of voice in their labor unions and whether these perceptions might provide benefits for both the organization and the union workers themselves.
A quantitative study was designed to test the hypotheses proposed, and union workers were asked to complete a questionnaire related to their perceptions about union-management environment, employee voice in labor unions, satisfaction of collective bargaining results, social desirability and employee compliance. Later, statistical analysis was applied to test the hypotheses and to draw the conclusions.
Research Procedure
The research procedure consists of all the steps that the researcher followed to complete the study. Figure 3.2 provides a description of all the steps that were followed in order to conduct this research.
Literature review was one of the most important steps in the research procedure, as it allowed the researcher to develop a research topic. After the development of the topic, the researcher designed the framework and developed the related hypotheses. After this, investigation on the instruments to be used was made. However, no validated instruments to measure perceptions of employee voice in labor unions, satisfaction of collective bargaining results and employee compliance were found. Because of this, the researcher developed and adapted the instrument to measure the above-mentioned. A pilot test was conducted to ensure the validity and reliability of the instruments.
After the instruments were tested, the data was collected, and after this, analyzed.
Moreover, the researcher presented the findings and provided recommendations for future research.
28 Figure 3.2. Research Procedure
Sample
Initial contact was made with an agricultural multinational organization in Honduras. In order to be granted access to potential respondents, the researcher agreed not to disclose the name of the company.
The company was founded in the northern coasts of Honduras, in 1899. The Honduran labor movement started in 1954, when employees from the company went on a strike that lasted around three months. Nowadays, the company has some of the largest and strongest labor unions in the country (Sanchez, 2012). The company employs around 10,000 people in Honduras, of which 2,110 are permanent union workers.
Review Literature
Develop research topic
Develop and adapt the instrument
Expert review and pilot test
Data Collection
Data Analysis
Present findings, conclusions and recommendations Design framework and develop hypotheses
29
The population of main interest is targeted at permanent union workers. However, the different labor unions from the company are dispersed along the northern coast of the country. Because of this geographic isolation, the researcher only has access to one of the labor unions. Created in 1955, this labor union is the oldest and largest labor union in the company, with 795 permanent union workers.
Sample Profile
For the purpose of this study, a total of 205 questionnaires were distributed, completed and returned by the permanent union workers of SUTRASFCO.
From the 205 respondents, there were 51 (24.9%) females and 154 (75.1%) males, and the ages ranging from 31 – 40 years old had the highest frequency (81 respondents, or 39.5%). Moreover, the majority of the respondents (134, or 65.4%) have a High School Degree education. In this sample, 61 respondents have been working in the company for more than 10-15 years (29.8%) and 61 respondents have been working in the company for more than 15 years (29.8%). In addition, the majority of the respondents have been part of the union for 7 -9 years (32.7%).
Regarding their union participation, 88 of the respondents reported attending union meetings frequently (42.9%) and 90 of the respondents reported voting on union elections frequently (43.9%). However, when it comes to serving on union committees, 78 respondents (or 38.0%) reported only serving occasionally. Moreover, 83 respondents (40.5%) reported having held union office only occasionally.
Table 3.2 lists detailed information on the descriptive statistics of the sample for this study.
30 Table 3.1.
Descriptive Statistics of the Sample (N=205)
Variable Category Frequency Percentage (%)
Gender Female
31
Data Collection
Data was collected through the distribution of paper-and-pencil questionnaires. A labor relations expert, who collaborated with this research, personally contacted the union workers to ask them to fill the questionnaire. Union workers filled the questionnaire during their free time, and the labor relations expert was always near to answer any questions about the questionnaire.
The questionnaires were translated by a qualified expert with a TOEIC score of 990 into Spanish, the native language of the union workers. Later on, expert review was conducted by a qualified Labor Relations expert in Honduras. After this, back translation was conducted, by a qualified expert with a TOEIC score of 925 to make sure that the original meaning from the English version was kept when translated into Spanish.
Data collection was conducted throughout the months of March and April, 2014.
Because the questionnaire used to collect data was designed by the researcher based on literature, a pilot test was designed to ensure the validity and reliability of the questionnaires. The pilot test was conducted on 42 union workers. The study conducted Exploratory Factor Analysis (EFA) to test validity of the measures and Cronbach’s Alpha to measure internal consistency of the measures.
Data Analysis
This study utilized SPSS and SmartPLS as the statistical tools. SPSS was used for descriptive analysis, Exploratory Factor Analysis (EFA), reliability, and correlation.
SmartPLS was used for Confirmatory Factor Analysis (CFA) and model testing.
Descriptive analysis was conducted to analyze the demographics of the respondents in order to reveal the sample profile. Moreover, descriptive analysis was utilized to check for errors, outliers and the distribution of the data. Descriptive Analysis was also used to examine the means and standard deviations of the research variables in this study. EFA was applied to the data collected in order to determine the factor structure of each research construct to ensure construct validity. EFA was also used to run Harman’s one factor test to detect whether the final data had a serious CMV problem. In order to measure the reliability of the scales, the scales’ Cronbach’s Alpha was tested.
Moreover, this study utilized correlation to find initial relationships on the hypotheses
32
and any patterns among the variables. Correlation was also performed to understand the strength of the relationship between the variables.
EFA and CFA were conducted through Measurement Model Validation using the SmartPLS software. CFA was conducted to ensure construct validity of the data. Model testing through Structural Equation Modeling (SEM) using the SmartPLS software was utilized to test the research framework and the relationship among the variables. SEM allows simultaneous testing of all relationships in the research framework. Moreover, SEM allows a better test on mediation.
Instruments and Measurement
A survey questionnaire was designed to collect the data for this research. The questionnaire was adapted from reference literature and questionnaires that were used for similar topics. The questionnaire will be organized into the following parts:
demographics, positive union-management environment, perceived employee voice, satisfaction of collective bargaining results, employee compliance, social desirability and control variables.
Demographics
Participants were asked to provide their age, education, gender, tenure in the organization, tenure in their respective labor union, and level of participation in the labor union in order to better understand the respondents’ profile and the impact these have on the study variables.
Positive Union-Management Environment
Union-management environment is defined by Angle & Perry (1986) as the
“setting in which labor unions and management interact over time”. Positive union-management environment can be defined as how positively employees perceive this interaction. To measure positive union-management environment, this study will utilize Angle & Perry’s ‘Labor-Management Relationship Climate’. The instrument has a reliability of .95, which shows a high-level of reliability. Some items include
“management is reasonable when dealing with unions” and “the union and management
33
are natural enemies”. In total, 25 items are utilized using a 5-point Likert Scale. The respondents were asked to rate their level of agreement with each statement from 1 (strongly disagree) to 5 (strongly agree). A high score represents a positive union-management environment. A low score represents a negative union-union-management environment.
Because the measurement was very long, some items that were not appropriate for the research context were eliminated. The researcher got permission beforehand from the one of the scale’s original authors to delete some of the items. Originally, the scale was described as a uni-dimensional scale. However, Romero and Yeh (2014) did a research of a sample of 92 respondents, and utilized the SPSS software to conduct exploratory factor analysis, it revealed to be multidimensional. So in order to ensure construct validity, EFA was conducted on the original scale. The items with communalities lower than 0.6 were deleted. Another factor analysis was conducted, and after checking the rotated matrix, items with a factor loading below 0.6 were also deleted.
The final factor analysis revealed 2 dimensions with 10 remaining items in the positive union-management environment scale (as shown in Table 3.1).
Table 3.2.
Rotated Component Matrix of Positive Union-Management Environment.
Component
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
34
Perceived Employee Voice in Labor Unions
Employee voice is defined as an opportunity to ‘have a say’ (Townsend et al., 2013, p.4). Perceived employee voice can be defined as employees’ perceptions of whether or not they ‘have a say’ in their work environment. Moreover, literature states most union workers are concerned with issues related to their wages, pension, health benefits, work conditions, working hours, job responsibilities and safety (Budd, Warino,
& Patton, 2004). Utilizing this definition and literature as reference, 7 items were developed to evaluate employees’ perceptions of their voice, utilizing a 5-point Likert scale. Some items include “the labor union provides an opportunity for me to give voice to wage related issues” and “the labor unions provides an opportunity for me to give voice to safety related issues”. The respondents were asked to rate their level of agreement with each statement from 1 (strongly disagree) to 5 (strongly agree). A high score represents a higher level of perceived employee voice. A low score represents a lower level of perceived employee voice.
Satisfaction with Collective Bargaining Results
Aydin and Ceylan (2009) define employee satisfaction as the combination of affective reactions to the differential perceptions of what the employee wants to receive with what he/she actually receives. Utilizing the definition above, satisfaction with collective bargaining results can be described as the differential perceptions of what the employee wants to receive through collective bargaining, with what he/she actually receives. Utilizing this definition and literature as reference, 7 items were developed to evaluate satisfaction of collective bargaining results, utilizing a 5-point Likert scale.
Some items include “I am satisfied with the wage agreements reached during collective bargaining negotiations” and “I am satisfied with the safety issues reached through collective bargaining negotiations”. The respondents were asked to rate their level of agreement with each statement from 1 (strongly disagree) to 5 (strongly agree). A high score represents a higher level of satisfaction with collective bargaining results. A low score represents a lower level of satisfaction with collective bargaining results.
35
Employee Compliance
Podsakoff, MacKenzie, Paine and Bachrach (2000) define employee compliance as employees’ internalization and acceptance of the organization’s rules, regulations and procedures. Acceptance will result in a conscientious adherence to these regulations and procedures, even when no one is observing or monitoring the employee. Utilizing this definition and literature as reference, 5 items were developed to evaluate employee compliance, utilizing a 7-point Likert scale. Some items include “I usually comply with the organization’s regulations” and “I generally conform with the organization’s regulations”. The respondents were asked to rate their level of agreement with each statement from 1 (strongly disagree) to 7 (strongly agree). A high score represents a higher level of employee compliance. A low score represents a lower level of employee compliance.
Social Desirability
Because the researcher feared a social desirability bias related to employee compliance, a validated scale measuring social desirability was added. Social desirable responding is defined as the tendency for participants to present a favorable image of themselves (Van de Mortel, 2008). Moreover, the author argues that socially desirable responding is most likely to occur in responses to socially sensitive questions and that social desirability response bias can affect the validity of an instrument. The social desirability measure was taken from Strahan and Gerbasi (1972), with a Cronbach’s alpha of .85. It is a short version of the Marlowe-Crowne Social Desirability Scale consisting of ten true or false items. The items are either socially desirable but untrue for most people, for instance “I am always willing to admit when I make a mistake”; or very socially undesirable but very common, for instance “I have never being annoyed when people express ideas very different from my own”. To compute a person’s score, the number of true responses was counted after reversing the negatively keyed items. A high score represents high social desirability while a low score represents low social desirability.
36
Control Variables
The demographic variables included in this study were also utilized as control variables: age, education, gender, tenure in the organization, tenure in their respective labor union, and level of participation in the labor union.
Research states that there is a strong relationship between age and satisfaction.
According to DeSantis and Durst (1996), there are two general theories that have been put forth in regards to this relationship. The first theory is that the relationship is best represented by a U-shaped curved. First, satisfaction decreases, however, satisfaction then increases with age. The second theory, well-documented in literature, is that job satisfaction increases with age. Older employees may have a stronger sense of achievement, and might gain esteem just by the virtue of time on the job. As age might affect satisfaction with collective bargaining results, the researcher decided to hold it as a control variable.
Moreover, Bedeian, Ferris and Kacmar (1992) argue that tenure is a consistent and stable predictor of satisfaction. In fact, Lee and Wilbur (1985) argue that as tenure increases, workers are better able to adjust their expectations to the returns that can be provided by their jobs, or in this case, the returns that can be provided by collective bargaining. Because of this, the researcher decided to hold tenure in the organization and tenure in their respective labor unions as control variables.
In addition, there is evidence that gender might affect satisfaction. Numerous studies report a ‘masculinity culture’ well-within labor unions (Cooper, 2006). In fact, Reilly (2013) argues that in order to gain acceptance and avoid marginalization, women might be forced to play ‘accepted’ feminine supporting roles. For instance, Jung, Moon and Hahm (2007) state that women tend to have lower expectations and are generally more satisfied than men. As such, gender might affect satisfaction with collective bargaining results, so the researcher decided to hold it as a control variable.
Likewise, education can have an influence on workers’ satisfaction. DeSantis and Durst (1996) argued that the effects of education on satisfaction are negative. The younger generation, especially those workers who have more formal education, may have lower levels of satisfaction. Since education can have an effect on satisfaction, the researcher decided to hold it as a control variable.
37
Kelloway and Barling (1993) define union participation as the expenditure of time on union’s affairs. Level of participation might affect satisfaction with collective bargaining results, as it has been shown in past research that employees who felt had an effective communication with management might be associated with higher levels of satisfaction (Thornton, 2009). The scale to measure union participation is based on Kelloway and Barling (1993) literature on participation, which states that formal participation measures involve: meeting attendance, voting in union elections, serving on union committees, and holding union office.
As stated before, social desirability will be held as a control variable for employee compliance, as the researcher fears that the tendency to answer questions in a more favorable manner might affect this variable.
Validity and Reliability
The researcher feared that self-report might lead to Common Method Variance (CMV). According to Podsakoff and Organ (1986) CMV happens when measures of two or more variables are collected from the same respondents and an attempt is made to interpret any correlation among them. Because both measures come from the same source, then any defect in that source contaminates both measures. In order to minimize CMV, different scale points were utilized for dependent and independent variables for this study.
According to Podsakoff, Mackenzie and Podsakoff (2003) Harman’s single-factor test is one of the most widely utilized techniques by researchers to address the issue of CMV.
As such, Harman’s single-factor test was conducted, and the largest factor only explained 21.2% of the variance, which shows there is not a serious CMV problem with the measurements.
In addition, because most of the questionnaires used to collect the data for this study were designed by the researcher based on literature review, EFA and CFA were necessary to ensure that the instrument was stable and effectively measured what they were intended to measure. As this was an important step before going through the model assessment process, convergent and discriminant validity were assessed.
The Smart PLS software was utilized to confirm the validity and the reliability of the measurements. According to Hulland (1999) items with a factor loading below 0.5 are
38
not considered valid, therefore they should be dropped. As such, items below 0.5 were deleted. The internal consistency of the indicators for each construct is shown through the composite reliability. According to Hair, Black, Babin and Anderson (2010) the standard score for composite reliability is >0.7. Moreover, according to Fornell and Larcker (1981) the Average Variance Extracted value (AVE) should be greater than 0.5 to indicate good convergent validity.
Convergent Validity
Table 3.3 shows the means, standard deviations, factor loadings, composite
Table 3.3 shows the means, standard deviations, factor loadings, composite