This chapter describes the research framework, hypotheses and methodology that are used in the study. It outlines the research procedure, sample, instrument, data collection procedure, and statistical methods for data analysis.
Research Framework
The research framework was developed in accordance with the purpose of the study and reviewed literature. The Research Framework presented in Figure 3.1 shows the overview of the components of this study. Based on the research hypotheses, this study explored the relationship between multitasking job demand and job stress, and revealed how the polychronic individual fit or misfit with multitasking job environment influenced the job stress. All of the variables included in the framework are unidimensional variables.
Figure 3.1. Research framework.
Demographic factors and control variables
Gender Age
Years of working experience Job position
Industry Self-efficacy
Polychronicity
Multitasking job demand
Job stress
H1
H2
26
Research Hypothesis
There were 2 hypothesis developed. These hypotheses are competing, meaning that if Hypothesis 1 is supported, Hypothesis 2 should be rejected, and vice versa if Hypothesis 2 is supported, Hypothesis 1 should be rejected.
H1: Multitasking job demand is positively and significantly related to job stress.
H2: Polychronicity moderates the relationship between multitasking job demand and job stress: when polychronicity is high, high multitasking job demand will result in less job stress; when polychronicity is low, high multitasking job demand will result in more job stress.
Research Procedure
The research procedure included nine steps that were followed to complete the study (See Figure 3.2).
At the very beginning the research problem was identified. Then, the research objectives were identified and defined, and they served as the basis of the current research.
The literature review was conducted in order to get the insights into the topic. Based on the literature review, the framework was developed. The instruments developed by other researchers were adapted in the current study. The pilot study was used in order to analyze the reliability of the instrument, the main study was conducted. After conducting the main study, the collected data was analyzed and hypotheses were tested. Finally, the report of the findings of the study was provided, and the conclusion based on the findings was drawn.
27 Figure 3.2. Research procedure.
Research Design
In the current study, quantitative research method was used in order to identify the relationships between multitasking job demand, job stress and the moderating effect of polychronicity on the relationship between multitasking job demand and job stress. Self-report survey was utilized to obtain the data. Convenience sampling was used, because the target population was employees in different job sectors in the Russian Federation, and no
Conclude Research Findings Analyse Data
Collect Data Run Pilot Study
Construct Research Instruments Construct Framework
Review Literature
Identify Research Objectives
Identify Research Problem
28
sampling frame was available for the population. Before the data collection for the current study, the pilot study was conducted to ensure validity and reliability of the measurement scale in questionnaire. Inferential statistics were applied to test hypotheses.
Research Sample
In order to test hypotheses of this study, convenience sampling was utilized. The targeted population was employees in different job positions in the different organizations in Russia.
The sample consisted of the employees in different job positions in diverse organizational sectors. The respondents were working for companies situated in different cities of the Russian Federation (e.g., Saint-Petersburg, Moscow, Chelyabinsk, Volgograd, Pskov, Perm, etc.). Organizational sectors included banking sphere, medical equipment sales industry, computer industry, wholesale of household products sphere, manufacturing industry, etc. A total number of 451 questionnaires were received, but 49 questionnaires were found to be invalid, because some respondents failed to meet the sample criteria for job location in Russia or current job position (some of the respondents appeared to be students or unemployed). The number of valid questionnaires entered into the analysis was 402.
Data Collection
Data collection process included several ways to distribute questionnaires. First, since online questionnaires were used, the questionnaire link was sent through emails to people that the researcher was familiar with, or to the managers of companies who were responsible of spreading questionnaire’s link within the company. The connections with managers were built through personal contacts. A week after the questionnaire link was sent, the phone or email reminder was done. Second, the questionnaire link was posted in different groups of social networks, such as vk and facebook, so that any working person in Russia had a chance to fill in the questionnaire. Third, hard copies of questionnaires were handed in to some companies in Moscow region, which the researcher had personal contacts with. Hard copies of questionnaires were collected a week after they were handed in to the companies.
29
Questionnaire Design
The questionnaire items included measures of research variables that were chosen from the previous studies. The questionnaire, including variables’ items and demographics, was originally designed in English. Because the target population was Russian employees, the questionnaire was translated from English into Russian. To ensure that questionnaire, translated into Russian, matched the original meaning of the questionnaire’s statements in English, back translation was done by the Russian native speaker, who was working as a Russian-English interpreter. The translated version of questionnaire underwent a pilot study with 31 respondents; the pilot study sample was similar to the sample in the current study.
The pilot study revealed a translation problem for some of the items, so those items were cross-checked and changed. In general, the results of pilot study showed good face validity and internal consistency reliability.
Questionnaire included 5 parts, the first part had the items for job stress, the second part included questions about multitasking job demand, the third part contained polychronicity items, the fourth part had self-efficacy questions and, finally, the fifth part included demographics.
There were two types of questionnaires: online and hard copies. The online questionnaires were made through Google Forms.
Measurement
The research instrument for this study is a questionnaire as shown in the Appendix. It consists of five sections, where the first section includes job stress items, second section contains a multitasking job demand items, the third section comprises polychronicity items , the fourth section includes self-efficacy and the last section includes the measurement for the demographics.
Job Stress
Job stress is defined as an employee’s awareness or feeling of personal dysfunction as a result of perceived conditions or happenings in the working place, or employee’s
30
psychological and physiological reaction, that are caused by workplace environment (Montgomery, Blodgett, & Barnes, 1996).
Job stress questionnaire items were derived from the 12-item General Health Questionnaire by Goldberg (1972). It included 12 items, where 6 of them were reversed-coded. The questions from 12-item General Health Questionnaire were modified into statements. For example the original item was: “Have you been able to concentrate on whatever you are doing?” , and it was modified into “I have been able to concentrate on whatever I am doing.”
Participants were asked to describe their psychological and emotional health over the past few weeks. Table 3.1 shows the measurement items. The Cronbach’s alpha was 0.91 (Hecht & Allen, 2005).
A Likert scale ranging from “1=Strongly Disagree” to “7=Strongly Agree” was utilized.
Table 3.1.
Measurement for Job Stress
Construct Code Questionnaire item
Job Stress JS1 I have been able to concentrate on whatever I am doing. (reversed coded)
JS2 I have felt that I was playing a useful part in things. (reversed coded) JS3 I have felt capable of making decisions about things. (reversed coded) JS4 I have been able to enjoy my normal day-to-day activities. (reversed
coded)
JS5 I have been able to face up to my problems. (reversed coded)
JS6 I have been feeling reasonably happy, all things considered. (reversed coded)
JS7 I have lost much sleep over worry JS8 I have felt constantly under strain
JS9 I have felt that I could not overcome my difficulties JS10 I have been feeling unhappy and depressed
JS11 I have been losing confidence in myself
JS12 I have been thinking of myself as a worthless person
31
Job stress items underwent Confirmatory Factor Analysis in the Amos software to ensure construct validity of the measurement. It was suggested by researchers that CMIN/DF in a ratio between 2 and 5 indicates a reasonable fit (Marsh & Hocevar, 1985), and ratio smaller than 2 are considered to represent a minimally plausible model (Byrne, 1991). As for RMR, the values lower than .08 are thought to be acceptable (Hu & Bentler, 1999), but ideally it should be less than .05 (Stieger, 1990). According to MacCallum, Browne and Sugawara (1996), RMSEA less than .01 is excellent fit, less than .05 is good fit and less than .08 is mediocre fit. In order to indicate a good fit for GFI and AGFI, it should be equal or greater than .90 (Byrne, 1994). The values of acceptable fit indices are shown in Table 3.2.
The results of Confirmatory Factor Analysis for original measurement of job stress are shown in Table 3.3. Therefore, none of the indices in Table 3.3 show good fit, thus, items JS2, JS3, JS4, JS5, JS6, JS7, JS8, and JS9 were deleted.
The results for new measurement model are shown in Table 3.4. CMIN/DF is 1.571, which shows that it is a minimally plausible model, RMR is .025, which is an ideal fit, RMSEA is .038, which is a good fit, GFI and AGFI are higher than .90, these indices show good fit. Thus, the modified model shows good fit and can be used, and a modified model could go through cross validation. Therefore, the new job stress measurement model was tested by cross validation. First, the data set was randomly split into two sample groups, and second, the new model of job stress was tested by the multi-group comparison in IBM SPSS Amos.
Table 3.2.
Summary of Acceptable Fit Indices for Measurement Models
Fit indices Value Interpretations Citations CMIN/DF 2-5
32 Table 3.3.
Summary of Goodness-of-fit for Original Job Stress Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Job stress 9.265 .000 .212 .144 .805 .719
Table 3.4.
Summary of Goodness-of-fit for Modified Job Stress Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Job stress 1.571 .208 .025 .038 .996 .981
The modified job stress measurement model passed the criteria for cross validation, where the structural covariances and measurement weight should have insignificant p values (p>.05), while the p value of the measurement residuals can be less than .05 to ensure validity of the measurement (Little, 1997). Criteria for cross validation are shown in Table 3.5.
Table 3.5.
Summary of Model Comparison for Modified Job Stress Measurement Model
Model DF CMIN P NFI
Multitasking job demand is a job demand that requires “accomplishing multiple-task goals in the same general time period by engaging in frequent switches between individual tasks” (Delbridge, 2000, p.1) or by performing several tasks at a time.
Multitasking job demand questionnaire was derived from Hecht and Allen (2005).
The variable that they measured in their study with these items was called polychronicity supplies, but since the concept of multitasking job demand and polychronicity supplies are the same in the current study, multitasking job demand was measured with polychronicity
33
supply items. Multitasking job demand measurement included 10 items, where 3 of them were reversed coded. Table 3.5 shows the measurement items. The Cronbach’s alpha in that study for polychronicity supplies was 0.81 (Hecht & Allen, 2005).
A Likert scale ranging from “1=Strongly Disagree” to “5=Strongly Agree” was utilized.
Table 3.6.
Measurement for Multitasking Job Demand
Construct Code Questionnaire item
Multitasking Job Demand MJD1 On the job, I am required to complete entire projects every day, rather than completing parts of several projects. (reversed coded)
MJD2 When doing this job, work must be done one thing at a time. (reversed coded)
MJD3 This job requires me to complete one task before starting another. (reversed coded)
MJD4 This job demands that I juggle several activities at the same time.
MJD5 It is typical of this job to have many tasks to complete.
MJD6 On this job, I am required to complete parts of several projects every day, rather than completing an entire project.
MJD7 This job requires people to do many things at once.
MJD8 On the job, I am frequently asked to start new tasks when other tasks have not yet been finished.
MJD9 This job often requires that I spend a little bit of time on several tasks—moving back and forth from one thing to the other.
MJD10 The demands of this job are such that I repeatedly have to switch gears from one task to another.
Multitasking job demand items underwent Confirmatory Factor Analysis in the Amos software to ensure construct validity of the measurement. According to Table 3.2 none of the fit indices for original multitasking job demand measurement model in Table 3.7 show good fit, thus, items MJD2, MJD3, MJD4, MJD5, MJD7 were deleted.
The results for new measurement model are shown in Table 3.8. CMIN/DF is 2.628, which shows a reasonable fit, RMR is .031, which is an ideal fit, RMSEA is .064, which is a mediocre fit, GFI and AGFI are higher than .90, and these indices show good fit. Thus, the
34
modified model shows good fit and can be used, and a modified model could go through cross validation. Therefore, the new multitasking job demand measurement model was tested by cross validation. First, the data set was randomly split into two sample groups, and second, the new model of multitasking job demand was tested by the multi-group comparison in IBM SPSS Amos.
Table 3.7.
Summary of Goodness-of-fit for Original Multitasking Job Demand Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Multitasking job demand 10.529 .000 .113 .154 .828 .729
Table 3.8.
Summary of Goodness-of-fit for Modified Multitasking Job Demand Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Multitasking job demand 2.628 .000 .031 .064 .988 .964
The modified multitasking job demand measurement model passed the criteria for cross validation, where the structural covariances and measurement weight should have insignificant p values (p>.05), while the p value of the measurement residuals can be less than .05 to ensure validity of the measurement (Little, 1997). Criteria for cross validation are shown in Table 3.9.
Table 3.9.
Summary of Model Comparison for Modified Multitasking Job Demand Measurement Model
Model DF CMIN P NFI
Polychronicity is “a non-cognitive variable reflecting an individual’s preference for shifting attention among ongoing tasks” (Poposki & Oswald, 2010, p. 250) within the same period of time or preference for focusing attention on several tasks at the same time.
35
Polychronicity was measured with 7 items taken from Inventory of Polychronic Values (Bluedorn, Kalliath, Strube, & Martin, 1999). The other three items were taken from Hecht and Allen (2005). Polychronicity measurement includes 10 items, where 3 of them were reversed-coded. Table 3.9 shows the measurement items. The items from P1, P2, P3, P6, P7, P8, P10 are items taken from Inventory of Polychronic Values (Bluedorn, Kalliath, Strube, & Martin, 1999), the items P4, P5, P9 are taken from the study by Hecht and Allen (2005). The Cronbach’s alpha for those combined 10 items was 0.88 (Hecht & Allen, 2005).
A Likert scale ranging from “1=Strongly Disagree” to “5=Strongly Agree” was utilized.
Table 3.10.
Measurement for Polychronicity
Construct Code Questionnaire item
Polychronicity P1 I would rather complete an entire project everyday than complete parts of several projects. (reversed coded)
P2 I prefer to do one thing at a time. (reversed coded)
P3 I believe it is best to complete one task before starting another.
(reversed coded)
P4 It is hard for me to start something new, if there are other things I have not finished. (reversed coded)
P5 I find it difficult to switch gears from one task to another.
(reversed coded)
P6 I like to juggle several activities at the same time.
P7 I would rather complete parts of several projects every day than complete an entire project.
P8 I believe people should try to do many things at once
P9 When I have several things to do, I prefer to spend a little bit of time on each—moving back and forth from one thing to the other.
P10 I believe people do their best work when they have many tasks to complete.
Polychronicity items underwent Confirmatory Factor Analysis in the Amos software to ensure construct validity of the measurement. According to Table 3.2 none of the fit indices for original polychronicity measurement model in Table 3.11 show good fit, thus, items P1, P2, P3, P4, P5, P6 were deleted.
36
The results for new measurement model are shown in Table 3.12. CMIN/DF is .887, which shows a minimally plausible model, RMR is .011, which is an ideal fit, RMSEA is .000, which is an excellent fit, GFI and AGFI are higher than .90, these indices show good fit. Thus, the modified model shows good fit and can be used, and a modified model could go through cross validation. Therefore, the new polychronicity measurement model was tested by cross validation. First, the data set was randomly split into two sample groups, and second, the new model of polychronicity was tested by the multi-group comparison in IBM SPSS Amos.
Table 3.11.
Summary of Goodness-of-fit for Original Polychronicity Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Polychronicity 9.497 .000 .080 .146 .851 .765
Table 3.12.
Summary of Goodness-of-fit for Modified Polychronicity Measurement Model
Variable CMIN/DF P RMR RMSEA GFI AGFI
Polychronicity .887 .001 .011 .000 .998 .989
The modified polychronicity measurement model passed the criteria for cross validation, where the structural covariances and measurement weight should have insignificant p values (p>.05), while the p value of the measurement residuals can be less than .05 to ensure validity of the measurement (Little, 1997). Criteria for cross validation are shown in Table 3.13.
Table 3.13.
Summary of Model Comparison for Modified Polychronicity Measurement Model
Model DF CMIN P NFI
37
Demographic Factors
Job location.
The respondents were asked to provide information on the location of their job. It was needed to identify the percentage of people that were working in the following parts of Russian federation: Northern, Southern, Central, Western, and Eastern part. The job location was dummy coded as central part of Russia=1, the rest of the locations=0.
Current job position.
The information on current job position was needed to identify which positions more often have multitasking job demand. The respondents were asked if their current job position is: business owner, office worker, laborer, etc. The current job position was dummy coded as office worker=1, the rest of the positions=0.
Industry.
The respondents were asked to indicate which industry they were currently working in.
It was needed for calculating the percentage of people working in a particular industry. The options were provided (i.e. information technology, wholesale, manufacturing, education etc.). The industry was dummy coded as manufacturing industry=1, the rest of industries=0.
Control Variables
Gender.
Gender might have an effect on job stress (Indik, Seashore, & Slesinger, 1964).
Gender variable was used to identify the percent of the female and male respondents. Gender was measured with one item: What is your gender? The options were dummy-coded as male
= 0, female = 1.
Age.
It was found by some studies that age is related to job stress, that job stress decreases with the age (Indik, Seashore, & Slesinger, 1964). Age data was collected to know the
38
percentage of each age range. Age data was measured with one item: What is your age? This was an open-ended question.
Years of work experience.
Work experience may also have influence on the job stress. It was found that work experience is negatively related to job stress (Hellman, Morrison, & Abramowitz, 1987).
Therefore, the respondents were asked to provide information about the period that he had been working in the current job position.
Self-efficacy.
Self-efficacy refers to an individual’s belief in his capabilities to perform well in a particular situation (Bandura, 1997).
Self-efficacy may have the effect on the person-job fit in the current study. As individuals with high levels of self-efficacy perceive complex tasks as a challenge to overcome rather than a threat to avoid (Bandura, 1989), therefore there may be a fit of efficacious individual in the multitasking job environment. Moreover, it was found that
Self-efficacy may have the effect on the person-job fit in the current study. As individuals with high levels of self-efficacy perceive complex tasks as a challenge to overcome rather than a threat to avoid (Bandura, 1989), therefore there may be a fit of efficacious individual in the multitasking job environment. Moreover, it was found that