Quantitative And Qualitative Job Insecurity And Its Impact On Well-being In Belgian Banks

Sample size

The population size of the target population, on which a study is based is the first attribute that is important for the determination of the size of a sample that will be required to study that population. Selecting a sample size smaller than what is required will not be able to give a proper representation of the population, whereas, by selecting a sample size higher than is what is required will be responsible for involving greater costs and higher time of data collection. Both of these has to be reduced to the minimum, while conducting a research. Thus, a smaller sample size might give unreliable results while a larger sample size will involve greater cost and time. Hence, selecting the proper sample size is important (Chow et al., 2017).

This study is conducted on 63 Belgian Banks in which the total number of employees are 69,000. 15,000 employees are sampled for the study from the total number of employees which is approximately 21% of the employees in the banks. In order to study a population, the standard margin of error and confidence interval that is accepted is 5% and 95% and according to these standards, in order to study a population of 69,000, a sample of 384 units will be sufficient. Thus, the sample size selected for this study is quite large and such a huge sample size might not have been necessary.

This research is regarding the job insecurity and well-being of the employees in the Belgian Banks. There are in total 63 Belgian Banks in which the total number of employees working are 69,000. From all these employees, 21 percent of the employees were selected to conduct the study. For the purpose of the study, it was important that each of the 63 banks participated in the survey. Thus, the sampling was conducted from all the banks. As in total, 21 percent of the employees were sampled, 21 percent of the employees were selected randomly from each of the banks. Other factors such as gender, age and designation of the employees were not given any priority at the time of the sampling. This type of sampling is known as Stratified Random Sampling. The population in this case was divided into 63 strata, which were the 63 Belgian Banks and 21 percent of the units (employees) were sampled from each of the stratum.

A particular type of probabilistic sampling is the Stratified Random Sampling. This sampling technique has several advantages and disadvantages which are discussed as follows (Shipman, 2014):

Sampling method

Advantages:

• Stratified random sampling is a better method of estimation of the population estimates than simple random sampling. In this sampling technique, the population is classified into several homogeneous strata. Thus, the estimates obtained from the samples are better representation of the population as the samples come from every stratum. The results are thus better even when the sample size from each stratum are small.
• The technique of sampling clearly states that the sampling will be done from every stratum individually. This indicates that the sample considered by this technique is a better representation of the whole population that is the target of the research study that is to be conducted.
• With the help of stratified random sampling, along with obtaining estimates about the population as a whole, estimates for each of the stratum is also obtained.
• The convenience of this this type of sampling is extreme to the administration of certain studies. Several zones can be defined by the administration and the results can be obtained from these several zones separately and the zones can be treated according to the results obtained.
• There can be several studies in which all the units of the population are not convenient to be treated in the same way. For example, a student in the hostel might not be treated in the same way a criminal will be. In case of any studies involving all types of people, this sampling technique has been found to be extremely helpful.

Disadvantages:

• Stratified random sampling unlike simple random sampling can only be conducted if there are no missing units in the selected population. If there are any missing units in the population, this sampling technique will become obsolete in that scenario. It is not always possible to target a population in which all the sampling units will be present without any default. The access to all the population units might be restricted on the basis of several privacy policies and access to these units can be obtained but the procedure will be time consuming. Thus, the method will be extremely inconvenient in those scenarios.
• Each stratum must always be defined in a proper manner so that which sample units come from which population stratum can always be identified. All the population units must belong to unique stratum. There should not be any repetition of the population units in different stratum. For this study, the strata are the Banks. The population is divided into 63 strata and no employees can work parallel in two banks. Thus, no strata can overlap in this scenario.
• In case of a situation, where the sample requirements are altered or exceeded for the purpose of the study, the strata that has been predefined has to be defined again on a fresh note. Introduction of newer strata might also be necessary. In these cases, the whole sampling procedure will have to be redone which is again extremely inconvenient for any research.  

In order to conduct a survey, it is always important to evaluate whether the data collected or the questionnaire prepared is valid or reliable. To test for this reliability and validity of the data, a statistic known as the Cronbach’s alpha is used. The value of this test statistic lies between 0 and 1. The closer the value of the test statistic to 1, the more reliable the data is (Bonett & Wright, 2015).

In this research, the value of the Cronbach’s alpha statistic has been found to be 0.89 for the variable quantitative job insecurity. The value as can be seen is extremely close to 1. Thus, it can be said that the data on the quantitative job insecurity is extremely reliable.

In this research, the value of the Cronbach’s alpha statistic has been found to be 0.87 for the variable qualitative job insecurity. The value as can be seen is extremely close to 1. Thus, it can be said that the data on the qualitative job insecurity is extremely reliable.

In this research, the value of the Cronbach’s alpha statistic has been found to be 0.89 for the variable psychological distress. The value as can be seen is extremely close to 1. Thus, it can be said that the data on the quantitative job insecurity is extremely reliable.

While conducting any research, the demographic profile of the respondents is always collected as data. Whatever the research can be about, it is always important to check the demographic profile to which the respondents belong. In this research also, the study is about the job insecurity and well-being of the workers of the Belgian banks. But in this case also, data on the demography of the respondents have been collected. Thus, it has been observed that there is an extreme importance of the data on the demographic profile to which the respondents belong in every study. Based on the requirements of the study, the population in target are distributed into several groups. This distribution or classification is conducted on the basis of several demographic factors. The most common demographic factors are gender, age and education, the expected effects of which are discussed as follows:

Gender: While filling up any survey form for any research, one question is always found to be common and that is the gender or the sex of the respondents. This particular demographic factor is extremely important for any study. It is always known that the opinions of men and women are different. According to Ingalhalikar et al. (2014), in the brain of women, both the right hemisphere and the left hemisphere are equally balanced in terms of strength but in the brain of a man, the hemisphere to the left is stronger than the hemisphere to the right. This is the major cause of the difference in the opinions across gender. This is why, women can think more emotionally while men can think only practically and no emotions are involved in their thoughts usually.

Age: Another question is usually come across while filling up of survey forms is the age of the respondents. The thoughts of a person change as they grow old. This is because, with the increase in the experience with life, people know and learn a lot of things and based on all these learnings, their thoughts change. Thus, there are expected to be difference in the opinions of a teenager from a middle aged or an older person.

Education: One more question that is most common to be asked in surveys is the education of the respondents. Literacy plays a very important role in the opinions of people. The thoughts of a person who is highly educated is bound to be different from a person who is not that much. Thus, these differences might also be important to notice while conducting a research. Based on these differences, recommendation based on a research will be developed.

This study on the 63 Belgian Banks has been conducted with the help of a questionnaire and thus the design involved in this research is a survey design. The total number of employees in the 63 Belgian Banks is 69,000, which is quite high and it is not possible to conduct the survey on all the employees. Thus sampling was important. This research design of survey sampling has several positive and negative effects for the research. These are discussed as follows (Heeringa, West & Berglund, 2017):

Positive Effects:

• In a survey design, the sample that is selected from the population is highly efficient in representing the total population. Thus, studying the whole population is not necessary anymore. The population parameters can be estimated from the selected sample only. With the help of this design, extraction of data is much more appropriate than any other research design.
• The cost of conducting a survey design is lower than other research designs. In this research design, developing the questionnaire and distributing it is the only cost that is involved. There might be cases where the sample unit is large and thus, in order to get responses from them, incentives can be provided which can be in a very little amount but this will satisfy the motto of data collection. However, involving all these costs is still less than the cost of conducting a personal interview or focus groups and the data might not be reliable as this.
• The distribution of the survey questionnaires can be conducted in various ways. Distribution of the questionnaires can be done via email, local mail or fax. Distributing the questionnaires over the internet has been very common nowadays. In this case, the cost involved in distribution is also minimized. Participants across the globe can be obtained very easily with this distribution technique.
• The results obtained from analysing the data obtained with the help of this research design is usually valid, reliable and the has a good statistical significance.

Negative Effects:

• In case of any emergencies, the survey design cannot be modified. What has been made at the beginning of the research has to be continued to the end.
• In case of questions, answers to which are controversial, the respondent might not be comfortable in answering such question. This will lead the questionnaire to be incomplete and there will be problem at the time of analysing a data with missing values.
• Standardization of the questions is necessary before distributing the questions to the respondents, otherwise the respondents will be facing problems while answering.

References

Bonett, D. G., & Wright, T. A. (2015). Cronbach’s alpha reliability: Interval estimation, hypothesis testing, and sample size planning. Journal of Organizational Behavior, 36(1), 3-15.

Chow, S. C., Shao, J., Wang, H., & Lokhnygina, Y. (2017). Sample size calculations in clinical research. Chapman and Hall/CRC.

Heeringa, S. G., West, B. T., & Berglund, P. A. (2017). Applied survey data analysis. Chapman and Hall/CRC.

Ingalhalikar, M., Smith, A., Parker, D., Satterthwaite, T. D., Elliott, M. A., Ruparel, K., … & Verma, R. (2014). Sex differences in the structural connectome of the human brain. Proceedings of the National Academy of Sciences, 111(2), 823-828.

Shipman, M. D. (2014). The limitations of social research. Routledge