Case study: Job Satisfaction in the Banking Industry: Or the logistical nightmare of conducting
large scale quantitative research
Job characteristics are believed to have an impact on stress and well-being at work (Karasek&
Theorell, 1990). The demands of the job on the one hand and theextentto which you have control
over your own activities (decision latitude) on the other, are two factors which together define
how stressful a job is. Those jobs which are high demand, but offer limited control, are
considered to be high-strain and carry an increased risk of job dissatisfaction, stress and burnout.
Based on this theoretical framework, the Union of Belgian Banks sent out a research call to
several institutions, with a bidding process based on criteria such as quality of the proposal,
timing, and – above all – budget. The aim of the research was to carry out quantitative research
to measure the relationship between job characteristics and job satisfaction in all Belgian banks
(seeCambréet al., forthcoming). But in order to do this effectively, several methodological issues
needed to be resolved during the research process.
First of all, aresearch consortiumwas selected to conduct the research, or more precisely, the two
highest ranked bidders were asked to jointly undertake the research. This was the outcome of a
political decision by the banks (see also p. 142, ‘Affiliation and conflicts of interest’),since the
employers preferred one partner and the unions (employee representatives) preferred the other.
The two competing research institutes, a private company specialising in stress at work and
theKatholiekeUniversiteitLeuven (Belgium), were required to co-operate and develop a level of
trust in order to conduct the research. For example, both research institutes had different ideas as
to which scale should be used in the questionnaire. They could not just combine the scales or
include both scales, simply because they are supposed to measure the same concept.
Furthermore, this would also make the questionnaire too complex. Therefore, the research
institutes had to combine their knowledge, look for compromises and jointly work on a shared
vision, which is, to say the least, rather time consuming.
A second obstacle that needed to be overcome was thesample(see Chapter 7). Intotal, 69,000
employees work for Belgian banks and it was decided that questioning all employees would be
too complicated and too expensive. Therefore the research committee, consisting of
representatives of the banks, the unions and the research consortium, opted for a cross-sectional
design (p. 45) with a fixed sample of 15,000 employees (roughly 21%; see p. 187 ‘Absolute and
relative sample size’).
In this sample, the small banks were over-represented in order to be able to make conclusions at
the level of each bank. Within each bank, the respondents were selected at random with no
particular quota for gender, age or employee level. In the postal survey (see p.231 ‘Self-
co.
Case study Job Satisfaction in the Banking Industry Or the logisti.pdf
1. Case study: Job Satisfaction in the Banking Industry: Or the logistical nightmare of conducting
large scale quantitative research
Job characteristics are believed to have an impact on stress and well-being at work (Karasek&
Theorell, 1990). The demands of the job on the one hand and theextentto which you have control
over your own activities (decision latitude) on the other, are two factors which together define
how stressful a job is. Those jobs which are high demand, but offer limited control, are
considered to be high-strain and carry an increased risk of job dissatisfaction, stress and burnout.
Based on this theoretical framework, the Union of Belgian Banks sent out a research call to
several institutions, with a bidding process based on criteria such as quality of the proposal,
timing, and – above all – budget. The aim of the research was to carry out quantitative research
to measure the relationship between job characteristics and job satisfaction in all Belgian banks
(seeCambréet al., forthcoming). But in order to do this effectively, several methodological issues
needed to be resolved during the research process.
First of all, aresearch consortiumwas selected to conduct the research, or more precisely, the two
highest ranked bidders were asked to jointly undertake the research. This was the outcome of a
political decision by the banks (see also p. 142, ‘Affiliation and conflicts of interest’),since the
employers preferred one partner and the unions (employee representatives) preferred the other.
The two competing research institutes, a private company specialising in stress at work and
theKatholiekeUniversiteitLeuven (Belgium), were required to co-operate and develop a level of
trust in order to conduct the research. For example, both research institutes had different ideas as
to which scale should be used in the questionnaire. They could not just combine the scales or
include both scales, simply because they are supposed to measure the same concept.
Furthermore, this would also make the questionnaire too complex. Therefore, the research
institutes had to combine their knowledge, look for compromises and jointly work on a shared
vision, which is, to say the least, rather time consuming.
A second obstacle that needed to be overcome was thesample(see Chapter 7). Intotal, 69,000
employees work for Belgian banks and it was decided that questioning all employees would be
too complicated and too expensive. Therefore the research committee, consisting of
representatives of the banks, the unions and the research consortium, opted for a cross-sectional
design (p. 45) with a fixed sample of 15,000 employees (roughly 21%; see p. 187 ‘Absolute and
relative sample size’).
In this sample, the small banks were over-represented in order to be able to make conclusions at
the level of each bank. Within each bank, the respondents were selected at random with no
particular quota for gender, age or employee level. In the postal survey (see p.231 ‘Self-
2. completion questionnaire or postal questionnaire’) several steps were taken to improve the
response rate (see p. 234 ‘Steps to improve response rates to postal questionnaire’; see also
suggestions byDillman, 1983). The survey was based on addresses which had been provided by
the banks (name, language, address) and each employee randomly selected in the sample
received a personalized envelope through regular mail, sent to him/her by the employer. The
completed questionnaire needed to be returned (free of charge) through the internal post within
each bank. This caused two problems: (1) a perceived lack of anonymity, because the employees
received a personalized envelope (p.136, ‘Invasion of privacy’); and (2) potential bias to the
reliability (p.41; p.157) and the response rate (see Key concept 7.5 on p. 189) because the
completed questionnaires were collected by the banks themselves. The researchers were able to
overcome this by communicating clearly in a letter (1) that although the data collection was not
completely anonymous (home address on envelope), the data analysis would be completely
anonymous; and (2) that the completed questionnaires were collected by the bank but were
transferred immediately to the researchers without being opened or read.
The latter already presented variouslogisticalproblems. The researchers had to travel to each
bank to collect the completed questionnaires and due to the fact that in Belgium, part of the
population speak Flemish (Dutch) and part speak French, two versions of the questionnaire
needed to be available and then carefully translated and tested for the accuracy of their
translation (see also Tips and skills, p.488: ‘Translating interview data’). The questionnaires
were sent to the respondents’ home addresses, a French version if the respondent lived in the
French part of Belgium, a Flemish version when living in Flanders. This prompted a series of
angry calls when Flemish people, living in the French part, or vice versa, received a
questionnaire that was not in their native language. Furthermore, Brussels is officially bi-lingual
and, to complicate matters even more, contains many headquarters in which the main language
spoken is ... English! In order to minimise attrition, it was important that these respondents
received a questionnaire in their preferred language. Another logistical issue was the co-
ordination and control of the distributed information. The Belgian banks, who were the research
financers, chose a decentralised way of working, hence organizing a ‘sensibilizationcampaign’
within each bank whereby the researchers had to visit all the banks to explain the theoretical
framework and the outline of the research to representatives of both employers and employees.
Additional initiatives to prompt a higher response rate were taken up by individual banks, or,
more precisely, by some of the banks. The researchers were required to carefully follow-up on
those initiatives implemented by the banks, to ensure that these initiatives remained both neutral
and valid for the research. Some of these initiatives proved difficult to deal with due to the
selective use of information that had been employed (e.g. letters forcing the employees to
participate; or union campaigns to guide certain answers). Hence, the researchers had to be
3. sensitive for the respective organisational cultures, while making sure they kept a neutral
position towards all partners involved in the research.
Once the data collection was completed (response rate of 47.6%), thedata handlingneeded much
attention. A comprehensive check and double-check was conducted on wrong entries, filters,
missing cells ... just to increase the reliability as well as calculating a weight-factor to
compensate for the over-representation of small banks. One issue was the major difference in
response rate between the banks. Due to a strong campaign, some banks reached a response rate
of over 60%, whilst others barely reached 20% because they did nothing to increase the response
rate. To what extent did this bias the reliability? At the request of the financers, a reference group
of Belgian employees needed to be defined, comparable on gender, age, educational level and
employee level which would allow all banks to be compared not only with the other banks, but
also against an overall group of employees outside of the banking sector. However, since the
questionnaire was also the result of a negotiation, some of the items used were new and not
included in previous research on economy-wide employee well-being. Obviously, there existed
no previous data for these new items. As a consequence, these new items could not be compared
with larger populations and therefore only an internal comparison (every bank compared to all
other banks) was chosen. A final issue here was the choice for a minimal threshold for analysis
and reporting. It was decided that every cell needed to contain at least 15 observations in order to
allow for a graphical representation in the analysis or the final report. As a result, for some
smaller banks the final report was not exactly fine-grained, since some of these banks only
contain 25 employees. For these banks, only overall results were presented, whereas for larger
banks the results were split up for gender, age, department, etc.
A final issue occurred whenpresentingthe results. As mentioned above, the language issue is
particularly important in Belgium to the extent that one even has to be concerned with the order
of reporting and presenting (in terms of which language first).
A discussion arose concerning the graphs used in the report: using different axes can result in
different perspectives, despite the fact that, statistically, the results obviously remain the same. In
both figures below, the amount of people with stress is 5, whereas 10 have no stress. So the
appearances can be deceptive.
QUESTIONS
4. Solution
1) The small banks were over-represented in the sample. Afterwards, it was calculated a weight
factor in order to correct for this. However, as we could expect, for some small banks the
response rate was still too small to allow for organizational conclusions. So what was the point
of this over-representation? Discuss its appropriateness in contemporary organizational research
with many small firms. What alternatives could we have been used to allow for conclusions at
organizational level?
The banking sector anywhere in the world is often an industry that tends to oligopoly, ie, there is
little competitive, banks have mechanisms of division of its geographical demand by type of
customer, corporate, individuals, entrepreneurs, among other criteria segmentation. Also in the
banking sector often have few large entities and several other smaller to serve customers who are
not of particular interest to large banks.
The over-representation of small banks questionnaire sought to capture the different realities that
are observable as more cases are examined, while the reality of the big banks is probably more
homogeneous. Small banks in the role of an employee is more flexible, there is more diversity of
situations and organizational dynamics, the big banks are more structure and functions tend to be
better defined, there are more control mechanisms, the picture is better defined.
It may be valid to use the same instrument for both sizes of organization, because what is sought
measure is the satisfaction at work, and if the fact that an organization of the banking sector
determines to some extent this variable, if there specificities industry relevant to influence an
employee's satisfaction.
When making general conclusions, the weighting factor is useful to bridge the differences
between large and small banks. However, I think that although the cuestionari has been the same
for large banks to small, the analysis must necessarily be different for one than for others.
In large banks are more likely to apply heramientas of descriptiba, the inferential statistics even
generate significant statistical modes. However in small banc, the analysis must be
mascualitativo, make assumptions about the content of the responses above val of aggregate
data.
The response rate, in my opinion, should be a variable to consider it, and relate with others such
5. as efforts to increase it (campaign). Perhaps questionnaire should be complemented in small
banks with other tools such as focus groups.
2) Due to a strong campaign, some banks reached a response rate of over 60%, while others
hardly reached 20%. To what extent did this bias affect the reliability of the results? What can
we say about the generalization issue?
I consider the variables and dimensions of the questionnaire should be classified if they are
suitable for generalizations or not. Aspects of the questionnaire are transverse and independent of
the size of the bank, things like working hours, days off, salary, among other must exist.
These separate variables more sensitive to the size of the organization and control mechanisms,
effectiveness or conformity with the union, among others. Generalize variables that lend
themselves to it and make pormenorisado analysis of specific cases of each institution.
The response rate, in my opinion, should be a variable to consider it, and relate with others such
as efforts to increase it (campaign). Perhaps questionnaire should be complemented in small
banks with other tools such as focus groups.