MBM and MIBM Students Assignment 2022

Assignment for MBM and MIBM students.
Artefact topic: Applied Research Design
Please prepare an answer for the following scenario. You need to answer all
three sections. Your answer should fit in the boxes provided. The word
limit is 400 words.
Q. You are working for an insurance company. In the last financial year, you had
110,000 customers but the number of your customers has dropped about 9% this
In this survey, I would use random sampling to formally draw and define random
probability samples on the targeted population of insured people. Consequently, little
is known on the consequences of the survey estimates quality when we use different
sampling frames. For example, the frames that list individuals insured verses their
addresses. Despite this, the initial choice for sampling frame may be based on
different a diverse criteria, there are different results for the quality of the survey
estimates. Researchers hypothesizes that knowing the respondent’s names upfront
would decrease coverage error and nonresponse. This can therefore be accomplished
using person-based sampling frames or alternatively using the augmenting non-person
based frames with the insured names. In the research, we can therefore
systematically compare the designs in the sampling frames in the face-to-face surveys
with the help of the databases from the insurance company (Bell, Bryman,, & Harley, 2018).
The research population will be found from the company’s database and a random
sampling of individual’s names will be given. The research population will be based on
1000 people randomly selected. The use of even the most conservative measures
support my research hypothesis that the presence of the insured names in the
sampling frame would the targeted population response rates , the rates of
cooperation during the survey and the ineligibility rates by an approximately 3.5-7
percentage points. Perhaps, even the accuracy of the results in the population
estimates would increase. The results show that using individual-level data would be
suited in a well maintained augmented sampling frame with people’s specific
information at hand.
Socio-demographic data for respondents:
Age (in the form of age groups 18-24 years old, 25-34 years old, 35-44 years old, 45-
59 years old, 60 years old and older)
Education (primary, incomplete secondary, secondary, specialized secondary,
incomplete higher and higher)
Income
Employment
Family size
Subjective assessment of financial situation
Where the respondent resides
Methodical report, which describes the main parameters of reachability
To build a sample, we take all ranges of landline and mobile phone numbers from the
insurance website. The total capacity of these ranges is more than 100,000 phone
numbers (Coolican, 2017)..
For the survey, we randomly (using a random number sensor) select from this
database 40-45 thousand numbers depending on the season (more in the summer and

year to 100,000. Given the acute situation, you want to understand the reasons for
this and want to investigate whether you are losing more customers in a certain
segment of the market and what might be the reasons for that.
Please answer the following questions:
Section A. Assuming that you want to implement a survey design, describe
the sampling strategy that you will use and explain how you plan to conduct
the study (describe research population, sampling frame, sampling strategy
and execution plan in detail) 40 points
Section B. To incentivise participation in the survey you are offering
participants a 10% reduction for their next annual subscription. What is the
minimum number of people that you need to contact, to give you the required
sample size to be able to make a claim about your customer base with 95%
confidence level and accept a 5% margin of error? Explain the basis of your
decision and the assumptions you made (20 points)
To obtain a percentage error margin of a 90 percent confidence level, it requires a sample size of
about 800 participants. For a confidence level of 95%, the sample size in the research would be
around 1000 people. The margin error acceptable would be 5% in this case. In the insurance
survey, the number of insured is 100,000. A 1% of the number give the appropriate sample size of
1000 participants. But since the margin of error is 5%, the number should be slightly higher in
random sampling. If a higher sampling size is picked, the margin of error should be slightly lower
and the confidence level would be higher (Hennink, Hutter, & Bailey, 2020).. However, in this
case, the margin of error is 5%.
To find the number of sample in the unknown population, firstly
1. Divide the confidence interval by two. Then look at the area in the z-score table. (Za/2)
95%/2= 0.475
2. Error of margin divided by two
5%/2= 0.025
Use the given percentage which is 25%
3. Subtract from 1.
Alternatively, one can use this formula to find the sample population
Sample size= Distribution of 50% / (Margin of error %/ confidence level)) Squared (Lindlof, &
Taylor, 2017).