, PLEASE USE THIS DOCUMENT AS A GUIDE ONLY
Question 1
1. Which specific type of probability sampling method is most appropriate for Sarah’s study?
Give reasons for your answer.
1. The Core Requirement: Representation of Key Subgroups
Sarah's research is not just about any unemployed youth; she is specifically interested in how key
variables like age group, gender, and level of education relate to (or 'drive') youth unemployment. To
understand these relationships, her sample must accurately represent the different categories within
these variables.
A stratified random sampling method is the most effective way to ensure this representation.
According to the HMDVA81 Study Guide, this technique involves "selecting units from each layer
or stratum in an organisation" (HMDVA81, Study Guide, p. 46). The key advantage of this method
is that it "can be used to ensure the adequate representation of minority subgroups of interest"
(HMDVA81, Study Guide, p. 46). In a simple random sample, these subgroups might be
underrepresented or even missing, which would skew the results of her statistical analysis.
2. Why Other Probability Sampling Methods are Less Appropriate
To understand why stratified random sampling is the best choice, it is useful to compare it to the
other probability sampling methods mentioned in the guide.
Simple Random Sampling: This method involves "drawing names from a hat or assigning a
number to each unit in the sampling frame and then using a random number generator to select
the units" (HMDVA81, Study Guide, p. 46). While it is a valid probability method, it is most
useful "when the sampling frame is small and homogeneous" (HMDVA81, Study Guide, p. 46).
Sarah's population of unemployed youth in a whole province is neither small nor homogeneous.
Using simple random sampling would be a gamble, as it might not capture enough young
women, or enough individuals from specific age brackets or education levels, to allow for
meaningful statistical comparisons between these groups. This would directly undermine her
aim to investigate the impact of these factors.
Systematic Random Sampling: This technique involves "selecting every 10th element from the
sampling frame" (HMDVA81, Study Guide, p. 46). The key issue here is that it requires a
comprehensive and ordered list of all unemployed youth aged 18-30 in the North West
province. Such a comprehensive sampling frame is highly unlikely to exist. Even if a list
existed, there is a risk of periodicity, where the list's order might contain a hidden pattern that
could bias the sample (e.g., if every 10th person on a list of households were all from a similar
socio-economic background). Therefore, this method is not practical for Sarah's context.
Cluster Sampling: This is a "two-stage technique where the researcher first samples a
geographical area, and in the second stage samples units within the area" (HMDVA81, Study
Guide, p. 46). While this is more practical for large areas, it is often chosen to reduce costs and
logistical difficulties, not to maximize representation of key demographic variables. The guide
clarifies that "unlike strata that are homogenous, clusters are heterogeneous" (HMDVA81,
Study Guide, p. 46). If Sarah used this method, she might end up with a sample heavily
concentrated in a few specific towns or areas. This could make it difficult to generalise her
findings to the entire province and, crucially, would not guarantee the representation of her
chosen subgroups (e.g., all gender and age groups).
Question 1
1. Which specific type of probability sampling method is most appropriate for Sarah’s study?
Give reasons for your answer.
1. The Core Requirement: Representation of Key Subgroups
Sarah's research is not just about any unemployed youth; she is specifically interested in how key
variables like age group, gender, and level of education relate to (or 'drive') youth unemployment. To
understand these relationships, her sample must accurately represent the different categories within
these variables.
A stratified random sampling method is the most effective way to ensure this representation.
According to the HMDVA81 Study Guide, this technique involves "selecting units from each layer
or stratum in an organisation" (HMDVA81, Study Guide, p. 46). The key advantage of this method
is that it "can be used to ensure the adequate representation of minority subgroups of interest"
(HMDVA81, Study Guide, p. 46). In a simple random sample, these subgroups might be
underrepresented or even missing, which would skew the results of her statistical analysis.
2. Why Other Probability Sampling Methods are Less Appropriate
To understand why stratified random sampling is the best choice, it is useful to compare it to the
other probability sampling methods mentioned in the guide.
Simple Random Sampling: This method involves "drawing names from a hat or assigning a
number to each unit in the sampling frame and then using a random number generator to select
the units" (HMDVA81, Study Guide, p. 46). While it is a valid probability method, it is most
useful "when the sampling frame is small and homogeneous" (HMDVA81, Study Guide, p. 46).
Sarah's population of unemployed youth in a whole province is neither small nor homogeneous.
Using simple random sampling would be a gamble, as it might not capture enough young
women, or enough individuals from specific age brackets or education levels, to allow for
meaningful statistical comparisons between these groups. This would directly undermine her
aim to investigate the impact of these factors.
Systematic Random Sampling: This technique involves "selecting every 10th element from the
sampling frame" (HMDVA81, Study Guide, p. 46). The key issue here is that it requires a
comprehensive and ordered list of all unemployed youth aged 18-30 in the North West
province. Such a comprehensive sampling frame is highly unlikely to exist. Even if a list
existed, there is a risk of periodicity, where the list's order might contain a hidden pattern that
could bias the sample (e.g., if every 10th person on a list of households were all from a similar
socio-economic background). Therefore, this method is not practical for Sarah's context.
Cluster Sampling: This is a "two-stage technique where the researcher first samples a
geographical area, and in the second stage samples units within the area" (HMDVA81, Study
Guide, p. 46). While this is more practical for large areas, it is often chosen to reduce costs and
logistical difficulties, not to maximize representation of key demographic variables. The guide
clarifies that "unlike strata that are homogenous, clusters are heterogeneous" (HMDVA81,
Study Guide, p. 46). If Sarah used this method, she might end up with a sample heavily
concentrated in a few specific towns or areas. This could make it difficult to generalise her
findings to the entire province and, crucially, would not guarantee the representation of her
chosen subgroups (e.g., all gender and age groups).
