POL 2156 Study Guide - Quiz Guide: Stratified Sampling, Sampling Error, Sampling Frame
85 views6 pages
22 Nov 2012
School
Department
Course
Professor
QUIZ 2 NOTES – chapters 6, 7
Sampling
Three factors for idenfying a populaon to sample:
Unit of analysis (individual MPs)
Geographic locaon (Canada)
Time period (1993-2000)
Advantages of sampling:
Ecient
Less expensive
Restricted to a certain meframe
Less data collecon/entry
Can provide accurate esmates within set parameters we are interested in the populaon,
sampling is a means to this end
Representaveness of the sample can be aected by:
Accuracy of the sampling frame
Sample size
The method by which the sample is chosen
Your sampling frame should minimize any ineciency/inaccuracy
Sample size: Larger sample sizes represent more accurately than smaller ones. To determine eecve
sample size we must look at:
The homogeneity of the sample – how similar a populaon is in regards to the variable of
interest. More homogenous, lower sample size.
The number of variables to study – the more complex our number of variables and relaonships,
the larger a sample size needed. Also, if looking at race specically we might need to skew our
samples in a way which doesn’t represent percentages in the actual populaon
The desired degree of accuracy – the researcher can state a margin of error they are willing to
accept
The method of random sampling used – Simple Random Sample, Straed Sample, Cluster
Sample
Simple Random Sampling:
All cases are listed and assigned numbers. Cases are then randomized (through a computer etc.)
unl proper sample is found
Less random/accurate version uses even intervals between cases to select sample group
Straed Sampling:
Breaks populaon into mutually exclusive subgroups and then randomly samples each
MP example could be broken up by party aliaon
Increases homogeneity and reduces sampling error, allows focus on small subgroups
Cluster Sampling:
Divide the populaon into a number of subgroups, then randomly select from these subgroups and
select a few to randomly sample. Sampling error increases with every sample taken, larger sample size
can parally remedy this.
Sampling in Qualitave Research:
It is not possible and also not necessary to sample EVERYONE in order to get accurate ndings
qualitavely.
3 Common Sampling Methods in Qualitave Research:
Purposive – one of the most common strategies, parcipants are selected according to pre-
selected relevant criteria. Sample size depends on objecves, as well as me alloed. Oen
determined on the basis of “theorecal saturaon” – the point where no new info will be added
despite new parcipants. It is most successful when data review and analysis are done in
conjuncon with data collecon.
Quota – somemes considered a type of Purposive sampling. We designate quotas for each
characterisc before selecng parcipants. This allows a focus on specic groups/issues to help
nd relevant informaon. Once quotas are selected, we nd people who t the criteria unl
quota is met.
Snowball – Parcipants use their social networks to refer the researcher to others who may t
the criteria to parcipate in the study. Oen used to nd and recruit “hidden populaons”
What is the dierence between Qual. And Quant. Sampling?
Name + Describe 2 types of sampling techniques used in Quantave research
Qualitave Research Methods:
Elite Interviews – “elite” refers to an individual or group with access to necessary specialized
informaon. In most cases, the size of this group is very small and hard to access. Random sampling
cannot be used, we must include as many members of the elite sample as possible.
Case Studies – Studying in-depth of a specic individual, program, or event. Can be useful to learn more
about a poorly-understood situaon, invesgang how an individual or program changes over me. We
cannot be sure it is generalizable, however.
Report layout:
• Introducon - A raonal for studying the case & who they are and the riding they represent, etc.
• Data Collecon - A descripon of the data you collected - websites, newspapers, etc. Time
period covered etc.
• Paerns or Themes: A detailed descripon of the facts related to the case.
• Subsecon A - Acvies in Riding
• Subsecon B - Acvies in Parliament
• Subsecon C - Acvies in Party if applicable
• Subsecon D - Non-Polician Acvies - Volunteer & Acvism
• Discussion - A discussion of the main paerns you found.