FIC0154 Basic Comm
Research
Online
Discussion Questions
Reference
materials: Lecture on Sampling and relevant text book chapters
________________________________________________
Question 1
What
is the difference between
probability and nonprobability samples? Explain in your own words and give
examples of each type of sample.
Answers:
Probability sample
- is you selecting according to mathematical guidelines in which the sample is selected using random method and every unit in my population have an equal chance to be picked as part of my sample.
- To get a probability sample is difficult because everyone must be in my population so that everyone has an equal chance to be selected.
- You need a sampling frame for probability sample; if not, you can obtain this sample as there has no equal chance on everyone.
Nonprobability sample
- is you select by not following mathematical guidelines.
- For example, the population is the students in a class, then you pick whoever is present in the class and you do not care of those who are absent.
Type of probability sample ->
1. Simple random - For example, you have a box and you choose from it. It's like a lucky draw.
2. Systematic Random Sampling
- For example, you need to have a complete list of members of your population and then you jump or skip names and choose them. (Eg. No.15, 25, 40, 65, 80...)
- For example, you need to have a complete list of members of your population and then you jump or skip names and choose them. (Eg. No.15, 25, 40, 65, 80...)
- You also need to ensure that the list of members are not arranged in meaning manner (Eg, the girls come first in the list then only boys, results will be having more girls in the sample and this leads to inaccurate results.)
3. Stratified Sample
- You pick sample that represent the proportion and it allows us to manipulate the proportion.
- For example, your population is Malaysian. You don't want only teenagers or girls or boys only in your sample. So, you choose sample that is the representatives of your population. The steps will be started from having the list of Malaysian. Then, you categorize them into different races (For example) such as Chinese, Malay, Indian and Others. Next, you randomly pick 10 members from each of the list so everyone has a chance to be picked.
4. Cluster Sample
- You reduce a huge amount of population into different small groups. It is you cluster them and randomly pick the clustered group and then keep repeating the process until you have randomly picked the numbers of sample you want in a very fast way but still ensuring random.
- For example, your population is Malaysian. First, you need to cluster the population, for example you categorized according to the 14 states of Malaysia, then you close your eyes and randomly pick 4 states out of 14 states and throw the remaining 10 states away. After that, you continue to cluster or re-cluster and randomly pick, for example you re-cluster them into different educational level such as PMR, SPM, Diploma and Degree, then you randomly pick 2 out of 4 educational level. Then you keep re-clustering and randomly pick from the clustered groups and reducing the numbers of Malaysian until the numbers of sample you want in your research.
- You cluster our population with not important categories but not with something like gender or age.
Types of Nonprobabilty Sample ->
1. Available/ Convenience Sample
- You pick anyone who is available or convenient.
- You pick anyone who is available or convenient.
- For example: you want to have sample of teenager, so you go to different colleges or universities and pick the first 15 students at the entrance.
- This is not a probability sample because if you stand at the entrance at 8 a.m. in the morning and pick, how about the students who come for afternoon class? So, it's not random sampling.
2. Volunteer Sample
- It's subjective, who ever volunteer then will have the chance to be in the sample.
- It's subjective, who ever volunteer then will have the chance to be in the sample.
- it's easier and cheaper
3. Purposive Sample
- You pick sample for specific purpose (to suit your purpose)
- For example, you want to conduct a research on shampoo. So, you will go and choose people with hair instead of people that are bald for not wasting time, money and effort. (If you randomly pick, bald people will have chance to be picked but there's no point of picking bald people.)
Question 2
What
are the advantages of choosing a
probability sample in your research?
What
are the advantages of choosing a
nonprobability sample in your research?
Answers:
Advantages of probability sample is
- you can calculate the amount of sampling error of your sample in your research.
- It is a good choice to choose probability sample because it's results from the sample you randomly pick can represent the whole population.
Advantages of nonprobability sample is that
- the sample are easier to pick whereby you pick anyone from the population without concerning on whether all the units are in that population.
- Besides, nonprobability sample requires lower cost and less time to be picked and done. The reason of this is because since you are easier in picking the sample, so there are no need much to be concerned in regards in using money and more time.
Question 3
If
you want to do a random cluster sampling
of 10 people out of this entire class (100 people), how should you go about it?
Answers:
First, you need to have the sampling frame of the entire class which is 100 people. Then, you cluster the population into 2 groups or categories which one of the groups are people who own cars and another one are people who do not own cars. Then, you randomly pick from the two groups or cluster. For example, you randomly picked a cluster that is people who own car(s). After that, you continue to cluster or re-cluster the population into the colors of the cars they own. For example you cluster them into cars of white, black, silver and red color. Then you randomly pick 2 out of the 4 colors of car and throw away the other 2 colors of car. Keep repeating the clustering and randomly picking processes until you pick the last 10 people out of this entire class.
Question 4
If
you want to do a random stratified
sampling of 10 people – 7 females, 3 males – out of this entire class, how
should you go about it?
Answers:
First, you need to obtain the sampling frame of the whole class, which is your population (100 people), then you divide or categorize the sampling frame into gender, which is male and female, Now, you have two lists whereby one of the lists is the male and the other one is the female. After that, you randomly pick 7 female from the female list and 3 male from the male list.
Question 5
What
is ‘sampling error’, in your
understanding?
Answers:
Sampling error is the degree to which the results of the selected sample differ from the whole population. It is the gap between sample results and the population results.
- For example, you want to calculate the average age of the class of 60 students. So, you randomly pick 40 out of 60 students in your population and you calculate and got the average age of the 40 students you have selected. If you calculate the average age of the whole class which is 60 students, the average age that you got may be different from the average age of 40 students. Therefore, there is a gap in between the sample of 40 students and the whole population - sampling error.
- The bigger the difference between the sample results and the population results, the bigger the sampling error and vice-versa.
- With a bigger sample size, you will obtain a results which have smaller sampling error.
Posted by Law Cheng Jing (Eileen) [0309527]
