Tuesday, April 17, 2007

Cluster Sampling - Group work


Definition: Cluster sampling is a sampling technique where the entire population is divided into groups, or clusters and a random sample of these clusters are selected. All observations in the selected clusters are included in the sample.

One-stage, two-stage and multi-stage cluster sampling:
The population is divided in non-overlapping groups called clusters. Samples of clusters are selected. The clusters are the primary units of sampling. The members of the clusters are the secondary units.
1. If all the members of each selected cluster are included in our sample (of secondary units), the method is called a one-stage cluster sampling.
2. If we take a random sample of each selected cluster, the method is called a two-stage cluster sampling.
3. The secondary units may themselves be groups of tertiary units, and we carry on into sub sampling tertiary units from the selected secondary unit, etc. this is called a multi-stage cluster sampling scheme.

Examples:
1. Suppose an organization wishes to find out which sports Year 11 students are participating in across Australia. It would be too costly and take too long to survey every student, or even some students from every school. Instead, 100 schools are randomly selected from all over Australia.These schools are considered to be clusters. Then, every Year 11 student in these 100 schools is surveyed. In effect, students in the sample of 100 schools represent all Year 11 students in Australia.

2. Suppose that the Department of Agriculture wishes to investigate the use of pesticides by farmers in England. A cluster sample could be taken by identifying the different counties in England as clusters. A sample of these counties (clusters) would then be chosen at random, so all farmers in those counties selected would be included in the sample. It can be seen here then that it is easier to visit several farmers in the same county than it is to travel to each farm in a random sample to observe the use of pesticides.

3. For example, in surveying the performance of school children, the country may be divided into areas (which form the primary units), schools within the areas form the secondary units, the classes within the schools form the tertiary units and the children within the classes form the main objects of the study population.

Advantages:
1. Reduced costs,
2. Simplified field work and administration is more convenient.
3. Instead of having a sample scattered over the entire coverage area, the sample is more localized in relatively few centers (clusters).

Disadvantage
1. Less accurate results are often obtained due to higher sampling error than for simple random sampling with the same sample size. In the above example, you might expect to get more accurate estimates from randomly selecting students across all schools than from randomly selecting 100 schools and taking every student in those chosen.
Site reference:

Friday, April 13, 2007

Day 6- Is randomness only of ONE type?

Hello Hemangi,
Good to see you after a long time. Im sure malaysia in all its splendour is keeping you busy not to mention the amazing work that you are doing there. Trust you and chitra are doing great. Do send pictures if you can...Well back to sampling...

Eg 1
A researcher wants to find out the opinion of primary school teachers on the increasing amounts of size and weight of children’s’ schoolbags. She decides to carry out a survey for this purpose. Out of the 1000 primary school teachers in Mumbai, she decides to select 300 teachers randomly for her study.

-- Points of Difference
1. The researcher is conducting a survey
2. The sample is chosen by the simple random method. As she simply selected 300 teachers randomly out of a list of 1000.

Eg 2
A teacher at a university wants to find out whether there is any difference caused in the academic achievements of students after using 2 different kinds of CAI packages - one with linear interactivity and the other with non-linear interactivity. She decides to carry out the study on her own batch of 50 students. Since there are 2 kinds of CAI packages to be implemented, she makes 2 groups in the following way:She lists everyone’s names on chits, and draws one chit out of the lot and writes that name under a column named linear CAI and puts that chit aside. Then she picks another one, this time writing the name under a column named non-linear CAI. She proceeds in this way until all chits have been randomly distributed under the 2 columns and she is ready with 2 groups of students for her study.

-- Points of Difference
1. The research is conducting an experiment where by 2 independent variables have been selected.
2. The sample chosen here is also randomly selected but not by the simple random selection. This looks like a systematic sampling method as she is randomly selecting the names from the lot but systematically alotting those name individually to one group and then to the other.

Ref: http://metca6research.blogspot.com/ Is randomness only of ONE type?

Monday, April 9, 2007

Day 4 - Let's get sampling

Hie,
Sorry for the delay, my net was down and was extremely busy for Easter.
Anways the two differences between Simple Random Sampling and Convenience Sampling are:

Simple Random Sampling:

1. The fact that it is random allows for anybody to have equal chance of being selected.

2. Hence can be generalized to the whole population





Where as Convenience Sampling:

1. It is based on the easy accessibilty of people that can partake in the study which could result in biasness.

2. It cannot be generalised to the whole population but can be generalised to that research study of that particular group.

Tuesday, April 3, 2007

Lets bridge the gap!!

Hi,

I totally agree. Lets bridge this communication gap and get everybody on the same page...

Like Anthony Robbins quotes "To effectively communicate, we must all realise that we are all different in the way we perceive the world and use this understanding to as a guide to our communication with others"
Now if only people listened, it would make a world of difference...

Monday, April 2, 2007

Day 3- GO Measure Practical Economy (Criteria)

Hi again,

I liked the acronym you used to help us remember the 4 important points of criteria of sampling design. "GO Measure Practical Economy"

GO- Goal Orientation

A research study is aimed at a particular goal and has a set of objectives to help reach that goal. To achieve this goal a good sample has to be selected that is accessable and that represents the population (or in case of the researchers discretion may not represent the population). Samples are of different kinds (sampling design). To be able to choose the one design/strategy that best suits the research problem is important. Hence one criteria of sampling design is to be oriented towards the goal of the study.

Measureability

Sample size is important such that the results should be measureable.

Practicality

The selection of a particular sample design should be practical in relation to the research study. Depending on the length of the study, the sample should be chosen on its representativeness and accessability (ref to previous 2 posts) There maybe more than one sample designs that maybe suitable to the study. Hence along with the other criterias, its practicality should be considered.

Economical

The last but definietely not the least, the economical criteria. The funds available for the research study is an important concern for the researcher. Keeping this in mind the researcher must carefully select a sample, if need be a few research assistants depending on the length of the study.

Following these criterian carefully the researcher will be able to conduct a practical study resulting in reliable outcomes.

(i hope im on the right track, although i was initially confused about the measurability criteria. Do correct me if im wrong.)

Ref: http://metca6research.blogspot.com/ Day 3- Criteria of sampling design

Day 2- Accessibility of a sample

Hie,

I have read the post on accessiblity of a sample. I have also referred to the Url that you suggested. Great resource.

3 examples where a researcher might have a problem accessing a sample are:

1. Sensitivity : If a researcher is conducting a study on breast cancer of new mothers and its impact level on child care/feeding. The accessibility to those mothers with breast cancer may be available but because of the trauma and the overwhelming factor of the cancer and being a sensitive issue, those mothers may not be willing to be part of the study.

2. Timing is everything: The sample participants may not have the time to participate in the study, although they are willing to be part of the study ex: lawyers, policemen, journalists etc.

3. Permission: If a sensitive study on a rape victim or child molesting was the topic of research, accessibility to that sample may be a problem as many parents or local guardians may not give permission to those victims who have been under such an ordeal. So not only is it a sensitive issue but a sensitive issue to the fact of being socially and morally disturbing. or

If a study were in the corporate field where certain information were to be disclosed about a new product for the study, companies (sample) may not have the time, may not authorise the persons in charge to be part of the study or may simply treat this as a sensitive issue.

Hence its agreed upon that amongst other factors, accessibility of a sample is important and should be thought of prior to the commencement of the study.

Ref: http://metca6research.blogspot.com/ Day 2-Accessibility of a sample

Friday, March 30, 2007

Day1 - What would the researcher do?

No. Firstly it will be highly impossible for one researcher to take on the task of gathering data on the opinions of every citizen. Secondly s/he will spend years and probably decades to gather data as India is a phenomenally large country. It can be possible if the researcher has a great number of research assistants with him/her then the task of gathering data can be distributed to these assistants. However there is a possibility of threats.

Another alternative: The researcher can choose a sample that represents the entire population then s/he can administer the test on that sample and whatever results gathered from that study can be generalized to the whole population of India. Thus making the possiblity of carrying out this study simpler.

But to gather data on the different opinions of politicians from every citizen of India is a mighty big task, but not generalizable. As every citizen will have a different opinion and is entitled to his/her opinion. Just by gathering opinions from a handful of people cannot generalise that the entire population thinks that way about politicians.