I have a market research exam in 25 mins. Can someone explain methods of sampling to me?

You have about 10 minutes to memorise the following:





Sampling is simply taking data gained from surveying a small portion of a population and using it to predict certain features about the entire population. We use sampling for several reasons:


(1)to save money,


(2)to save time,


(3)to be more accurate,


(4)to avoid destruction or contamination of a product.





Definitions


Element: unit about which information is sought e.g. individuals, products etc.





Population: All the elements defined prior to selection of the sample. It has four parts: elements, sampling units, extent, and time.


Elements: males or females aged 8 – 80 who have watched an existing TV programme


Time : in the past 30 days.


Units: households


Time: March 2003-03-07





Sampling unit: The element or elements available for selection at some stage of the sampling process, e.g. females 20-40, males 18-24 etc.





Sampling frame: A list that identifies the target population. It is a list of all sampling units available e.g. Telephone Directory. The frame defines the unit. Target population minus operational population = sample gap. Combine lists to minimise sample gap.





Study Population: The group of elements from which the sample is drawn.





Stages Involved In Sampling


Who is to be surveyed?


How many should be surveyed?


How should the sample respondents be chosen?


Screening questions define who should and should not be included.





1.Define the target population


2.Identify the sampling frame


3.Determine sample size: Study objectives, time constraints, cost constraints, data analysis procedures.


4.Select sampling procedure – various techniques available


5. Select the sample





Probability Sampling Designs


Each element in the population has a known chance of being selected in the sample.





a.Simple Random Sampling – Each element in the population has an equal chance of being selected. It is used in all complex probability sample designs. For example, suppose we want a simple random sample of two cases from a population of five:


Five Cases: A, B, C, D, E. Ten Possible Pairs: AB, AC, AD, AE, BC, BD, BE, CD, CE, and DE.


Each combination has a 1/10 chance of being selected. Each case has a 4/10 chance of being selected.








b.Multi-Stage Sampling - Where it would cost too much to send researchers all over the country, multi-stage sampling is used to reduce the cost. The procedure is as follows:


1.Divide the country into dozens of regions.


2.Select three or four areas using random sampling methods.


3.Divide these into dozens of areas again and select three or four again.


Continue dividing and selecting until the areas are small enough to make Random Sampling methods possible.





c.Systematic Sampling – The researcher selects every nth in the frame after a random start, i.e., uniform intervals. Example: a telephone book of 400 pages with 400 listings per page has 160,000 numbers. If we require a sample of 1,000, we take every 160th item and begin with a random number between 1 and 160. There is high likelihood of bias if some unknown pattern occurs in the process.





a.Stratified Sampling – Defined population is divided into mutually exclusive and collectively exhaustive sub-groups. Let’s look at the difference in wages between male and female workers. Across the UK, there are four female workers for every five males. The procedure is:


1.Decide the total sample size (say 1000)


2.Divide this into groups (strata) with the same proportion as the population (400 females and 500 males).


3.Select the 400 females and 500 males using random numbers.


4.Add the results together.


Benefits: We can do analysis within strata with a smaller standard error than that available for the whole sample. Thus, our confidence interval is smaller, too. Most companies use stratified sampling. Example: suppose we are asked to monitor the retail sales of Folger’s coffee. To do this, we want to measure the unit sales level of Folger’s in a sample of stores. What stratification variables should we use? We should first answer the question: What factors contribute to the variability in the variable we intend to measure. (Size of store, day of the week, region of the country)





b.Cluster Sampling - The country is divided into areas as with Multi-Stage Sampling. Researchers are sent to the areas with instructions to interview every person who meets the desired criteria, e.g., every mother with a child aged under one year old.





Designing Cluster Samples


How homogeneous are the clusters?


Shall we seek equal or unequal clusters?


How large a cluster shall we take?


Shall we use a single-stage or multistage cluster?


How large a sample is needed.





c. Area Sampling - is a sampling of areas. Suppose we want to run an in-home test of a new shampoo. We decide to run this test in Birmingham. An accurate list of all households is not available. So we select an area sample as follows:


1. List all city areas in Birmingham.


2. Choose a simple random or systematic sample.


3. Attempt to place the product in all households in the chosen area.





Non-probability Sampling


The selection of the population element is based in some part at the discretion of the researcher or interviewer.





When to use


Procedure satisfactorily meets the sampling objectives


Lower Cost than Probability sampling


When time is limited


Not as much human error as selecting a completely random sample


Total list population not available








Types of Non-probability Sampling





c.Convenience Sampling – Samples are selected on the basis of convenience of the researchers. No sample design is used. Choice left to interviewer. Example: having employees test a new product. Unlikely to be any more than of some help in designing a research project and sample design.





d.Judgement Sampling – Samples are selected on the basis of what some expert thinks these particular sampling units will contribute to the research problem. Used more often in industrial than consumer research. Example: if 80% of business comes from 20% of customers, judgement will take a sample of half of the 20% and half of the 80%. Judgement cancels out sampling errors as long as sub-groups are represented among the 80%.





e.Quota Sampling – Researcher takes explicit steps to obtain a sample that is similar to the population on a specified ‘control’ characteristics. These characteristics might be knowledge of proportions of sub-groups among a population. Interviewers fill in proportional “cells”. Leads to substantial bias that cannot be measured. Students research ACORN types and groups.





Response Errors


Such errors occur when :


Giving an answer that will please the interviewer


Wishing to appear socially acceptable or politically correct


Faulty memory, fatigue or nature of questions


Unfamiliarity with subject matter


Interviewer influence


Avoid with screening questions that prove respondent has enough information


Avoid bias in wording of questions


Test questionnaire in a pilot study


Train interviewers to avoid influencing respondents’ answers





Small Population Samples


Used when researching the effects of prescription drugs among a small number of patients.


Direct screening – Using telephone or in a shopping centre. Waste of interviewer time, slow and expensive.


Two-phase Screening - The first identifies people who qualify and the second gets the required data from them. Use Cluster sampling, omnibus studies or panels.


Stratification – using demographic characteristics to identify areas of concentration then adopting a main technique.


Multiframe Sampling – a list of the desired population is incomplete or not available so use two frames.


High yield Clusters – using Systematic sampling methods to identify clusters by telephone.


Snowballing – having found one qualifying respondent, we ask him or her to identify others to develop a network of contacts. Each network becomes the sampling unit.

I have a market research exam in 25 mins. Can someone explain methods of sampling to me?
You take something. Thats it. Good luck!!
Reply:So, for your high-level course exam, this is your preparation?


Gotta love that!
Reply:oh dear, sampling is simply where you get your sample of interviewies from. so you may be using RDD random digit dialing, or u may be using a database used for more specific cases, such as specific groups i.e for a survey on truck tyres you only want business with trucks. or a survey on radio listening trends amongst teenagers would require a teenage sample. etc. good luck is it for a job? if so which company is it? is it IpsosMori?