Question

Topic: Research/Metrics

How To Estimate Popultion Based On The Sample?

Posted by Anonymous on9/5/2005 at 1:48 PM ET 350 Points
你好,

I need very urgent reply as I have interview tonight. How can we estimate the number of cars in the market? What question or technique should be used if we do not know the population? For .e.g suppose we surveyed 1000 respondents and we came to know that 40% own honda and 50% own toyota and 10% other. Now how can we tell that there are X number of cars in the state or in the country based on our survey? Is there any formula to size the market? Please help. This question is very crucial for my interview. Thank you so much.
To continue reading this question and the solution, sign up ... it's free!
SUBSCRIBE
或,SIGN IN

RESPONSES

  • Posted on9/5/2005 at 1:59 PM Member
    You've got your answer right there, marketing research is based on assumptions, that help you take a certain decision. Your key indicator would be in the type of respondents you recruit during your survey. Make sure you have the chosen individuals homogoneously picked out. And that's all! .. Moreover... questions like "What type of vehicle did you previously have" .. and "Why type of vehicle do you think you'd own next?" ...

    Good luck with your interview.

    Shady
  • Posted byPaul Linnellon9/5/2005 at 2:06 PM Accepted
    Hi MS

    For a formula and explanation, check out the CONFIDENCE worksheet function in Microsoft Excel help.

    Good luck with the interview

    PGL
  • Posted byPaul Linnellon9/5/2005 at 2:28 PM Member
    This information should be available from your licensing authority.
  • Posted byChris Blackmanon9/5/2005 at 11:12 PM Accepted
    OK, if you survey a REPRESENTATIVE sample of the population sized at 1000 persons, and you discover (say)

    30% don't own a car
    20% own brand A
    40% own brand B
    10% own brand C

    Then you can extend that to the whole population... So say there are 1 million people in the population, you can say 300,000 don't own a car, 200,000 own brand A, etc.

    But you must ensure your survey is representative of the population. Beware - if you survey people by telephone, is there a relationship between telephone ownership and car ownership (i.e. more affluent households may have both a car and a phone). If you survey people in the street near a railway station, you may get a correspondingly higher level of non-car owners...

    One reliable way to size the market is to see if the Government releases statistics on new car registrations, or any kind of car registrations... If you know how many cars are registered, that's the size of the national "fleet". If you can separately get the number f new car registrations (maybe from the government or vehicle manufacturers industry associations) then you can determine the approximate fleet renewal rates.

    Say there are 1M cars on the road, and 10,000 new cars registered every month. Then 120,000 new cars registered each year, means fleet is being renewed approx every 8 years...

    Hope this helps.

    ChrisB
  • Posted bymgoodmanon9/6/2005 at 8:58 PM Accepted
    I would simply ask survey respondents how many cars they own (in the household). Then I'd take the average (example: 1.7 cars per household) and multiply that by the total number of households in my universe.

    In that example, if my universe had 2 million households, that would translate to 3.4 million cars.

    As others have said, it's critically important that your survey sample be representative of the universe. If you're doing the study in Manhattan the results are not projectable to Albany, etc.; and urban residents are not representative of suburban (or rural) residents, etc.
BACK TO TOP

Post a Comment

Related Discussions

Community Info

Top 25 Experts
(Research/Metrics)