Hello Luca,
Thank you for the support provided. It is the first survey i
conduct and your advice helped me a lot.
I found that a fairly random sample could be attained by selecting
users from the ISPs clients database. The survey should be conducted
through email questionnaires sent by the ISPs in order to keep
confidentiality over client's adresses. I still have to look over the
representativeness of the sample. I plan to research "classical"
variables: demographics, internet use, attitudes towards ISPs.
Still, with respect to the rate response to email,
How should I send the questionnaire?
a as text in email
b as link to the HTML page
c both
d other
Thank you.
All the best,
Cosmin
vineri, 14 mai 1999, you wrote to me:
LM> Hello Cosmin,
LM> My answer to your questions is below. Good luck with your project.
LM> Ciao,
LM> Luca
LM> Cosmin Ciocan wrote:
>> Hello!
>>
>> I plan to conduct a survey to define the internet user profile in my country.
>> It is part of my graduation dissertation project.
>>
>> I will place a questionnaire on a web page and allow submission for 20 days.
>> I have a low budget and I don't think I can afford placing more than 1
>> banner on a page with aprox 2000 visitors daily.
>>
>> QUESTIONS ARE:
>> 1.Do you know of other effective and methodological correct ways of promoting
>> the survey to a specific group of users (Romanian speaking users)?
>>
LM> You should try to put a link to most frequently visited pages in Romanian. For
LM> instance internet provides always have a so-called portal. Try to contact the
LM> webmasters of such sites and ask them to put a link to your web page. Since you
LM> are a student it may work.
>> 2.What response rate should I expect?
LM> Difficult to say. Of the people will visit your page, between 10% and 50% will
LM> usually complete the survey. The influencing factors are the topic you are
LM> surveying, navigability of the questionnaire, incentives you provide to the
LM> users, lenght of the questionnaire, etc.
>> 3. The overall population is estimated at 300,000 users. What should
>> be the sample's size for the study to be representative?
LM> It depends from how variegate are the characteristics of your reference
LM> population. The more variability on the variables you intend to study, the
LM> greater the number of respondents you shall need. For instance, suppose you
LM> desire to study the attitudes of your target population with respect to a
LM> political candidate. Well, if there is only one predominant opinion about the
LM> candidate you should be able to obtain a reliable estimation of the population's
LM> attitudes by drawing a small sample. However, since may not know the
LM> characteristics of the 300,000 users, then I suggest you draw a sample as larger
LM> as possible given time and budget constraints. Having said this, I think that if
LM> you reach more than 500 users you should get a good picture of what you want to
LM> measure. Also, do not forget that web sampling can hardly been reconducted to the
LM> principles and rule of probabilistic sampling (see e.g. Cochran's "Sampling
LM> Techniques" - 1977 John Wiley & Sons). You should therefore be very careful to
LM> enlarge your findings to the entire population of Romanian web users.
LM> Hope this helps,
LM> Luca
LM> 13 May 1999
LM> --
LM> Mr. Luca Meyer
LM> Internet Market Research Consultant
LM> http://lucameyer.com/
LM> email: [EMAIL PROTECTED]
-----------------------------------------------------------------
IMAFDI-E: Internet Marketing infos for students
subscribe IMAFDI-E:
mailto:[EMAIL PROTECTED] TEXT: subscribe imafdi-e
unsubscribe IMAFDI-E:
mailto:[EMAIL PROTECTED] TEXT: unsubscribe imafdi-e
FAQ: http://www.imafdi.de/imafdi_thanks.htm
---------------------------------------------------------
