feb 6th.docx

16 views7 pages
30 Mar 2012
First Theorem
Tells us the shape of the sampling distribution and tells us its mean and
standard deviation
If we start off with some trait that is normally distributed across the
population (IQ, height) and take an infinite number of equally sized random
samples from that population, the sampling distribution of sample means
will be normal
Central limit theorem
For any trait or variable, even those that are not normally distributed in the
population, as the sample size grows larger the sampling distribution of
sample means will become normal in shape
Sampling Error
The difference between the statistic (sample) and the parameter
Theoretical construct- it is impossible to know the true population
Should theoretically be reduced with a representative sample
Confidence intervals allow for the possibility of error
Don’t generally take thousands of samples
Other issues related to sampling:
Hidden populations
Sample size- rules of thumb
o Small population (less then 1000)= 30%
o Medium population (10,000) = 10%
o Large population (150,000) = 1%
o Extremely large population (more then 10 million) = 0.025%
o Subgroup analysis = 50/grop
o Small representative sample is better
o As sample size grows you don’t get the same accuracy
Other things that effect sample representativeness
Under-coverage (certain groups missing from sample)
Non-response/ missing data (people not answering certain questions,
particularly with income)
Response bias (respond how they think the researcher wants them to
respond rather then true beliefs or opinions, especially with moral judgment
The instrument itself (wording of questions, order of questions)
Survey research
Most widely used method of gathering data and information
Exploratory, descriptive, explanatory research
Unlock document

This preview shows pages 1-2 of the document.
Unlock all 7 pages and 3 million more documents.

Already have an account? Log in
Great for large samples
Sampling methods are very important- if you have a method with thousands
of people results will not be very good
Great for measuring attitudes
Reliability- which you can measure
Can describe large populations
Standardized (allows you to compare across countries/ time)
Validity (hard to know if you are measuring what you thing you are
measuring, reader may not understand)
Cannot modify questionnaire in field
Types of surveys
Self administered (mail, phone, e-mail, web) often people just want to get it
done and don’t pay attention
Interview surveys (not as common, face-to-face) takes away anonymity, can
be sure participant understands the question
A questionnaire should include
Contact letter first (giving info about study what you’re asking about, why
you’re doing it, why its important)
Information letter (include consent)
A method of return if necessary
Questions on surveys
Surveys should be
o Mutually exhaustive (everyone fits into a category)
o Mutually exclusive (categories don’t overlap
Open ended questions must be coded
o Issues: misunderstanding, researcher bias
When designing questions
Must be voluntary
Make items clear (avoid abbreviations, vagueness, slang)
Avoid double barreled questions (ask two things at once, no and)
Must be possible to answer all questions accurately
Unlock document

This preview shows pages 1-2 of the document.
Unlock all 7 pages and 3 million more documents.

Already have an account? Log in

Get OneClass Notes+

Unlimited access to class notes and textbook notes.

YearlyBest Value
75% OFF
$8 USD/m
$30 USD/m
You will be charged $96 USD upfront and auto renewed at the end of each cycle. You may cancel anytime under Payment Settings. For more information, see our Terms and Privacy.
Payments are encrypted using 256-bit SSL. Powered by Stripe.