Textbook Notes (280,000)
CA (170,000)
York (10,000)
OMIS (50)
Chapter 5

Notes of Chapters 5-.docx


Department
Operations Management and Information System
Course Code
OMIS 2010
Professor
Alan Marshall
Chapter
5

This preview shows half of the first page. to view the full 3 pages of the document.
Statistics
Chapter 5- Data Collection and Sampling
5.1- Methods of Collecting Data
- A direct observation is the simplest, data obtained is called
observational, difficult to observe useful data this way
-Experimental data is produced vis-à-vis experiments
oA survey is any means by which you solicit information from people
about things/opinions (can be public or private)
The response rate is the proportion of people who were selected
for the survey who actually completed it
If it is low- can distort validity of the conclusions
Kind of survey is a personal interview- interviewer solicits
information from respondents by asking prepared questions, high
response rate, less misunderstandings of the questions, need to be
careful of bias in explanations, expensive
A telephone interview is less expensive, less personal, has a lower
expected response rate- generally only people with strong opinions
respond
A self-administered survey is a questionnaire that is mailed to a
sample of people; inexpensive (good for large population); low
response rate+ lots of incorrect responses (need to make sure well
designed)
5.2- Sampling
- Reason- cost, use sample to estimate population the target population is the
population about whom we wish to draw inferences; the sampled population is
the actual population from which the sample has been drawn
oCan come close to parameter- designed to estimate it
*presidential election when Roosevelt won even though predicted that he wouldn’t
5.3- Sampling Plans
- A simple random sample is selected in such a way that each person is equally
likely to be chosen
- A Stratified random sample separates the population into mutually exclusive
sets/ strata (single = Stratum) and then drawing simple random samples from
each stratum (only Stratify if connection between the survey and breakup; can
make inferences about/ compare different strata); can be done in 2 ways: 1. Can
draw same number for each strata, 2. Can draw based on group’s percentage of
the total population
-Cluster sampling is a survey of selected groups within a population (might not
be reliable)
oSampling error is high- can be partially offset by making sampled group
larger
5.4- Sampling and non-Sampling errors
-Sampling error is the difference between sample and population because of
amount that deviates by chance- reduce- make sample larger;
- Non-sampling error- serious because increasing size of sample doesn’t decrease
it
You're Reading a Preview

Unlock to view full version