brm class presentation
TRANSCRIPT
-
8/7/2019 BRM class presentation
1/19
-
8/7/2019 BRM class presentation
2/19
Simple Random Sampling: A samplingprocedure that assures each element in thepopulation an equal chance of being included
in the sample. E.g. drawing names from a hat,winning raffle ticket from a large drum etc.
If sample is large computer based randomsampling maybe used for sample selection.
-
8/7/2019 BRM class presentation
3/19
e.g. A researcher may be interested in selecting asimple random sample of all presidents of savingsand loan associations in New Mexico.
For this each presidents name is assigned anumber from 1 to 95 and each no is written on apiece of paper.
All slips are placed in a drum and thoroughlymixed.
One is selected for each sampling unit.
If sample size is 45, the selection procedure isrepeated 44 times after the first slip has beenselected.
-
8/7/2019 BRM class presentation
4/19
Def: A sampling procedure in which an initialstarting point is selected by a random process,and then every nth number on the list is
selected. E.g. selecting every 23rd name from a rural
telephone directory that does not separatebusiness listings from household listings
-
8/7/2019 BRM class presentation
5/19
Random results if the arrangement of the itemsin the list is random in character.
Periodicity: Occurs if list is not random incharacter.
E.g. collecting retail sales every seventh daywould result in a distorted sample because
there be a systematic pattern of selectingsampling units, sales for only one day of theweek would be sampled
-
8/7/2019 BRM class presentation
6/19
Def: A probability sampling procedure inwhich simple random subsamples are drawnfrom within different strata that are more or
less equal on some characteristic.Advantages:
More efficient sample
Random sampling error reduced Assurance that sample will accurately reflect
the population on the basis of the criterion
-
8/7/2019 BRM class presentation
7/19
Step 1: Identify a variable as an efficient basisfor stratification.
Characteristics of variable chosen:
a) characteristic of the population elementsknown to be related to dependent variable.
b) increase homogeneity within each stratum
c) increase heterogeneity between strata.d) easily convertible into sub groups.
E.g. pharmaceutical company
-
8/7/2019 BRM class presentation
8/19
Step 2: For each separate sub group, a list ofpopulation elements must be obtained.
Step 3: A separate simple random sample istaken within each stratum by using a table ofrandom numbers or some other device.
-
8/7/2019 BRM class presentation
9/19
Def: A stratified sampling in which the numberof sampling units drawn from each stratum isin proportion to the population size of that
stratum. A proportionate sample would have the same
percentages as in the population.
-
8/7/2019 BRM class presentation
10/19
Def: A stratified sample in which the samplesize for each stratum is allocated according toanalytical considerations.
The general logic is that as variability increases,sample size must increase to provide accurateestimates.
-
8/7/2019 BRM class presentation
11/19
Def: An economically efficient sampling technique inwhich the primary sampling unit is not the individualelement in the population but a large cluster ofelements.
Area Sample: Def: A cluster sample in which the primary sampling
unit is a geographic area. Most popular type of cluster sampling. E.g. a grocery researcher may randomly view all, or a
sample of, grocery stores within the geographicclusters. Interviews are confined to these clusters, no interviews
occur in other clusters.
-
8/7/2019 BRM class presentation
12/19
Cluster sample are utilized when no lists ofsample population are available.
Cluster should be homogenous as the
population itself.
-
8/7/2019 BRM class presentation
13/19
Def: Sampling that involves using acombination of other probability samplingtechniques.
E.g. A political pollster investigating an electionin Arizona may follow the following steps:
Step 1: Choose counties within the state
Step 2: Precincts within the selected counties
maybe chosen. Step 3: Blocks within the precincts might be
chosen, then all the blocks within thegeographic area would be interviewed.
-
8/7/2019 BRM class presentation
14/19
Degree of accuracy
Resources
Time
Advance knowledge of population
National versus Local project
Need for statistical analysis
-
8/7/2019 BRM class presentation
15/19
Rapid reach to a large sample
Instantaneous
Day-of-the-week effect
Lack of Internet penetration
A select sample segment
-
8/7/2019 BRM class presentation
16/19
-
8/7/2019 BRM class presentation
17/19
A sample from a panel
High response rate
Incentivized
Assurance
Propensity-weighting scheme
-
8/7/2019 BRM class presentation
18/19
Recruited ad hocSamples
Opt-in Lists
-
8/7/2019 BRM class presentation
19/19