survey and sample design

Question 1

population

Answer 1

depending on the research question

Question 2

sample

Answer 2

selection of elements of the population

Question 3

survey unit

Answer 3

elements of the population to which the sample selection relates and which are generally eligible for inclusion

Question 4

surveying the whole population…

Answer 4

is expensive and time consuming - that is why a sample is drawn and examined

Question 5

aim

Answer 5

is to select the sample in such way that is representative as possible of the population

Question 6

it is desirable that..

Answer 6

the measured values of the variables do not differ to much in terms of their statistical measures in the sample and in the population

Question 7

simple random selection

Answer 7

all n sample elements habe the same probability of being selected
Urn model: the population must be homogenous w with respect to the feature of interests
based on findings of probability theory, statements can be made about the accuracy and safety of the results

Question 8

stratified random selection

Answer 8

population is first derived into different strata and then from each strata some elements for the sample are taken
useful if the total is heterogenous with respect to the feature of interest
ensure that certain parts of the population are not over represented

Question 9

cluster sampling

Answer 9

Clusters of adjacent elements are included in the sample. within randomly selected cluster all are interviewed
this makes sense if there is no reliable list of all units of a population
cost saving

Question 10

quota sampling

Answer 10

characteristics are distributed differently
population is divided into quotas, percentage of quota are determined - the sample is distributed among the quota, interviewers are given instructions on how many are to be interviewed within the quota

Question 11

systematic selection

Answer 11

e.g. from an address list with consecutive numbering, each I-th is selected (e.g. every twentieth)
advantage: no element can occur multiple times
does not entirely correspond to the essential aspects of simple random selection

Question 12

arbitrary selection

Answer 12

selection is the responsibility of the interviewer or the person of investigation
selection without rules
inferential statements are not possible
repräsentative statements are not possible
checking correlation is possible

Question 13

clarification of the field access

Answer 13

Which population is assumed?
Which form of sampling?
How many people should be reached?
How should the survey be conducted?
how should people be approached and motivated to participate?
how is the return of results organized

Question 14

determination of the sample size

Answer 14

the greater the accuracy of the statistical test of the hypotheses an d the lower the likelihood of error the larger the sample must be
to determine the necessary sample size - desired estimation accuracy and the error probability