Lecture note - Sanpling

Survey

• Tool for measuring attitudes and orientations  in a large population

• Best method available to collect original data for describing a population too large to observe directly

Non-probability sampling – for obtaining information about specific groups

・Purposive / judgemental sampling

e.g. researching cancer patient 

・Snowball sampling

e.g. researching gang group, underground organization, terrorist organization..

・Quota sampling : the researcher selects sample from some fixed quota

When it is difficult to conduct probability sampling or it has no meaning to select respondent randomly.

Probability sampling

• Primary method of selecting large, representative samples

• Provide useful descriptions of the total population

• Sample of individuals from a population containing essentially the same variations that exist in the population

・Simple random sampling
Selecting sample at random
All samples have equal chance to be chosen
It is useful if the researcher does not know the demographics of the population.    
- Costly and time consuming

- Researcher need to have the list of sample which cover all sample 

・Stratified sampling

- Appropriate numbers of elements are drawn from homogeneous subsets of the population
- To avoid to include something not related to your research
- This method is useful if the researcher is familiar with the population.
Need the complete list with info (sex, age..)

・Cluster sampling

- Each cluster represents the population.
- Initial sampling of groups of elements (clusters), followed by the selection of elements within each of the selected clusters – can be multi-stage
- The clusters are ‘natural’ in a population
cost efficient since sampling method.

Need to know population of each cluster
Data from a cluster which is not selected was not included at all in the analysis. 

•Representativeness – if the aggregate characteristics of the sample closely approximate those same aggregate characteristics in the population

• Population – aggregation of elements from which the sample is selected

• Sample – selected element or respondent

• Sampling bias – those selected are not typical or representative of the larger population they have been chosen from

For making questionnaire

Operationalization – measurement

 > validity – what we intend to measure

 > reliability – extent to which measures give consistent results

Open-ended questions – respondents provide answers

Close-ended questions – respondents elect an answer from the list of options

Make items clear – “Do you agree with the peace agreement?”

Make relevant questions Answers should match the questions and set of options – “Are you satisfied with Ateneo education?” “barely, sometimes, often, very”

Answers should be mutually exclusive – “How often do you study?” “everyday, everynight, every weekend, once a week, more than once a week, monthly, never”

NO double-barreled questions –

 “Do you think that K-12 will make our students globally competitive and our education standards globally at par?”

NO questions with possibly different standards – “Do you think that students today are liberal?”

NO leading questions – “Due to changing times, do you agree that government policies must adapt to the changes?”

NO assuming questions – “How often do you go out with your gf/bf?”

NO difficult words - Embryonic Stem Cell

NO negative questions – “Do you NOT agree..”

NO ambiguity -  e.g. Which one you support, big government or small government?