Presentation on theme: "SAMPLE VS. POPULATION DESIGN METHODS CONSTRUCTION ERRORS Chapter 2 Samples and Populations."— Presentation transcript:
SAMPLE VS. POPULATION DESIGN METHODS CONSTRUCTION ERRORS Chapter 2 Samples and Populations
Sample vs. Population Population – the totality of subjects under consideration Target Population – consists of all subjects considered in the study Sample – a portion or a subset of the population for data collection and analysis Population/Target Population Sample
Sample vs. Population Kalamazoo House -holds Young-adults and older Population Target Population Sample
Census vs. Sample Survey Census – collection of data using all subjects in the population Sample Survey – collection of data from a representative sample of the population Population/ Target Population Sample Note: Random Samples should be representative of the population
Study Design or Protocol Design Steps involved in solving problems How do I solve this problem? ? ? Study design is done prior to data collection. It involves methods in data collection, analysis of the data and conclusions to be made.
Probability vs. Non-Probability Sampling Probability Sampling – subjects are chosen by chance Non-probability Sampling – can be used for informal and less scientific studies Note: Non-probability sampling tend to be less representative of the target population
Methods in Probability Sampling Simple Random Sampling (SRS) – samples are randomly selected from the population K-in-1 Systematic Sampling – Every kth subject is chosen Stratified Random Sampling – population is divided into subgroups called strata and SRS chosen from each strata Cluster sampling – population is divided into subgroups called clusters and clusters are randomly chosen as samples.
Example: Household Expenditures in Michigan Target Population : Households in Michigan Simple Random Sampling – randomly selecting the sample from a list of households Systematic Sampling – every 10 th household Stratified Sampling – take samples from each county Cluster Sampling – selecting counties in Michigan
Factors to be considered in a Survey Money Time Content/Information
Types of Surveys Type of SurveyAdvantagesDisadvantages Face-to-face Interview Explain questions, explore issues, make observations, use visual aids Expensive, need interviewer training - at home or work Accuracy, better sampling Expensive - in public areasCheaper, more people in less time Less representative sample Telephone Interview Accurate, cheapNo personal observation Written Questionnaire Cheapest per respondentBias from low response rate - by mailAllows anonymitySlow - by e-mailCheaper, quicker resultsLess representative sample - web surveyQuicker data processingNeed computing expertise
Construction of Questionnaire Is the question understandable? Are you gathering knowledge or attitude? Are the questions loaded? Do the questions ask for sensitive information? Note: An accurate answer leads to a good study and it starts from asking important questions correctly.
Types of Survey Errors Coverage errors – sampling frame excludes some segments of the target population Non-response errors – can cause bias in survey results Measurement errors – occurs when respondents answer ‘incorrectly’