Research Methods - Sampling Techniques w154y
This document was ed by and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this report form. Report 3b7i
Overview 3e4r5l
& View Research Methods - Sampling Techniques as PDF for free.
More details w3441
- Words: 827
- Pages: 10
UNIVERSITY OF ARTS, BELGRADE Masters program in management in culture
Course in social research methods
Sampling techniques Slobodan Cvejic, PhD Belgrade University, Faculty of Philosophy [email protected]
SAMPLE
The word is used in a variety of ways in scientific and everyday language (e.g. textile industry, reporter in the street, public opinion researcher) Sample defined: a segment of population that provides reliable information about investigated features of that population; a subset of cases selected in certain way to provide reliable conclusions on investigated topics Function of sample: to save time and money, and yet provide solid grounds for improvement of knowledge and policy making History of sampling: when statistics spread over sciences the need appeared to extend data collection about different issues; Neyman defined how to conclude from a part to the whole; from mid 1900s on the most of the statistics is about inference from sample to population; even some censuses done on large samples RESEARCH METHODS: sampling
NON-PROBABILITY SAMPLES
Subjective judgement in selection Chosen for convenience or on the basis of systematically employed criteria Types
Convenience sample (most easily approached cases) Most similar/most dissimilar sample (cases with similar features) Typical case sample (usefull cases, not extreme) Critical case sample (cases that are key or essential) Snowball sample (cases from first selection identify additional cases to be selected) Quota sample (cases selected according to the population proportions)
Subjective selection – external valididty and credibility at risk RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Probability: apriori and aposteriori, mathematical and statistical – statistics deals with mass of cases, stohastic probability established empiricaly, after the occasion Relative frequences to probabilities (e.g. throwing the dice) Importance of statistical probability – allows for inference from a part to the whole In statistical experiment (a study) it is not important to have equal or certain probabilities for outcomes it is only important to know which those probabilites are RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Each unit in the population has a known, nonzero probability of being included in the sample Random selection necessary to eliminate subjective judgements – either using a lottery procedure or using computer program to generate random numbers Random does not mean arbitrary or haphazard – random selection is very careful ensuring independence of selections Completely random selection is an ideal – our goal is not to avoid all possible risks of subjectivity, but to be able to measure their impact on inference Sometimes we give certainty of selection to some cases, or make corrections to probability post-hoc RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Types Simple random Systematic Stratified Cluster Multistage
RESEARCH METHODS: sampling
3R RULE
3 principles for making a good probability sample
Realism – adjusting goals, definitions and selection procedures to realistic framework (e.g. audience of a movie vs. cinema audience of the movie) Randomization – at least at one step selection should be random in order to eliminate subjective judgments and increase external validity and credibility (e.g. choose cities ‘at hand’, choose households in those cities systematicaly, but choose persons inside the households at random) Reprezentativness – population in minimizing mirror
Target population – research population Target sample – research sample RESEARCH METHODS: sampling
INFERENCE BASED ON SAMPLE SAMPLE DESIGN
FEATURE X
SAMPLE SELECTION
MEASUREMENT OF FREQUENCES FOR EACH VALUE OF X
DISTRIBUTION OF FREQUENCES
S A M P L E
INFERENCE
EXPERIMENT DESIGN
DEFINITION OF OUTCOMES
DEFINITION OF SAMPLING FRAME
AG A NUMBER TO EACH OUTCOME
RANDOM VARIABLE X
RESEARCH METHODS: sampling
DEFINITION OF PROBAB. FOR EACH VALUE OF X
P O P U DISTRIBUTION L OF A PROBABILITY T I O N
PRACTICAL SAMPLING DESIGN TARGET POPULATION To whome we want to genralize findings STUDY POPULATION Operational definition of target population & measurement instruments SAMPLING DISTRIBUTION The distribution of an estimator computed from many samples centered around its expected value SAMPLE The subset of subjects or units for which data is obtained
Components and Sources of Total Error NONSAMPLING BIAS - Listing & Frame - Nonresponse - Measurement Error SAMPLING BIAS - Selection Bias - Estimation Bias SAMPLING VARIABILITY - Sample Size - Sample Homogeneity
RESEARCH METHODS: sampling
T O T A L E R R O R
POST-SAMPLING CHOICES
Weighting – to adjust sample distributions to probabilities FEMALE
MALE
POPULATION
50%
50%
SAMPLE
80%
20%
WEIGHT
0.625
2.5
80 * 0.625=50
20 * 2.5=50
FINAL
Evaluation of nonresponse (potential for nonsampling bias)
Plan to minimize nonresponse – follow up Post-hoc solutions: forced face to face interviews with sample of nonrespondents on restricted number of variables that are of utmost interest; e.g. Average score on test 94 out of 100 points; if response rate 90%, average score of 34 by nonrespondents needed to reverse conclusions, if rate 80% - 64 points, if 75% - 70 points RESEARCH METHODS: sampling
Course in social research methods
Sampling techniques Slobodan Cvejic, PhD Belgrade University, Faculty of Philosophy [email protected]
SAMPLE
The word is used in a variety of ways in scientific and everyday language (e.g. textile industry, reporter in the street, public opinion researcher) Sample defined: a segment of population that provides reliable information about investigated features of that population; a subset of cases selected in certain way to provide reliable conclusions on investigated topics Function of sample: to save time and money, and yet provide solid grounds for improvement of knowledge and policy making History of sampling: when statistics spread over sciences the need appeared to extend data collection about different issues; Neyman defined how to conclude from a part to the whole; from mid 1900s on the most of the statistics is about inference from sample to population; even some censuses done on large samples RESEARCH METHODS: sampling
NON-PROBABILITY SAMPLES
Subjective judgement in selection Chosen for convenience or on the basis of systematically employed criteria Types
Convenience sample (most easily approached cases) Most similar/most dissimilar sample (cases with similar features) Typical case sample (usefull cases, not extreme) Critical case sample (cases that are key or essential) Snowball sample (cases from first selection identify additional cases to be selected) Quota sample (cases selected according to the population proportions)
Subjective selection – external valididty and credibility at risk RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Probability: apriori and aposteriori, mathematical and statistical – statistics deals with mass of cases, stohastic probability established empiricaly, after the occasion Relative frequences to probabilities (e.g. throwing the dice) Importance of statistical probability – allows for inference from a part to the whole In statistical experiment (a study) it is not important to have equal or certain probabilities for outcomes it is only important to know which those probabilites are RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Each unit in the population has a known, nonzero probability of being included in the sample Random selection necessary to eliminate subjective judgements – either using a lottery procedure or using computer program to generate random numbers Random does not mean arbitrary or haphazard – random selection is very careful ensuring independence of selections Completely random selection is an ideal – our goal is not to avoid all possible risks of subjectivity, but to be able to measure their impact on inference Sometimes we give certainty of selection to some cases, or make corrections to probability post-hoc RESEARCH METHODS: sampling
PROBABILITY SAMPLES
Types Simple random Systematic Stratified Cluster Multistage
RESEARCH METHODS: sampling
3R RULE
3 principles for making a good probability sample
Realism – adjusting goals, definitions and selection procedures to realistic framework (e.g. audience of a movie vs. cinema audience of the movie) Randomization – at least at one step selection should be random in order to eliminate subjective judgments and increase external validity and credibility (e.g. choose cities ‘at hand’, choose households in those cities systematicaly, but choose persons inside the households at random) Reprezentativness – population in minimizing mirror
Target population – research population Target sample – research sample RESEARCH METHODS: sampling
INFERENCE BASED ON SAMPLE SAMPLE DESIGN
FEATURE X
SAMPLE SELECTION
MEASUREMENT OF FREQUENCES FOR EACH VALUE OF X
DISTRIBUTION OF FREQUENCES
S A M P L E
INFERENCE
EXPERIMENT DESIGN
DEFINITION OF OUTCOMES
DEFINITION OF SAMPLING FRAME
AG A NUMBER TO EACH OUTCOME
RANDOM VARIABLE X
RESEARCH METHODS: sampling
DEFINITION OF PROBAB. FOR EACH VALUE OF X
P O P U DISTRIBUTION L OF A PROBABILITY T I O N
PRACTICAL SAMPLING DESIGN TARGET POPULATION To whome we want to genralize findings STUDY POPULATION Operational definition of target population & measurement instruments SAMPLING DISTRIBUTION The distribution of an estimator computed from many samples centered around its expected value SAMPLE The subset of subjects or units for which data is obtained
Components and Sources of Total Error NONSAMPLING BIAS - Listing & Frame - Nonresponse - Measurement Error SAMPLING BIAS - Selection Bias - Estimation Bias SAMPLING VARIABILITY - Sample Size - Sample Homogeneity
RESEARCH METHODS: sampling
T O T A L E R R O R
POST-SAMPLING CHOICES
Weighting – to adjust sample distributions to probabilities FEMALE
MALE
POPULATION
50%
50%
SAMPLE
80%
20%
WEIGHT
0.625
2.5
80 * 0.625=50
20 * 2.5=50
FINAL
Evaluation of nonresponse (potential for nonsampling bias)
Plan to minimize nonresponse – follow up Post-hoc solutions: forced face to face interviews with sample of nonrespondents on restricted number of variables that are of utmost interest; e.g. Average score on test 94 out of 100 points; if response rate 90%, average score of 34 by nonrespondents needed to reverse conclusions, if rate 80% - 64 points, if 75% - 70 points RESEARCH METHODS: sampling
Related Documents 3m3m1z
Research Methods - Sampling Techniques w154y
March 2023 0
Business Research Methods [session 3- Sampling Techniques] 544s28
October 2019 58
Sampling Techniques In Quantitative Research 605j8
November 2019 74
Sampling Methods 2tw4r
July 2021 0
Non Random Sampling Techniques 5f3f54
October 2019 55
Sampling Techniques Worksheet 3w621j
April 2022 0More Documents from "zulejunior87" q71w