sampling

SAMPLING DESIGN

Geethakumary.V.P.

SAMPLING -TERMS

SAMPLING THEORY-is a mathematical method of decision making for determining the most efficient means of selecting a sample that represent the population under study

SAMPLING -TERMS

Population : Is a group whose members possess specific attributes that a researcher is interested in studying.Population is the entire aggregation of cases that meet a designed set of criteria.

May consist of events, places, objects, animals or individuals

Sampling -terms

Target Population : is the population under study / The population to which the researcher wants to generalize the findings.Is the entire set of individuals or elements who meet the sampling criteria.Accessible Population : is that part of target population that is available to the researcher.

SAMPLE

A portion of the population that has been selected by the researcher to represent the population of interest.

Sample consists of a subset of the units that compose the population.

Element : Is a single member of the population under study. (Subject / Participant )

SAMPLING -TERMS

Sampling criteria/eligibility criteria –sampling criteria list the characteristics essential for membership in target population.The criteria that specify characteristics that the people must possess to be included into the sample.

Inclusion criteria Exclusion criteria : The criteria that the

subjects must not possess.

Sampling criteria/eligibility criteria

Reflect considerations other than substantive or theoretical interest

Cost Practical constraints Peoples ability to participate in the study Design considerations

SAMPLING -TERMS

REPRESENTATIVENESS-means that the sample must like the population in as many ways as possible

Representativeness – is how well the sample represents the variables of interest in the target population

Representative sample is one whose key characteristics closely approximate those of population

If sample is not representative ,external validity is at risk

REPRESENTATIVENESS contd

No way make sure that a sample is representative without obtaining information from population.

Certain sampling procedures are less likely to result in biased samples than others ,but representativeness can never be guaranteed.

Minimise error and if possible estimate their magnitude.

SAMPLING -TERMS

Sampling frame : A comprehensive list of all the sampling elements in the target population.

Strata : A stratum refers to a mutually exclusive segment of a population established by one or more characteristics.

SAMPLING –TERMS contd

Random selection : Is a process of selecting a representative sample of the target population. The purpose of which is to ensure that every element in the target population has an equal, independent non-zero chance of being selected for inclusion in the study.

SAMPLING –TERMS contd

Sampling frame : A comprehensive list of all the sampling elements in the target population

Strata : A stratum refers to a mutually exclusive segment of a population established by one or more characteristics

SAMPLING -TERMS

Statistic and parameter information collected from a sample is called

statistic information collected from a population is called a

parameter

Data –data are information collected from a source

SAMPLING -TERMS

Sampling error –is the difference or error between the sample statistic and the population parameter-estimated statistically by calculating the standard error.

Standard error –is the measurement of the standard deviation between a sample measure and a population measure.

SAMPLING -TERMS

Standard error random variation –expected difference in value of different subjects from the same sample systematic variation –exclusion criteria tend to increase the systematic bias in sample

SAMPLING -TERMS

Sampling bias –occurs when the researcher shows a preference in selecting one participant over the another.

Systematic over representation of segment of the population in terms of a characteristics relevant to the research question.

Sampling Design

Why Sample?

Why not study everyone? Census vs. sampling

Why sampling? Sampling reduces demands on resources

-finance, manpower, materials. Reduces duration of study. Sampling may be the feasible method. Ethically sampling is the only permitted method Sometimes increases accuracy(non sampling

errors and non response rate can be kept

minimal)

Quantitative studies

Seek to select samples that allow them to generalize their results to broader groups

Sampling plan that specifies the number & method of selection, in advance

Qualitative studies

Not concerned with issues of generalizability

Concerned with holistic understanding of the phenomenon of interest

Sampling decisions are based on informational & theoretical needs

Do not develop a formal sampling plan in advance

Sampling plan Describes the strategies that will be used

to obtain a sample for the study Developed to

Enhance representativenessReduce systematic biasDecrease sampling error

Provides details of the use of a sampling method in a specific study

Must be described in detail for purpose of critique, replication & meta analyses

Types of SamplingProbabilityNon-Probability

Probability Sampling methods

Random selection in choosing elements Researcher can specify the probability

that each element of the population will be included in the sample

More respected of the two approaches because of greater confidence in their representativeness

Probability Sampling

Is a random method of selecting participants in a research study who are the most representative of the population

Involves a selection process in which each element in the population has an equal, independent chance of being selected

Probability Sampling

Cannot guarantee “representativeness” on all traits of interest

A sampling plan with known statistical properties

Permits statements like: “The probability is .99 that the true population correlation falls between .46 and .56.”

Probability Samples

A probability sample is one in which each element of the population has a known non-zero probability of selection.

Not a probability sample if probabilities of selection are not known.

Types of Sampling

Probability Simple Random Stratified Random Cluster sampling Systematic sampling

Types of Sampling

Probability Simple Random Systematic Random Stratified Random Random Cluster Complex Multi-stage

Random (various kinds) Stratified Cluster

Simple random sampling

The most basic of probability sampling designs

Each element has independent chance for being included in to the study group

Requires complete list of the accessible population

One stage selection process

Steps of simple random sampling

Advantages The randomness offers the advantage of being

the most representative The only viable method of obtaining

representative sample in quantitative studies Allows for the use of inferential statistics Allows more accurate generalization to the

target population Allows the researcher to estimate the magnitude

of sampling error

Advantages

Eliminates researcher bias

Requires limited knowledge about the population

Provides a means of estimating the sampling error

Stratified random sampling

A variant of simple random sampling in which the population is first divided into 2 or more strata or subtypes

Used to increase the representativeness of different groups within the population

Subdivides the population in to homogenous subjects from which an appropriate number of elements can be selected at random

Stratified Random Sampling-1

Divide population into groups that differ in important ways

Basis for grouping must be known before sampling

Select random sample from within each group

Stratified Random Sampling

For a given sample size, reduces error compared to simple random sampling if the groups are different from each other

Probabilities of selection may be different for different groups, as long as they are known

Over sampling small groups improves inter group comparisons

Proportional stratified random sampling

A method of increasing the representativeness of the variable in the sample

The number of subjects taken from each stratum would be proportional to the number in the population

AdvantagesMore representative

Greater accuracy

Greater geographic concentration

Decreased sampling error, power increased, data collection time reduced

DisadvantagesIt requires extensive knowledge of the population under study to stratify it accurately

A complete list of target population is needed

It can quickly become very complex

Knowledge of advanced statistical methods & or assistance of a sampling consultant will be needed

Disproportional stratified random sampling

Disproportional stratified random sampling

In this sampling method the number of subjects in each stratum would not be proportional to the number in the population

Cluster sampling Suitable when the study population is large

Takes place in stages –Multistage sampling

The researcher begins with the largest most inclusive sampling unit , then progresses to the next most inclusive unit until the final stage i.e.. The stage from which the study subjects are randomly drawn to the sample

Involves successive random sampling of the units

Research Example

Cluster sampling

Merits Economical and time saving It allows probability sampling for a

population that is not listed in the sampling frame

Flexibility in the sampling method-existing divisions/subdivisions can be used

Cluster sampling

Limitations less accurate possibility of sampling error in each stage

Systematic SamplingCan be a probability or non-probability sampling

To be a probability sampling design, the elements in the sampling frame need to be listed randomly

Involves the selection of every kth from the sampling frame

The sampling interval, k= N/nWhere N is the accessible population & n is the

sample size

Systematic Random Sampling

Each element has an equal probability of selection, but combinations of elements have different probabilities. Population size N, desired sample size n, sampling interval k=N/n.

Randomly select a number j between 1and k, sample element j and then every kth element thereafter, j+k, j+2k, etc.

Example: N=64, n=8, k=64/8=8. Random j=3.


Has same error rate as simple random sample if the list is in random or haphazard order

Provides the benefits of implicit stratification if the list is grouped


Runs the risk of error if periodicity in the list matches the sampling interval

This is rare. In this example, every 4thelement is red,

and red never gets sampled. If j had been 4or 8, ONLY reds would be sampled.

Random Cluster Sampling

Done correctly, this is a form of random sampling Population is divided into groups, usually

geographic or organizational Some of the groups are randomly chosen In pure cluster sampling, whole cluster is sampled. In simple multistage cluster, there is random

sampling within each randomly chosen cluster

Random Cluster Sampling

Population is divided into groups Some of the groups are randomly selected For given sample size, a cluster sample has

more error than a simple random sample Cost savings of clustering may permit larger

sample Error is smaller if the clusters are similar to

each other

Random Cluster Sampling -

Cluster sampling has very high error if the clusters are different from each other

Cluster sampling is NOT desirable if the clusters are different

It IS random sampling: you randomly choose the clusters

But you will tend to omit some kinds of subjects

Stratified Cluster Sampling

Reduce the error in cluster sampling by creating strata of clusters

Sample one cluster from each stratum The cost-savings of clustering with the

error reduction of stratification

Stratification vs. ClusteringStratification Divide population into

groups different from each other: sexes, races, ages

Sample randomly from each group

Less error compared to simple random

More expensive to obtain stratification information before sampling

Clustering Divide population into

comparable groups :schools, cities

Randomly sample some of the groups

More error compared to simple random

Reduces costs to sample only some areas or organizations

Stratified Cluster Sampling

Combines elements of stratification and clusteringFirst you define the clusters

Then you group the clusters into strata of clusters,putting similar clusters together in a stratum

Then you randomly pick one (or more) clusterfrom each of the strata of clusters

Then you sample the subjects within the sampledclusters (either all the subjects, or a simple random sample of them)

Multi-stage Probability Samples

Large national probability samples involve several stages of stratified cluster sampling

The whole country is divided into geographic clusters, metropolitan and rural

Some large metropolitan areas are selected with certainty (certainty is a non-zero probability!)

Other areas are formed into strata of areas (e.g. middle-sized cities, rural counties); clusters are selected randomly from these strata

Within each sampled area, the clusters are defined, and the process is repeated, perhaps several times, until blocks or telephone exchanges are selected

At the last step, households and individuals within household are randomly selected

Random samples make multiple call-backs to people not at home.

Non-probability Sampling

An alternative approach to probability sampling

Each element in this study does not have an independent chance of being included in the study

Non Probability Sampling methods

Elements are selected by non random method

No way to estimate the probability that each element has of being included

Every element usually does not have a chance for inclusion

Less likely to produce accurate & representative sample

Non-probability Samples

• Convenience • Purposive • Quota

Convenience Sampling

Uses participants who are easily accessible to the researcher & who meet the criteria of the study

Entails the use of the most conveniently available people / objects as subjects in a study

Also known as Accidental sampling


Also called CHUNK

refers to fraction of the population being investigated which is selected

neither by probability nor by judgement

but by convenience .

Convenience Sample

Subjects selected because it is easy to access them.

No reason tied to purposes of research. Students in your class, people on State

Street, friends


Advantages: Easy, Saves money & time

Disadvantages: Potential for sampling bias, Limited generalizability of the results

Purposive Samples

Subjects selected for a good reason tied to purposes of research

Small samples < 30, not large enough for power of probability sampling. Nature of research requires small sample Choose subjects with appropriate variability in what you

are studying Hard-to-get populations that cannot be found

through screening general population

Quota Sampling

Pre-plan number of subjects in specified categories (e.g. 100 men, 100 women)

In uncontrolled quota sampling, the subjects chosen for those categories are a convenience sample, selected any way the interviewer chooses

In controlled quota sampling, restrictions are imposed to limit interviewer’s choice

No call-backs or other features to eliminate convenience factors in sample selection

Quota Vs Stratified Sampling

In Stratified Sampling, selection of subject is random. Call-backs are used to get that particular subject.

Stratified sampling without call-backs may not, in practice, be much different from quota sampling.

In Quota Sampling, interviewer selects first available subject who meets criteria: is a convenience sample.

Highly controlled quota sampling uses probability sampling down to the last block or telephone exchange

•But you should know the difference for the test!!

Snowball Sampling

A particular type of convenience sampling

Useful for studies in which the criteria for inclusion in the study specify a specific trait i.e. difficult to find by ordinary means

In this early members are asked to identify & refer other people who meet the eligibility criteria

Also known as Network sampling

Non-probability Sampling

Advantages: Less complicated, less expensive& allows the researcher to be more spontaneous when research situation arisesDisadvantages: Sample may not be a representative of the population, Cannot be generalized beyond the study sample & sampling error can not be estimated

STEPS IN SAMPLING

IDENTIFY THE POPULATION –IDENTIFY THE POPULATION –TARGET,ACCESSIBLETARGET,ACCESSIBLE

SPECIFY THE INCLUSION ANDSPECIFY THE INCLUSION AND

EXCLUSION CRITERIAEXCLUSION CRITERIA SPECIFY THE SAMPLING PLAN –SPECIFY THE SAMPLING PLAN –

METHOD,SIZE.METHOD,SIZE. IDENTIFY THE ELEMENTIDENTIFY THE ELEMENT

STEPS IN SAMPLING-Contd RECRUIT THE SAMPLERECRUIT THE SAMPLE

USE A SCREENING INSTRUMENTUSE A SCREENING INSTRUMENT

IDENTIFY ELIGIBLE CANDIDATESIDENTIFY ELIGIBLE CANDIDATES

GAIN COOPERATIONGAIN COOPERATION

EnjoyableEnjoyable

worthwhileworthwhile

convenientconvenient

pleasantpleasant

nonthreateningnonthreatening

STEPS IN SAMPLING-Contd RECRUIT THE SAMPLERECRUIT THE SAMPLE RECRUITMENT METHODRECRUITMENT METHOD face to face more effectiveface to face more effective

CourtesyCourtesy

PersistencePersistence

IncentivesIncentives

Research benefitsResearch benefits

Sharing resultsSharing results

Endorsements, confidentiality assuranceEndorsements, confidentiality assurance

STEPS IN SAMPLING-Contd

RETAINING SUBJECTS

get contact details

reimbursement for participation

bonus payment

Use subjects time preciously

Do not take for granted

nurture subjects-refreshment, surrounding

Maintain a pleasant climate

STEPS IN SAMPLING-Contd

ENSURE THESE DO NOT INFLUENCE THE DATA

SAMPLE SIZE No simple formula Use the largest sample Larger the sample more representative it is

likely to be Smaller sample tend to produce less

accurate estimates than larger ones Larger the sample smaller the sampling

error

SAMPLE SIZE -contd Size increases probability of deviant sample

diminishes Larger sample provide opportunity to

counter balance atypical value Larger sample are no assurance of

accuracy Large sample can harbor extensive bias

HOW TO ESTIMATE SAMPLE SIZE ?

CONSIDER

Main research question, outcome measure statistical procedure

statistical and clinical assumption

type 1,type 2 error- less; more subjects

level of significance-high; more subjects

precision - high ;more subjects

one sides/two sides


CONSIDER study constraints

Availability of resources- finance

material, manpower, logistic support,

Time, Ethical consideration level of significance, power and effect size


LEVEL OF SIGNIFICANCE –the more stringent the greater the necessary sample size

POWER

- is the capacity of the study to detect differences or relationships that actually exist in the population

-is the capacity to correctly reject a null hypothesis


EFFECT SIZE

Effect is the presence of a phenomenon If a phenomenon exists;it exist to some

degree Effect size is the extent of the presence of

a phenomenon

HOW TO ESTIMATE SAMPLE SIZE ? It is easier to detect large difference Smaller sample can detect large difference Smaller effect size require large sample Effect size is smaller with small samples and

thus more difficult to detect Increasing sample size increase effect size

making it more likely that the effect will be detected

Extremely small effect size may not be clinically significant

Factors determining sample size Size of population The resource available The degree of accuracy or precision

desired Homogeneity / heterogeneity of population Nature of the study Sampling plan adopted

Factors determining sample size Nature of respondent/attrition Effect size –strength of relationship

between variables Subgroup analysis Sensitivity of measures

Type of study

Qualitative & case studies: use very small samples, comparisons are not being made and sampling error & generalizations have little relevance

Type of study

Descriptive studies using survey questionnaires & correlational studies often require very large samples

Multiple variables are examined, extraneous variables

Type of study

Quasi & experimental studies must use sample size sufficient to achieve an acceptable level of power (.08) to reduce the risk of Type II error

These studies use controls & refined instruments, thus improving precision

Type of study

The study design influences power, but the design with the greatest power may not always be the most valid design to use

Number of variables

If the variables are highly correlated with the DV, the effect size will increase & sample size can be reduced

Therefore the variables in a study must be carefully selected

They should be essential to the question Or should have a documented strong

relationship with the DV

Measurement Sensitivity

As variance in instrument scores increases, the sample size needed to gain an accurate understanding of the phenomenon under study increases

Data analysis techniques

Large samples must be used when the power of the planned statistical analysis is low

For some procedures having equal group size increases power, because the effect size is maximized

The more unequal the group sizes are, the smaller the effect size. Therefore, in unequal groups the total sample size must be larger

Attrition

Number of subjects decline over time More likely if time lag between data

collection points is great Mobile population, difficult to trace Vulnerable or ‘at risk’ population Anticipate loss of subjects over time

Sample Size

Heterogeneity: need larger sample to study more diverse population

Desired precision: need larger sample to get smaller error

Sampling design: smaller if stratified, larger if cluster Nature of analysis: complex multivariate statistics

need larger samples Accuracy of sample depends upon sample size, not

ratio of sample to population

Sampling in Practice

Often a non-random selection of basic sampling frame (city, organization etc.)

Fit between sampling frame and research goals must be evaluated

Sampling frame as a concept is relevant to all kinds of research (including non probability)

Non probability sampling means you cannot generalize beyond the sample

Probability sampling means you can generalize to the population defined by the sampling frame

Evaluation of sampling design

Nonprobability

- rarely representative of the population

some section will be underrepresented

-Advantage lies in their

economy,convenience

Evaluation of non-probability sampling

Under or over representation of some segments

Convenient & economical Sometimes there is no option but to use

non probability approach

Evaluation of non-probability sampling

Need to be cautious in drawing inferences & conclusions from the data


Nonprobability

with care in selection of sample, conservative interpretation of the results and replication of the study with new sample NONPROBABILITY SAMPLE work well.


probability

-only viable method of obtaining

representative sample.

-helps to estimate sampling error

Problems in Sampling?

What problems do you know about? What issues are you aware of? What questions do you have?

Problems of data collection

Expect more than expected

time

difficulty level

changes

reaction of people

Problems of data collectionPEOPLE –unpredictable

bystanders /others

SAMPLE participation

disappearance

mortality

external influence in decision

passive resistance

Problems of data collectionRESEARCHER- Lack of staff

interaction

role conflict

control of emotion

INSTITUTIONAL –POLICY

Events-season ,climate,calamity, local events, national events

SUPPORT SYSTEM FOR DATA COLLECTION

PURPOSE TYPE OF ASSISTANCE

-Physical

-Money and material

-Emotional ACADEMIC COMMITTEE INSTITUTIONAL SUPPORT PERSONAL AND SOCIAL SUPPORT

KEY TASKS OF DATACOLLECTION

RECRUIT SUBJECT

MAINTAINING CONSISTENCY,CONTROL,

RELIABILITY

SOLVING PROBLEMS AS THEY COME

CRITIQUING SAMPLING PLAN

ISSUES

WHETHER DESCRIBED

ADEQUATELY

WHETHER THE RESEARCHER MADE GOOD SAMPLING DECISION


DESCRIPTION

description of main characteristics

number and characteristics of

potential subjects who decline to participate in the study


DESCRIPTION

the type of approach

population under study

eligibility criteria

setting

sample size,rationale

description of main charecteristics

sampling

Health & Medicine

sampling terms population

sampling terms sampling

sampling terms sampling

sampling terms sampling

sampling method

sampling elements

sampling decisions

sampling terms statistic