Sampling

Sample:

a group of participants in a study selected as being representative of a defined population (control variable defines population)

Population:

a large group to which the results of a study conducted on a sample of that group may be inferred (group of potential participants to whom you want to infer the results of a study)

Inference:

a conclusion reached on the basis of evidence and reasoning (generalization of results found in the sample to some larger group) (sample to pop)

Generalizability:

ability to generalize results to different populations with the same characteristics in different settings (pop to pop)

Statistic:

a value representing a characteristic of a sample

Parameter:

a value representing a characteristic of a population

Sampling:

the act process, or technique of selectin a suitable sample,

o   Specifically the act, process, or technique of selecting a representative part of a population for the purpose of determining parameters or characteristics of the whole population

Issues related to sampling

-        Larger sample, more representative, but may be too costly

-        Selection may be flawed (taking easy way out)

-        What kind of participants do you need

-        Age: gender : trained or untrained: experts or novices

-        Size – special types (athletes, cyclists, etc)

-        Can you get permission – can you find enough

Probability sampling:

likelihood of any one member of population being selected is known

Nonprobability sampling

-        likelihood of selecting one member of population is not known

Random sampling:

procedure for selection of participants for a study that provides an equal chance of election for all members of the population        

simple random sample procedure

o   Must obtain a complete sampling frame

o   Each person has same change of being selected

o   Suitable where population is relatively small and where sampling frame is complete and up to date

Systematic sampling:

procedure for selection of participants for a study in which every nth individual on a list is selected

o   Challenges: periodicity

o   Must first workout a sampling fraction by dividing population size by sample size and select every nth

Stratified sampling:

selection procedure for participants in a study that provides an equal chance of selection for all members of purposeful divisions of the population

o   Divide populations into relevant subgroups and randomly sample subgroups

Cluster sampling:

useful wen the population to be studied is infinite, a list of members of the population Is nonexistent, the geographic distribution of the population Is widely scattered

o   Randomly sample groups

advantages of non probability sampling

• cheaper,

• used when sampling frame isn’t available,

useful when population is so widely dispersed that cluster sampling wouldn’t be efficient,

• Often used in exploratory studies,

• some research not interested in working out but what proportion of population give a particular response but rather in obtaining an idea of the range of responses on ideas that people have

• Not always possible to undertake probability method of sampling

Convenience sampling:

group of participants for a study selected because of readiness of opportunity

Quota sampling:

non probability technique where the assembled sample has the same proportions of individuals as the entire population with respect to known characteristics traits or focused phenomenon

o   Divide population into subgroups, identify proportions inn the population, select subjects from various subgroups, ensure sample is representative of the entire population

o   Disadvantages: bias from interviews, impossible toe estimate accuracy because not random sample

Snowball sampling:

you initially contact a few potential respondents and then ask them whether they know of anybody with the same characteristics that you are looking for in your research

Periodicity

-        Bias caused by particular characteristics arising in the sampling frame at regular units

Sampling error

-        Inclusion of participants in a study that results in a sample not representative of the population due to chance error

sample size

• (Adult U.S. population/ 1000-1500 participants = typical)

• issues

  • o   Larger sample = smaller the sampling error

    o   Big is good, appropriate is better

What is effect size

Calculated value that serves as a relative index of the meaningfulness of results

o   Idea of standard deviation to contextualize the difference between the two groups

o   Standard deviation is a measure of how spread out a set of values are

Variance:

large variance as measured by the standard deviation the results you find may not represent what actually happened in your research

Validity:

in an experiment the researcher attempts to establish cause and effect relationships

o   Requires a logical process

• Good theoretical framework

• Use of appropriate subjects

• Application of correct type of design

• Use of correct statistical model

• Proper selection and control of independent variable

• Appropriate selection and measurement of dependent variable

• Correct interpretation of results