Sampling and Measurement

Inference Chain

Population → Sample → Measurment → Data → Statistics → Inference

Population

Entire group of individuals, objects, or observations a researcher wants to study

Target Population

Entire group of people or objects that a researcher wants to study and draw conclusions about

Accessible population

Part of the target population a researcher can realistically reach

Parameter

Numerical values that describes a characteristic of an entire population

Statistic

Numerical value that describes a characteristic of a sample taken from a population

Sampling Frame

A source that identifies members of a population from which a researcher selects a sample

Probability sampling

Individuals have a known probability of being selected. Included simple random, systematic, stratified, cluster, and multistage sampling

Simple random sampling

Every member has an equal chance of being selected

Systematic sampling

Researchers select every Kth individual from an ordered sampling frame

Stratified random sampling

divide population into meaningful strata, then randomly sample within each group

Disproportionate stratified sampling

researchers intentionally oversample a smaller subgroup

Cluster Sampling

Researchers randomly select naturally occurring groups instead of individuals

Multistage sampling

States → Counties → Schools → Classrooms → Students

Nonprobability sampling

probability to participate is unknown (volunteering)

Convenience sampling

Participants are recruited because they are easily accessible

Volunteer sampling

Participants chose whether to enter the study

Purposive sampling

recruit people with specific characteristics

Snowball sampling

participants refer other participants

Quota sampling

Recruit until category targets are filled

Volunteer bias

People who volunteer are different then those who don’t

Non-Response Bias

Selected individuals fail to participate

Attrition Bias

People drop out over time

Sampling error

Natural variability

Sampling bias

Systematic distortion

Standard error (SE)

Describes how much sample means vary from sample to sanple

Random sampling

Determines who enters the study

Random assignment

Determines which condition participants enter

Operational definitions

Specifies exactly how a variable will be measured or manipulated

Variables

Characteristic that can take different values

Independent/Dependent variables

IV-manipulated, DV-measured

Predictor and outcome variables

Predictor- sleep duration

Outcome- exam score

Categorical/Quantitative variables

Categorical - Group membership

Quantitative - Numerical Value

Discrete/Continuous variables

Discrete - countable

Continuous - many values within a range

Nominal

Categories without order

Ordinal

Categories with order

Interval

Values with equal intervals, no meaningful absolute zero

Range

Values with equal intervals, yes meaningful zero

Nominal/Dichotomous variables

Nominal - Diagnosis, group, major

Dichotomous - Treatment/control, yes/no

Interval measurement

Ordered values with approximately equal intervals

Ratio measurement

Equal intervals and meaningful zero point

Random error

Fluctuates unpredictability

Systematic error

Pushes measurements consistently in one direction

Reliability

Consistency or reproducibility of measurement

Discriminal validity

measures of different constructs should not correlate so strongly that they appear identical

Ceiling effects

Measure is too easy or incentive at the upper end

Floor effect

measure is too difficult or insensitive at the bottom

Range restriction

Sample contains a narrow range of scores

Student T test

Comparing mean in 2 groups

One-sample T test

comparing mean of a group against a known mean

Paired T test

Means come from same group

One way ANOVA

Comparing means between 3+ groups

Factorial ANOVA

Comparing means of groups defined by 2+ parameters

Type 1 error

False positive

Type 2 error

False negative