Chapter 3: Variables, Claims, Validities

Variable

Something that varies (often interested in measuring and manipulating)

Must have at least two levels

Constants

something that does not vary (one level)

not necessary

Characterizing Variables

1. Conceptual vs Operational

2. Measured vs Manipulated

Conceptual Variable

(Construct) an abstract entity we know exists but is not tangible (abstract)

ex: happiness

Conceptual Definition

Theoretical definition used to limit and define a conceptual variable

ex: concept variable= satisfaction with life

conceptual definition = a person’s cognitive evaluation of their life

Operational Definition

a particular way a conceptual variable is defined that allows it to be measured or manipulated with a tangible metric

• crucial to hypothesis formation/study design

ex: conceptual variable: happiness—> conceptual definition: a person’s cognitive evaluation of their life —→ operational definition: 1-5 scale questionaire

Measured Variables

a variable whose levels are simply observed (naturally) and recorded (tracked)

(may or may not change over the course of the study)

ex: mood, height, IQ, hair color, etc

Manipulated Variables

• a variable that the researcher controls/influences

• assign participants to a specific level of variable (control, experimental)

Variable and Claims

we use variables to make claims about the world

Types of research claims

1. Frequency claims

2. Association claims

3. Causal claims

Frequency Claims

claims that describe a particular rate or degree of a single measured variable

ex:

Association Claims

a claim about 2 or more variables and how they influence each other

(correlations/covariations)

ex: linked, associated with, increase liklihood

Positive Association

Directionality: variable x increases/decreases, y increases/decreases

Negative Association

Directionality: variable x increases/decreases, y decreases/increases

zero association

Directionality: no association

Causal Claims

claim that argues that one variable causes another variable

ex: increases, decreases

In order to establish causality you must:

1. Covariance/Correlation

2. Temporal Precedence

3. Internal Validity: eliminate all other possibilities

Claim

an argument someone is trying to make

Validity

the appropriateness of a conclusion

Validity claim should be:

1. reasonable

2. accurate

3. justifiable

4 types of validity

1. Construct Validity (association)

2. External Validity (association)

3. Statistical Validity (association)

4. Internal Validity (causal)

Point estimate

estimated value of a population

Confidence interval

a range designed to include true population value at the time

Construct Validity

How well a variable is measured/manipulated

are they measuring what they think they are measuring?

Challenges the operationalization of the conceptual variable.

External Validity

Is the data generalizable?

challenges the sample, study setting, and context

Statistical Validity

How strong and accurate are the stats?

challenges the extent that the studies stats are precise, reasonable, and replicable

(point estimates, confidence interval)

Internal Validity

Any alternative explanations?

Challenges a study’s ability to rule out alternative explanations for a causal relationship between two variables

• correlations have low internal validity

How to achieve internal validity

• keep all other factors constant

• Random Assignment

• Within subject manipulation

internal validity and external validity

maximizing internal validity can harm external validity

Covariance/correlation

establish relation between variables (association)

Temporal Precedence

causal variable comes before outcome variable

Prioritize Validities

based on what the goals of the researcher are