Notes on Variables & Research Design from Transcript
Independent and Dependent Variables
Independent Variable (IV) − CAUSE; the variable that researchers manipulate or consider as the cause in the study.
Dependent Variable (DV) − EFFECT; the variable that is measured and observed as the outcome.
In a typical experimental design: IV is manipulated and DV is measured.
In a correlational design: no variables are manipulated; the study describes relationships between variables.
Key definitions from transcript:
Independent Variable = MANIPULATED
Dependent Variable = MEASURED
Mathematical framing (typical linear model):
This expresses DV as a function of IV plus error.
When a study is correlational (no manipulation): both Interest and Motivations can be treated as DVs; no IV is actively manipulated.
Example interpretations from slides:
Page 5–6: Correlational study with Interests and Motivations; both treated as DV.
Page 7–8: IQ and Resilience among psychology students; no IV manipulated; both treated as DV.
Page 9–10: Salary Increase and Promotion affecting Work Performance; both Salary Increase and Promotion can be manipulated (as IVs) with DV = Performance.
Practical note:
In correlational studies, causality cannot be inferred from the observed association between variables.
Study Designs: Correlational vs Experimental
Correlational study characteristics:
No manipulation of variables (e.g., Interest and Motivations).
Relationship observed between two or more variables.
Both variables can be considered dependent because they are not experimentally controlled as causes.
Experimental study characteristics:
Manipulation of one or more independent variables (IVs).
Observation of effects on a dependent variable (DV).
Examples from transcript:
Page 6: Interest and Motivations in a correlational design; both are DV.
Page 8: IQ and Resilience in a correlational design; both are DV.
Page 10: Salary Increase and Promotion as IVs; DV = Work Performance.
Important caution:
In experiments, random assignment helps ensure that observed effects are due to manipulation of IVs rather than confounding factors.
Variable Types Overview (as covered in the transcript)
Constant Variables
Attribute Variables
Discrete Variables
Continuous Variables
Covariates Variables
Dichotomous Variables
Latent Variables
Manifest Variables
Extraneous Variables
Confounding Variables
Categorical Variables
Other Types Of Variables (graphic slide in the transcript; content not legible in provided text)
Constant Variables
Definition: Variables that do not change during the experiment.
Purpose: Control for potential alternative explanations by holding these factors constant.
Example context from transcript:
Breast Feeding and IQ study: to attribute IQ differences to feeding type, other factors (socioeconomic status, mother's age, family setup, etc.) are turned into Constants so they do not vary.
Another example:
Artificial Lights and Growth: factors like soil quality and watering are kept constant during plant growth experiments.
Extraneous and Confounding Variables (overview and examples)
Extraneous Variables: Variables not intentionally studied but that can affect results.
Examples from transcript: memory capacity affecting test performance; test-taking time of day; test anxiety; level of stress.
Confounding Variables: A special type of extraneous variable that directly affects how the IV acts on the DV.
Definition from transcript: An extraneous variable that could influence the DV and also correlates with or causally affects the IV.
Important distinction:
All confounding variables are extraneous, but not all extraneous variables are confounding.
Diagrammatic relation (as per Scribbr slide in transcript): IV → DV with Extraneous and Confounding variables interacting with the path from IV to DV.
Examples: Caffeine and Test Scores (illustrative study design)
Study setup (from transcript):
Experimental Group: 50 students who drink coffee.
Control Group: 50 students who drink water.
Preliminary result: Experimental group scored higher.
Important follow-up (lesson from later text):
True experiments require random assignment to control for confounding factors.
Subsequent difficulties (e.g., the same group failing later) may be due to confounding factors not initially controlled.
Possible confounding variables discussed:
Time spent preparing for the exam.
Sleep quality or amount.
Smoking status and nicotine effects.
Takeaway:
A true experiment should include random assignment to minimize confounds and establish causal inference.
Covariates and their role
Covariates: Variables included in a study to create interactions with IV and DV; they are not the primary treatments but can influence outcomes.
Example from transcript:
Plant drought tolerance study: level of drought is the treatment, but plant size is a covariate that also influences tolerance.
Practical use: Covariates are statistically controlled for to isolate the effect of the IV on the DV.
Attribute Variables
Characteristics that cannot be manipulated (inherent or pre-programmed).
Examples from transcript:
Gender, race, psychological condition, intelligence, creativity, anxiety, learning styles, etc.
Key point: These variables are fixed attributes of individuals and are not experimental manipulations.
Discrete vs Continuous Variables
Discrete Variables
Qualitative or quantitative values obtained by counting.
Measured as whole units; cannot take fractional values between adjacent values.
Examples from transcript: countable items such as number of people, coins, etc.
Continuous Variables
Have an infinite number of possible values within a range; can take fractional values.
Examples from transcript:
Age: 25 years, 10 months, 2 days, 5 hours, etc.; effectively continuous.
Practical examples from transcript:
Temperature (Celsius, Fahrenheit, Kelvin) is continuous because it can be subdivided into decimals.
Length of a film is continuous (time is effectively infinite for practical purposes).
Pages of a book are discrete because you can count them.
Number of males and females in a classroom is discrete because it is a count.
Latent and Manifest Variables
Latent Variables
Cannot be directly measured or observed (e.g., personality traits, anxiety).
Inference is made from their effects on observable variables.
Measurement error is a concern: scores may vary across occasions or forms.
Measurement model example (common in psychology):
Observable indicators: X1, X2, …, Xk relate to latent factor L via
Manifest Variables
Directly measurable or observable indicators used to infer latent constructs.
Example from transcript (customer satisfaction): observable metrics include sales numbers, price per sale, regional trends, customer demographics, return rates, and consumer ratings to infer the latent construct of customer satisfaction.
Categorical Variables
Definition: Qualitative variables that describe quality or category rather than numeric amount.
Examples from transcript:
Eye color (blue, green, brown, hazel)
States (Florida, New Jersey, Washington)
Dog breeds (Alaskan Malamute, German Shepherd, etc.)
Dichotomous Variables
Definition: Variables with only two possible outcomes.
Examples from transcript:
Heads or Tails; Male or Female; Rich or Poor; Pass or Fail; Alive or Dead; Day or Night.
Constant, Extraneous, and Confounding Review (quick recap)
Constant Variables: kept fixed to prevent alternative explanations.
Extraneous Variables: not of primary interest but can affect results (e.g., memory capacity, time of day, stress).
Confounding Variables: extraneous variables that systematically affect the DV and are related to the IV; threaten causal interpretation.
Random assignment is emphasized as essential for true experiments to minimize confounds.
Other Notes and Practical Takeaways from the Transcript
In the SFG example (Page 11–12):
The study asks whether a systemic language grammar approach will affect intercultural competence over one semester.
The independent variable is considered to be the SFG approach, which can be manipulated into formats such as Collaborative, Oral, Written.
Answer from slide: D. None of the Above; SFG is the IV here since it can be manipulated into those formats.
The transcript repeatedly reinforces that IVs are manipulable factors, while DVs are the outcomes measured to assess the effect.
For practical research planning, always:
Identify IVs and DVs clearly.
Decide whether the study will be correlational or experimental.
Consider potential extraneous and confounding variables and plan control measures.
If using covariates, determine how they will be accounted for in analysis.
Use random assignment in true experiments to support causal claims.
Quick Practice Questions (from transcript concepts)
In the motivation study, which variables are DV if Interests and Motivations are being observed?
Answer: Both are considered DV in a correlational design.
If a study manipulates Salary Increase and Promotion and measures Work Performance, which are the IVs and DV?
Answer: IVs = Salary Increase and Promotion; DV = Work Performance.
What is a key difference between Extraneous and Confounding Variables?
Answer: Extraneous variables are any variables not of interest that can affect results; confounding variables are a specific type of extraneous variable that correlates with or causally affects the IV and directly influences how the IV affects the DV.
Give an example of a latent variable and one or two indicators (manifest variables).
Example latent: Anxiety; indicators: self-report anxiety scale scores, observed stress behaviors, physiological arousal measures.
Why is random assignment important in experiments?
Answer: It helps ensure that observed effects are due to the manipulation of the IV rather than preexisting differences between groups, reducing confounding influences.
Closing reflections from the transcript
The material emphasizes clear definitions of IVs and DVs, proper study design, and the importance of controlling or accounting for extraneous and confounding variables to support valid conclusions.
It also highlights a range of variable types that researchers must consider in planning, executing, and interpreting research.
FAQ-style clarifications (based on the slides)
What is the IV in the 'systemic language grammar approaches (SFG)' study on intercultural competence?
The independent variable is the SFG approach, which is manipulated into Collaborative, Oral, Written formats.
Can a correlational study have a true IV?
In a correlational study, the variables are not manipulated as IVs; both are typically treated as DV with observed relationships, not causal manipulation.
How are latent variables typically studied?
Latent variables are inferred from multiple observable indicators (manifest variables) using a measurement model such as Xi = \lambdai L + \epsilon_i.