Other methods (like observation or surveys) can show relationships, but they cannot prove causation because of one major issue:
👉 Confounding variables
⚠️ Confounding Variables (The Hidden Problem)
A confounding variable is something extra that influences the results without the researcher intending it to.
Simple Definition (Easy Read)
-
A confounding variable = a hidden influence
-
It changes the result without you realizing
Dog Example
You want to test:
“Does a stranger cause more barking than a friend?”
But what if:
-
The stranger is louder than the friend
-
The owner looks nervous
-
The test happens in a new place
👉 Now you don’t know what caused the barking.
Was it:
-
The stranger?
-
The noise?
-
The owner’s anxiety?
That’s the problem.
🎯 Controlling Extraneous Variables
Extraneous variables = all variables that are not the IV but could affect the DV
The goal in experiments is:
👉 Turn extraneous variables into constants
🧠 Key Control Methods (Clear + Practical)
1. Standardisation
Make everything the same every time.
-
Same room
-
Same distance
-
Same time of day
-
Same instructions
✔ This reduces differences between trials
2. Randomisation
Randomly assign participants (or animals) to groups.
✔ Prevents bias
✔ Spreads confounding variables evenly
3. Control Groups
Compare against a baseline condition
-
No “treatment”
-
Normal situation
✔ Helps isolate the effect of the IV
4. Counterbalancing (Advanced but useful)
Change the order of conditions
Example:
-
Some dogs see the stranger first
-
Others see the friend first
✔ Prevents order effects (learning, fatigue)
5. Blinding
Hide information from participants or observers
-
Single-blind: participant doesn’t know
-
Double-blind: neither participant nor researcher knows
✔ Reduces bias and expectation effects
🧪 Why Observational Methods Are Not Enough
Observation = watching behavior without control
Problem:
You cannot control variables
Example:
You observe:
-
Dogs bark more at strangers
But you don’t know:
-
If the stranger looked threatening
-
If the owner reacted
-
If the dog had past trauma
👉 So you can say:
-
“There is a relationship” ❌
-
But NOT “This causes that” ❌
🐶 Real Research Link
The kind of structured testing you’re describing is similar to:
Originally used with human infants, later adapted to dogs to study:
-
Attachment
-
Security
-
Response to strangers
✔ Shows how controlled environments reveal emotional behavior
🧠 Ethical Concerns in Animal Experiments
This is especially important for your work (mental health, disability, advocacy).
⚖️ The “Four Rs” (Animal Ethics)
1. Reduction
Use the fewest animals possible
2. Refinement
Reduce stress and improve conditions
Example:
-
No real aggression
-
Use actors or mild stimuli
3. Replacement
Use alternatives if possible
-
Simulations
-
Observational studies
4. Responsibility
Ensure:
-
Proper care
-
No long-term harm
🚫 What Researchers Must Avoid
-
Causing trauma
-
Long-term anxiety
-
Physical harm
-
Breaking trust (especially in bonded animals like dogs)
❤️ Important Real-World Insight
Your earlier point about grief and pets is actually supported by research:
👉 The human–animal bond can be as strong as human relationships
This is why:
-
Loss of a dog can feel like losing a family member
-
Experiments involving pets must be especially careful
🧩 Bringing It All Together
To prove causation in psychology, you must:
-
Change one variable only (IV)
-
Measure its effect (DV)
-
Control everything else
-
Remove confounding variables
👉 That’s what makes the experimental method powerful—and difficult.
🧠 1. The Core Idea (Refined)
The experimental method tries to prove:
👉 “Does one thing cause another?”
But your example shows something deeper:
👉 Behaviour is often two-way (reciprocal)
-
The dog reacts to the stranger
-
The stranger reacts to the dog
-
The dog then reacts again
This creates a feedback loop, not a simple cause → effect.
🔁 2. Reciprocal Behaviour (Your Key Insight)
What you described fits with:
👉 Reciprocal determinism (from Social Learning Theory)
Simple Explanation
-
Behaviour affects environment
-
Environment affects behaviour
-
Both keep influencing each other
🐶 In Your Scenario
-
The stranger feels fear
-
That fear changes body language (tense posture, hesitation)
-
The dog senses this (dogs are highly sensitive to tone, movement, scent)
-
The dog becomes alert or defensive
-
The stranger becomes more afraid
-
The dog escalates further
👉 This is a cycle, not a single cause
⚠️ 3. Why This Creates Confounding Variables
This is where experiments become difficult.
You originally had:
-
IV = stranger vs friend
-
DV = barking
But now we must consider hidden variables:
New Confounding Variables
-
Stranger’s fear level
-
Body language (tense vs relaxed)
-
Tone of voice
-
Past experiences (dog and human)
-
Owner’s reaction
👉 These are not controlled automatically
🎯 4. Controlling This in an Experiment
To keep the experiment valid, researchers must separate variables carefully
✔ Improved Experimental Design
Independent Variables (split properly)
Instead of one IV, we refine it:
-
Familiarity
-
Emotional Behaviour of Stranger
Dependent Variable
-
Bark frequency
-
Body posture (tail, ears, stance)
-
Distance maintained
-
Stress signals
Key Control Strategy
👉 The same actor performs each condition
-
Same clothing
-
Same distance
-
Same movements (scripted)
✔ This removes variation in behaviour
Additional Control
-
Train the actor to display:
-
“Fear” in a consistent way
-
“Calm” in a consistent way
👉 This prevents random emotional differences
🧪 5. The Reality Problem
Even with control, one issue remains:
👉 You cannot fully control real fear
As you said:
“It is easier said than done not to show fear.”
That is completely accurate.
🧠 Why This Matters
-
Real fear is physiological (heart rate, smell, micro-movements)
-
Dogs may detect:
-
Adrenaline
-
Subtle tension
-
Eye contact
👉 So even trained actors may not fully replicate real fear
🔍 6. What This Means for Research
Because of this, psychologists often:
Combine Methods
-
Experiment (for control)
-
Observation (for realism)
✔ Example Approach
-
Lab experiment (controlled acting)
-
Natural observation (real-life dog encounters)
👉 Compare both sets of results
🐾 7. Ethical Considerations (Updated with Your Scenario)
Your version adds more ethical complexity
🚫 Risks
-
Dog becomes stressed or defensive
-
Stranger experiences genuine fear
-
Escalation could occur
✔ Ethical Adjustments
-
No real aggression (only mild cues)
-
Immediate stop if stress signs appear
-
Use trained dogs with calm temperaments
-
Professional supervision
❤️ 8. Real-World Insight (Very Important)
Your statement reflects real life:
Fear is not easily controlled—whether facing humans or animals.
This links to:
👉 Which means behaviour is not just situational—it’s personal and emotional
🧩 Final Understanding (Clear Summary)
Your improved model is:
👉 Behaviour is influenced by:
-
The situation (stranger vs friend)
-
The emotional state of both individuals
-
The interaction between them
✔ Key Takeaway
The experimental method tries to simplify reality:
But your example shows:
👉 Real behaviour is dynamic, emotional, and reciprocal
