Causality and the 3. variables Flashcards
Why is it critical to observe the causality for the observed change in the dependent variable?
To truly grasp the relationship between the independent and dependent variable we must investigate the causality for the change. We cannot determine the effect of the independent variable if we have not considered other potential factors. If the authorities put up several signs along a stretch of road to reduce speeding, followed by drop in the level of speeding we cannot say that it was the new road signs that generated the desired results without inspecting the causality and other potential factors more closely. People reducing their driving speed could also be down to increased presence of police officers, higher fuel prices, icy roads, etc…
Name some general factors which can affect the dependent variable
Some discrete event may alter the dependent variable (DV), e.g., a car crash or heavy snowfall reduces speeding.
Things change over time, this includes DVs. This is called maturation.
If test subjects know they are being tested they often change their behaviour. Test effects are very real and should be considered.
Faulty test equipment could be a concern.
The regression problem, i.e., not collecting a large enough and representative sample
What is the direction problem?
If we want to observe the effects A has on B, we cannot always be certain if it is actually B that influences A. One example would be the aggression levels in kids, and kids excessively watching violent videos. Which causes the other may not be immediately apparent.
What are 3rd variables?
3rd variables show strong correlation with the measured variable while there exist no causality. Consider the two highly correlated phenomena of ice cream sales and shark attacks.
How do you protect your research against 3rd variables?
Detection and elimination of 3rd variables can be conducted through statistical procedures to highlight the true effect of individual variables. You can also use control groups and structured well planned data-collection.
What is correlation?
Correlation is a statistical term describing the degree to which two variables move in coordination with one another.
NOT the same as causality!!!
