The Basics of Science and Experiments in Psychology Flashcards
Falsifiability
Any hypothesis needs to be testable to prove if its false
Why is psychoanalitics not a scientific theory of psychology
Psychoanalysts source all mental issues to childhood trauma, if the person does not remember their trauma, they say it’s repressed
- this is not falsifiable
Deductive reasoning
An argument which the conclusion necessarily follows
- It is impossible for the conclusion to be false given the premises are true
- Used in mathematics
Inductive Reasoning
You can never be 100% sure of the conclusion
Ex. “all swans that have observed have been white”
DOES NOT = all swans are white
Can you prove a positive?
No
Fact
observation of the world around us
Hypothesis
Questions about an observation
Theory
A well-supported explanation based on lots of evidence
Law (science)
A detailed description of how something happens, usually supported by math
Dependant Variable
The thing that is effected or caused to occur
- the phenomenon we’re interested in
Independent variable
The variable that is expected to cause the behaviour
Operational definition
A description of a variable in themes of the operations used to establish or measure that variable
- Precise
- Practical
- Quantitative
- Good interobserver reliability
Histogram
A graphical representation of ranges of a variable, like a bar graph, but shows a continuity
We like this in psychology
Gaussian distribution (both normal and uniform)
the graph looks relatively the same on the left and right of the graph
Central Tendency
The most typical value of a data set
- Commonly called the average
Mode (statistics)
The value that occurs with the greatest frequency
Ex, 3,4,2,5,4,7,4,3
Mode = 4
What are the pros of using the mode? (Statistics)
- Works well with nominal frequency data (not numbers) ex. Dominant handedness of people
- Can produce sensible values
What are the cons of using the mode? (Statistics)
- Dependent on how you “bin” data
- Low sampling stability (fluctuates a lot from sample to sample)
- May not exist for certain data sets (Could have multiple modes or none at all)
Median
The middle number in an ordered set of data
- If it’s an odd number, find the middle two and take the mean of them
What is the pro of using the median
Robust to extreme values
Mean (statistics)
Add all the values together and divide by how many values there are
Advantages of using the mean? (Statistics)
Best sampling stability
Works with many statistical methods
Cons of using the mean? (Statistics)
- Value may not exist in the real world
- Assumes data is on, at least, an interval measurement scale
- Heavily affected by extreme values
Is an average an accurate representation of the group?
No, you need to look at the whole data set
How do you find standard deviation?
Calculate the deviations
Square them
Take the mean
Remove the square
Standard deviation = ((the sum of( x - mu )^2/N)^(1/2)
Skewness (statistics)
If the graph is skewed more to the left or right
Means and standard deviations are only useful for Normally distributed data
Which “average” should you use if your dataset is skewed?
The median
Do psychologists test populations or samples
They usually can’t test populations
Descriptive research
- Data tries not to interfere with how data arises in the real world
- Describes characteristics of a population
Correlational Research
- Looks for relationships between variables
- Uses descriptive research methods to obtain data on variables
Experimental Research
Manipulates variables in a controlled manner to isolate causes of some phenomena
Naturalistic Observation
Recording behaviour in real world settings without trying to manipulate the situation
- Have high External validity (the extent to which we can extend scientific findings to real-world settings)
- Can’t make causal inferences
- Act of observing might affect the behaviour of the observer (reactivity)
Reactivity
Behaving differently when you know you’re being observed
Case Studies (definition)
A research design that examines one (or a few) person in depth over an extended period of time
- Useful for existence proofs
- Often pretty biassed
- Weakest external validity
- Provides information about rare phenomena that is hard or unethical to study in labs
- Can’t establish causation
Anecdotes
Short, personal experiences
- Does not establish causation
- Can be inaccurate or biassed
- May ignore contradictory claims
NOT EVIDENCE
Surveys and Self-Reports
Using a questionnaire or interview to gather information about specific aspects of a participant’s background, experiences, or behaviour
- Good for collecting large sets of data
- Does not establish causation
- Can be massively influenced by how a question is worded
- Assumes that people know what they’re asking
Positive Impression Management
Trying to make ourselves look better than we are
Malingering
Tendency to make yourself appear psychologically disturbed to achieve some goal
Random Sampling
You need to have random people to do a survey to help balance out biases. Every person within a population has an equal chance of being selected
- Smaller surveys are more likely to be inaccurate
- Valuable to all types of data collection
test-retest reliability
When a test is readministered it should produce similar results to the first time it was administered
What is the problem with Rorschach (ink-blot) tests?
It has poor test-retest and inter observer reliability
Interobserver reliability
Two or more people using the same measure should arrive at the same conclusion
- Also called “interrater reliability”
Validity
The extent to which a measure assesses what it purports to measure
Ex. Polygraphs (tends to give false positives)
Type I Error
False Positive
“…[When] researchers conclude that the is a causal relationship between two variables when in fact there is not”
Type II Error
False negative
“…[When] researchers conclude that there it not a causal relationship between two variables win there is”
The correlation coefficient (r)
Has a value between -1 and +1
+1 = perfect correlation (both variables have positive correlation)
0 = no correlation
-1 = perfect correlation (the variables have a negative correlation)
Illusory Correlation
Perception of a statistical association between two variables where none exists
(Ex. the full moon makes people act crazy)
Confirmation Bias
The tendency to seek out information that supports our hypothesis and deny other evidence
Availability Heuristic
When you estimate the likelihood of an occurrence based on the ease which it comes to mind
The Third Variable
Two variables may be related to one another only because of a third variable
Observational Study
Researchers watch people in two groups (cannot establish causation)
Experimental Study
Researchers randomly assign people to groups (can establish causation)
Why does data fluctuate across data sets?
All possible variables cannot be controlled
Ex. Learning experience, genetics, imperfect measuring tools
This makes unsystematic error
What does something need to be statistical significant?
If they obtain a “p-value” less than 0.05
- Often doesn’t matter as much as people think
Random Assignment
Randomly sorting subjects into the experiment’s groups
(Not the same as random sampling, which has more to do with hoe subjects are chosen)
Confounding variables
Any difference between the experimental and control groups other than the independent variable
The Placebo Effect
“Placebo” - I shall please
When people believe something good will happen, they are more likely to feel it happening
The Nocebo Effect
Feeling negative effects because they are told there will be some
Belief that an otherwise harmless thing is harmful can make it harmful to you
Ex. Mass Psychogenic Illness
How many standard deviations hold 99.7% of data in a normal distribution?
3
How many standard deviations hold 95% of data in a normal distribution?
2
How many standard deviations hold 68% of data in a normal distribution?
1
Negative skew
Skewed left
-long tail points left
Symmetrical Data
Symmetric normal
-tails are balanced
Positive skew
Skewed right
-long tail points right
Correlational designs
used to examine the relationship between variables
Does the p-value really tell you if an experiment is valid or not?
Nope
Confounding variable
Any difference between the experimental and control groups other than the independent variable
Double-blinding
When neither the researchers nor the subjects are aware of who’s in the experimental or control groups