Psych Stats Flashcards
Descriptive Statistics
summarizes or describes a set of data with numbers (means, standard deviation, median etc.)
Inferential Statistics
The conclusions you draw from numbers/the descriptive stats - inferences about the world
Independent Variable
- what you are manipulating
- on the x axis
- want this to be the only thing different between groups
Dependent variable
- what is measured (the outcome)
- the effect
Continuous Variable
not set categories; full range of values
examples: height, skin tone, weight
Discrete Variables
Specified values, whole numbers, yes or no, categories
Examples: yes/no, gender, colors,
Nominal
A category or name; always discrete
ex: colors, majors, types of sleep (REM, Alpha 1 etc).
Ordinal
Rankings of things: but may not be evenly numerically spaced (always discrete)
ex: positions in a race, levels of emotion
Interval
- Equally spaced rankings (continuous)
- can not multiply or divide them
- no meaningful zero
ex: IQ test, celsius or fahrenheit
Ratio
- no negative numbers
- meaningful zero
- often continuous
ex: height, time, length
Meaningful Zero
when 0 means the absence of the thing being measured (even if not attainable)
ex: 0 Kelvin is no temperature, 0 feet is no feet tall
Between Group Designs
Different people in different conditions – comparisons between groups
In Group Designs
All people do both conditions: comparisons are made within the same groups
Correlation Studies (NonExperimental)
- Methods: surveys, observations
- No manipulation of the independent variable
- No random assignment
- find associations not causation (correlation does not = causation)
Experiments
- random assignment
- manipulated IV and controls
- once variable measured (DV)
- purpose: to establish cause and effect – able to make causal statements
Extraneous variable
an outside variable that may influence the dependent variable
Operational Definitions
How we choose to measure or manipulate variables of interest
Measures of Central Tendency
- Mean
- Median
- Mode
Mean
- average of the data points
- add all up and divide by number of data points
- evenly balances deviations (data points add up to the same number on both sides)
- HIGHLY affected by outliers
Median
- the middle score in a data set
- number of points +1 / 2 to find the median number (not value but place in data set)
- not very effected by outliers
Mode
- What occurs most frequently in a data set
- highest bar in a histogram
- may not be a mode in a data set
- may be more than one mode
- not effected by outliers
Variability
How the data varies in a set
Range
- The min score to the max score
- heavily affected by outliers
IQR
- 75th percentile minus the 25th percentile
- the middle 50% of the data
- not as affected by outliers
Variance
Formula: sum of (a data point - the mean) squared – all divided by the number of data points
Standard Deviation (SAMPLE)
Square root of Variance
Sqrt of : sum of (a data point - the mean) squared – all divided by the number of data points
Standard Deviation (POPULATION)
Formula: sqrt ( sum of (a data point - the mean) squared – all divided by the number of data points + 1)
- used for inferential statistics
Population
- The group of people on the whole you are interested in studying
- mean is called muew (long u)
- results: parameters
Sample
- the group you are actually studying (from the population of interest)
- mean is X with a line over it or M
- Results: called statistics
- Sample mean: good, unbiased estimator of underlying population mean
Effect Size (EDIT)
- indicates the magnitude of an effect
- units: ???
Cohen’s d
tells us how two distributions overlap
- small effect: d= 0.2
- medium effect: 5= 0.5
- large effect: d = 0.8
report as positive
mean difference between groups
Point Estimate
Something we get from a data set (estimating population mean from sample mean)
- always in the middle of the confidence interval: has error bars on either side
Interval estimate
range of plausible values of the population parameter
confidence intervals
allows you to infer something about the mean of an unmeasured population
- smaller is better: more precise estimate
- can be used for 1) mean 2) standard deviation 3) difference between means of two populations 4) effect size
Line Graphs
- imply continuous data
- Scale IV and DV (interval, ratio)
- often time series: measurements taken over regular time intervals
Scatterplot
- 2 scale variables
- each dot is a data point
- specific to correlational / regression designs
Bar Graphs
- Nominal or Ordinal IV (category on x-axis)
- Scale DV (number on y-axis)
- Good to show differences between means
Normal Distribution
symmetrical bell curve
- unimodal (one peak) – half data above and below it
Bi Modal Distribution
more than one mode (2+ peaks)
Positive skew
Clump towards the bottom
- floor effect: everyone gets low measurement
Negative Skew
Clump towards top
- ceiling effect: everyone close to top of measurement scale
Skew Affects
positive:
- higher mean, slightly higher mode
negative:
- lower mean, slightly lower mode
Outlier
An extreme score
