A chart showing the standard deviation from the normed center of the plotting
Nominal Scale
Least Precise
Puts Variables into Categories
Like race, socio-economic
Chi Squared
Designed to measure the significance between 2 variables
Ordinal
2nd Least Precise measurement
Categorizes by magnitude highest to lowest
Class rank
No real detail about why the grades are the way they are, just that they are
Interval
The Most Used for of Measurement in Education
Catergories by magnitude and understanding the why of the intervals in between
No Absoulte Zero...there will always be something
Uses Pearson r for correlation
Be wary of using this to make predictions...not always cause & effect
Ratio
Most Precise
Categories of data showing magnitude
Intervals are the same size and there is true zero
Data can be manipulated
Descriptive Statistics
Looking at the numbers and doing things like:
Averages
Median
Mode
Correlation is not Causation
Cummulative Percentage Chart
Spearman is all about the rank order of things
Central Tendency
Single Value that is considered the most typical...a norm at which to determin standard deviation
Validity
Using the right tool to measure correctly
Content Validity- the items on the instrument refelct the content
Construct Validity- the extent to which a higher order construct is represented in a study (Help Seeking, stress, dyslexia...etc.)
Reliability
Having reliable scores
Having temporal stability- stability over time
Angoff Method- standard setting
Internal Validity
The ability to infer that a causal relationship exists between 2 variables
External Validity
The extent to which the study can be generalized and applied across populations
Errors of Measurement
Standard Deviations from the norm
Always going to have errors in samples, mathematical
The small the standard deviation the more accurate the measure
Standard Error of Measurement
The standard deviation of a sample population
Null Hypothesis
A statement about a population parameter some condition concerning the pop. parameter is true.
prediction of no difference in a study when a new treatment is given
Rejecting the Null Hypothesis
When there is a relationship between populations due to the given treatment you reject the null hypothesis
Accepting the Null Hypothesis
Means you are admitting that there is no relationship between populations given the treatment
Inferential Statistics
Inferring information from a sample to a larger population
Types 0f Samples
Random Sample
Sampling Interval
Stratified Sample
Cluster Sample
Convenience Sampling
Quota Sampling
Purposive Sampling (Judgemental)
Snowball Sampling
Type One Error
To say it had an effect, but it really didn't have one
Type 2 Error
To say the treatment did not have an effect when it really did
Power
The larger the sample size the more power a test has...it is more sensitive. The Bigger the sample size the better...less likely to have a Null hypothesis
Effect Size Indicator
A statistical measure to show strength of a relationship
Practical Significance
A conclusion made when a relationship is strong enough to be of practical importance
t Test
Used to compare 2 group means
Used to determine if the difference between sample populations was created merely by chance errors or really because of the treatment
ANOVA
ANOVA is used to compare one or more group means
Can compare more than 2 groups
Gives a p value
Scientific Method
Question
Research
Hypothesis
Test
Analyze
Test Again
Report
Alternate Hypothesis-Example
Male and Female population's means on SAT tests are different
Null Hypothesis- Example
Male and Female population means on not different on the SAT