Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
46 Cards in this Set
- Front
- Back
Alternative Hypothesis
|
The claim trying to be proved
|
|
Ha or H1
|
Alternative Hypothesis
|
|
Null Hypothesis
|
The opposite of the alternative hypothesis
|
|
Ho
|
Null Hypothesis
|
|
Reject Ho
|
Enough data evidence that Ha is proven beyond a reasonable doubt and Ho is false
|
|
Fail to reject Ho
|
Not proof beyond reasonable doubt that Ha is proven beyond reasonable doubt, does not imply Ho is true
|
|
Type 1 error
|
Reject Null Hypothesis (Ho) when it is in fact true-say you have proven something (Ha) that is not true-
More serious error than Type 2 |
|
Type 2 error
|
Failure to reject the Null Hypothesis (Ho)-What you are trying to prove (Ha) is true, but you did not find enough evidence to support-Less serious error than Type 1
|
|
Level of Significance (Alpha)
|
Maximum allowable chance of making a type 1 error. (.05 most common)
|
|
P-Value
|
Based on observed data, smallest level of significance that allows Null (Ho) to be rejected
|
|
If P-Value is smaller than Alpha
|
Proof beyond reasonable doubt that Ha proven and reject Ho
|
|
When is Alpha level set
|
Before looking at data
|
|
When is P-Value determined
|
After data is analyzed
|
|
Mean & Median are both measures of:
|
Center
|
|
Variance
|
Standard Deviation squared
|
|
Standard Deviation
|
Typical amount population values differ from population average-how spread out population is
|
|
Empirical rule for mound shaped populations
|
68% within 1 SD +/-
95% within 2 SD +/- 99.7% within 3 SD +/- |
|
Interquartile Range
|
25%ile to 75%ile
Distance from Q1 to Q3 |
|
Small Interquartile range signifies
|
Consistent population
|
|
Best way to describe spread when population not mound shaped
|
Median & Interquartile range
|
|
Best way to describe typical spread of values for mound shaped population
|
Mean & Standard Deviation
|
|
Best way to describe typical spread of values for skewed population
|
Interquartile Range & Median
|
|
Shapiro-Wilk P-Value
|
The probablity of seeing a histogram as skewed as your sample. Larger
S-W numbers are better. |
|
Shapiro-Wilk < or = 0.5
|
Proof beyond reasonable doubt of skewed population
|
|
Shapiro-Wilk > or = 0.5
|
Proof beyond reasonable doubt of Mound shaped population
|
|
Box plot symmetry signifies
|
Indication of mound shaped population
|
|
Hypothesis Tests: Three common forms:
|
1. Ha typical value of population not equal to the test value (2 tailed)
2. Ha typical value of population > test value. 3. Ha Typical value of population < test value |
|
2-Tailed: Difference can be on__side of test value
|
Either
|
|
1-Tailed: Observed data will support difference to be___
|
> or < but not both
|
|
SPSS: Sig=
|
P-Value
|
|
Both 1-Tailed P-Values should add up to___
|
One
|
|
Confidence Intervals for Mean and 2-Tailed Hypothesis test will always___with each other
|
Agree
|
|
1-Tailed Hypothesis test can have____
data to Confidence Interval (rare) |
Contradictory
|
|
1-Tailed Hypothesis tests are more___
|
Powerful
|
|
Confidence Intervals give a range that is_____
|
Easy to understand
|
|
1-sample T-Test assumptions
|
1. Random sample (ideal) or at least try to eliminate bias.
2. Mound shaped population or n.>/= 30 |
|
If n>/= 30 but sample is still skewed then better to use_______rather than T-Test
|
Sign Test
|
|
What does 5% trimmed mean do
|
Throws out largest and smallest of each 20 samples to eliminate outliers
|
|
Sign Test for population median assumption (only one)
|
Random sample
|
|
Practical significance is a _______of statistical significance
|
Subset
|
|
Confidence Intervals:
Assumptions (three) |
1. Random sample
2. n>/= 30 or mound shaped population 3. Same as T-Test |
|
To check assumptions of mound shaped population when n<30
(three things) |
1. Knowledge of researcher
2. Shapiro-Wilk Test P-Value >/= .05 3. Histogram/Boxplots |
|
Paired Samples T-Test:
Assumptions (two) |
1. Random sampling
2. Population of differences must be mound shaped or n>30 however, if sample is skewed the median is more meaningful |
|
Skewed populations of differences:
Best Test: |
Sign Test (median)
|
|
Matched pair data: which test to use:
1.) Mound shaped: 2.) Skewed: |
1. Paired Samples T-Test
2. Sign Test |
|
Confidence Interval for difference between matched pair populations:
What if zero is included in the range___ |
Plausible that two averages might be the same
|