Chapter 8 Part 2 Flashcards
What is effect size?
Quantifying size diff between two groups to judge the significance of the results
What is an issue with hypothesis testing?
Only using critical/p-values without considering effect size and its effect on results
What is more likely to produce a significant sample size?
Small absolute difference between two means > more likely to be statistically significant with large sample than small
What are two ways to explain the difference between the two groups being compared?
Absolute diff between group means, standardized diff btw n group means
What is absolute effect size?
Diff between means expressed in the distributions of measurements
When is absolute effect size useful?
When variables have intrinsic and well-understood meanings
What is a downside to absolute effect size?
comparing two raw scores w/o thinking of distribution spread
What is standardized effect size?
Diff between group means
When is standardized effect size useful?
measurements have no intrinsic meaning, different scales of measurement
What does the standardized effect size serve the basis for?
Comparing differences across studies, used extensively in post-hoc meta-analyses
What is the standardized effect size equation mean?
difference between means is divided by the standard deviation to yield the standardized mean difference between groups
What does the standardized mean difference provide?
common ground for comparison
Why does the standard deviation need to be included in the calculation?
To correct for differences in the spread
When Cohen’s d be used?
When comparing mean scores of two groups
What is cohen’s d simply?
Difference between the two group means divided by the average of their standard deviations
How is cohen’s d arranged by size?
.2 is small, .5 is medium, and a .8 is large
How is cohen’s d calculated?
nothing more than the difference between two group means divided by their pooled standard deviations
An effect size of zero would indicate __________.
No effect of IV
Cohen’s d is used to calculate a ____ effect size….
standardized
What is a Type 1 error?
Rejecting the null hypothesis when it is in fact true
What does the level of significance of a statistical test also set?
The maximum probability of making a Type 1 error
In the case of errors what does an alpha also determine?
Probability of making a type 1 error
What is a type 2 error?
probability of rejecting the null when It is in fact false
What is the relationship between the type 1 and the type 2?
As the probability of the type 2 increases the probability of a type 1 decreases
What is the type 2 inversely related to?
level of significance
What does the type 2 error depend on?
Population mean
What is the correct decision probability for a type 1?
1-a
What is the correct decision probability for a type 2?
1-b
What is the first error that can be made in hypothesis testing?
Null hypothesis is true but the decision based on the testing is the null should be rejected
What is the second error that can be made in hypothesis testing?
Null is false but testing concluded is should be accepted
What is the power?
Probability of making a correct decision and rejecting the null when it is in fact false (no making a type 2 error)
What is the first thing that can affect power?
sample size
What is the second thing that can affect power?
Larger the n the greater the power of the test
What is the third thing that can affect power?
Significance level
What is the fourth thing that can affect power?
Higher the value of a (Type 1), higher the power of the test,
What is the fifth thing that can affect power relating to the fourth?
Increasing size of rejection region and decreasing size of non-rejection region
What is the fifth thing that can affect power?
The true value of the parameter being tested
What is the sixth thing that can affect power?
Greater the difference between the true value of a parameter and specified null, greater the power of the test
What is an important use of power in hypothesis testing?
Ensure that the sample size being considered is large enough for the purpose of the test
What happens if power is not used in hypothesis testing?
results will be inconclusive and effort/resources will have been wasted
Other things being equal, what will increase the power of a hypothesis test?
Increasing sample size
A researcher is planning a study and is considering various options for alpha. Selecting a larger alpha value will result in decreased ____
Beta
What is step 1 in hypothesis testing?
State claim/identify hypotheses
What is step 2 in ht?
Determine alpha
What is step 3 in ht?
Determine test
What is step 4 in ht?
Critical value of test
What is step 5 in ht?
Calculate test stat
What is step 6 in ht?
Compare test stat to critical value
What is the p-value?
Determine probability of the value of the stat given null is true
What is step 7 in ht?
Interpret decision
What happens if you have a large enough sample?
ANY difference can be significant
What is effect size?
Estimate of the degree to which the effect is present in the population
What is effect size simply?
Difference between uhyp and u true
What is u true?
Sample mean
What is u hypothesis
Value when null is true
What are the standard measures of effect size?
Cohen’s d, hedge’s g, and Pearson’s r
When do you use cohen’s d?
When you know the standard deviation
Why do type 1 errors occur?
Possible to draw non-representative sample at random from a population
How to control a type 1 error?
Changing alpha will change type 1 error
How to control a type 2 error?
Can’t directly control it but there are factors that influence it (a, sample size)
What power value is considered adequate?
.80
What are the four general factors that influence power?
Alpha, effect size, sample size, population variability
What happens when you increase alpha?
Decreases beta (type 2)
What happens with one tailed tests in the correct direction?
decrease beta
What happens with one tailed tests in the correct direction?
They increase power (relation to alpha)
What happens when you increase effect size?
You increase power
What happens if you increase sample size/decrease variability?
Power increases (type 2 decreases)
What happens if n value increases?
Standard error decreases, type 2 error shrinks
What happens if sample standard deviation decreases?
Standard error decreases (type 2 error decreases)
When do you want to estimate power?
Before you begin a study
What can a power curve help with?
Determining sample size needed for sufficient power