Chapter Seven
Decision Table, P Value less than Alpha
Reject the null, statistically significan, may be clinically significant if the experts say, find a relationship. Correct or Type one (chance only)
Effect size/ Sample Size
as effect size increases, sample size decreases
Power size/ sample size
As power increases, sample size is increased, and the likelihood of rejecting the null hypothesis correctly is increased. The larger the sample size the greater the power of the study (more likely to accurately reject the null hypothesis)
Ho True, Accept Ho, Fail to Reject the Null
Correct acceptance of Null
Decision Table, P Value greater than Alpha
Fail to reject the null, not statistically significant not clinically significant, no relationship. Correct to Type Two (usually too small of sample, power error)
Power Analysis
How sample sizes are calculated. Determines what sample size will ensure a high probability that we correctly reject the null that there is no difference between groups. Pick power (usually 0.8), threshold for significance (usually 0.05), and then effect size which is the overlap
Power
The ability to find a difference or an association when one actually exists The likelihood of rejecting the null hypothesis correctly; that is you say there is a relationship and difference and you are correct
Alpha
The chance of making a type one error
Beta
The chance of making a type two error
Type Two Error
The error made when a researcher accepts the null incorrectly, missing an association that is really there (sometimes called a power error because the researcher may not have enough power to find an association that really exists). Ho False. Usually occurs because sample was not large enough
Type one Error
The error when a researcher incorrectly rejects the null hypothesis, when he or she concludes there is a significant relationship but there really is not. Ho true
Ho False, Accept Ho, Fail to Reject the Null
Type Two Error
Ho True, Reject the Null
Type one Error= Alpha
Example of Type One Error and Type Two Error
Type one error (false positive): to a male, saying they are pregnant Type two error (false negative): to a pregnant female, saying you're not pregnant
Type Two Power
You will either correctly identify (i.e, reject Ho correctly = power) or you will incorrectly miss it (type two error) Power (correctly rejecting) + Beta (chance of making a type two error) = 100% (Therefore if there is a 80% chance that you are correct and find the relationship (power), the chance of a type two error is 20%)
If The Null is not True
there is a relationship between the variables you can conclude one of two things: You can fail to reject it. In this case you are incorrect and are making a type two error. The probability of doing so is equal to Beta (usually 0.20 or 20%) You can reject it correctly. The probability of reaching this conclusion is 1- B (usually 80%), which is also the power of your study
Effect Size Numerical
weak is less than 0.3, moderate is 0.3 to 0.5, strong is 0.5
Ho False, Reject the Null
correct rejection of null= power= 1- Beta
Effect Size
The extent to which a difference or relationship exists between variables in a population (the size of the difference you are attempting to find)
Sample Size
an adequate sample size is largely determined by the size of the difference between group means within the population (effect size), you are attempting to find and the power needed to accurately find it
If the Null Is True
there is no relationship or difference you can conclude one of two things: You can fail to reject it. You are correct in this situation, and the probability of reaching this conclusion is 1 - alpha (usually 0.05) or 95% Or you can reject it. Then you are incorrect and are making a type one error. The probability of reaching this conclusion is equal to alpha, usually 5%
