Term
|
Definition
| find the best fitting straight line that relates 2 dependent samples in populations |
|
|
Term
|
Definition
| describe the strength of the linear relationship |
|
|
Term
| a correlation coefficient of +/- 1 indicates: |
|
Definition
|
|
Term
| involes the linear relationship btwn one one dependent variable and multiple dependant variables |
|
Definition
|
|
Term
| used often in epidemiology studies: |
|
Definition
|
|
Term
|
Definition
| score-mean/SD: Z score of +1 = 1SD above the mean |
|
|
Term
| which test? two independent grps w subjects randomly assigned to groups |
|
Definition
|
|
Term
| 3 assumptions in the T test: |
|
Definition
1. scores represent normal distribution
2. sbj randomly selected/assgnd grps
3. variences of the 2 grps are equal |
|
|
Term
| numerator of the T test represents: |
|
Definition
| difference btwn grp means |
|
|
Term
| denominator of the T test represents: |
|
Definition
|
|
Term
| formula for the degrees of freedom: |
|
Definition
| df= total # of df - # of grps |
|
|
Term
| one vs two tail tests: which is less likely to commit a Type II error? |
|
Definition
| less likely to commit a type II error using a one tailed test |
|
|
Term
| useful test for when the experimental design is one group that receives before/after tx: |
|
Definition
|
|
Term
| Z scores are appropriate for alrge samples but the _____ is more commonly used for small samples (n<30) |
|
Definition
|
|
Term
| max # of grps you can compare w a T test: |
|
Definition
|
|
Term
| a stat method used to test for differences among three or more tx means: |
|
Definition
|
|
Term
|
Definition
normally distributed data
no sig diff btwn grp variances |
|
|
Term
| instead of a T-test the ANOVA has an: |
|
Definition
|
|
Term
|
Definition
| F = btwn grp variance/within grp variance |
|
|
Term
| ANOVA used to investigate the effects of ea of several independent variables and the joint effect of these variables acting together: |
|
Definition
|
|
Term
| the difference btwn the tx variables are called: |
|
Definition
|
|
Term
| the ______ effect is due to the joint action of the tx variables: |
|
Definition
|
|
Term
| multiple comparison test conducted after the primary analysis has rejected the null hypothesis: |
|
Definition
|
|
Term
| multiple comparison test: comparisons of hypotheses specified before the analysis commenced; appropriate even when F test is not significant: |
|
Definition
|
|
Term
| 4 ex of multiple comparison tests: |
|
Definition
1. honestly significant difference
2. newman-keuls
3. multiple range test
4. scheffe's method |
|
|
Term
| analysis used when levels of an independent variable are ordered along a continuum. |
|
Definition
|
|
Term
| test most used for frequency or count data; ex: (# of men,# of deaths, # women w/wo dz) |
|
Definition
|
|
Term
| basic premise of Chi sq: number of sample observations in each category of ea variable: |
|
Definition
|
|
Term
| basic premise of CHi sq: # of sample observations in ea category if the null hypothesis is true: |
|
Definition
|
|
Term
|
Definition
1. goodness of fit
2. test of independence
3. test of homogeneity
4. McNemar test |
|
|
Term
| proportions of a normal curve: +/- 1SD = what% of scores? |
|
Definition
|
|
Term
| proportions of normal curve: what % is within 2sd of mean? |
|
Definition
|
|
Term
| proportions of normal curve: what % is within 3SD from mean? |
|
Definition
|
|
Term
| a Z score of 1 indicates: |
|
Definition
| the score is 1 SD above the mean |
|
|
Term
what % of 14yo have cholesterol of >240?
mean cholesterol @ 14 = 170+/-35 |
|
Definition
| 240-170/35 = 2SD above mean; 2.325% of 14yo have chol>240 |
|
|
Term
| likelihood that any one event will occur, given all possible outcomes: |
|
Definition
|
|
Term
| SD of a theoretical sampling distribution of means: |
|
Definition
| standard error of the mean |
|
|
Term
| null hypotheses also reffered to as: |
|
Definition
|
|
Term
| concluding that a difference exists when none actually does |
|
Definition
|
|
Term
| concluding that no difference exist when one actually does: |
|
Definition
|
|
Term
| used to estimate the sample size needed to obtain a desired level of power before data collection: |
|
Definition
|
|
Term
| probability that a test will lead to rejection of the null hypothesis: |
|
Definition
|
|
Term
| useful for determining if a Type II error was commited when you have a nonsignificant finding: |
|
Definition
|
|
Term
| 4 determinants of statistical power: |
|
Definition
1. significance criteria
2. variance
3. sample size
4. effect size |
|
|
Term
| probability that the true score falls outside the interval |
|
Definition
|
|
Term
| 2 ways to be precise and accurate: |
|
Definition
1. reduce variability
2. increase sample size |
|
|
Term
| a confidence interval of 99% will contain the population parameter __% of the time: |
|
Definition
|
|
Term
1. extensively used in polling data
2. frequently used in epidemiology
3. can be used to test hypothesis IF the pop mean is known |
|
Definition
|
|
