Term
|
Definition
| Analysis of variance. Tests the null hypothesis that the population means are all equal by comparing the variability between groups to the variability within groups. |
|
|
Term
|
Definition
| A method of inference for two means. |
|
|
Term
| Coefficient of determination (R2) |
|
Definition
Sample variance of Y hat / Samle variance of Y
The square of the correlation coefficient. The proportion of variance in Y that can be explained by the linear relationship with X. The closer to 1, the stronger the relationship. |
|
|
Term
| Correlation coefficient (r) |
|
Definition
-1 < r < +1
A unitless measure of the strength of the linear relationship between X and Y. The father from zero, the stronger the relationship. |
|
|
Term
|
Definition
| Multiple measurements made on the same individual, or linked individuals such as identical twins. Reduces variability. Easier to isolate the effect of interest. |
|
|
Term
|
Definition
| The counts we would get on average if the null hypothesis was true. For a test for independence, it is (row total x column total) / overall total. |
|
|
Term
|
Definition
| Independent variable. Used to predict Y. |
|
|
Term
|
Definition
MST / MSE
Predicts the p-value in an ANOVA test. |
|
|
Term
|
Definition
| A chi-square test for a one-way table. Observations are classified according to one categorical variable. |
|
|
Term
|
Definition
|
|
Term
|
Definition
| Use pooled-variance procedure or Welch procedure. |
|
|
Term
|
Definition
| Testing if there is a significant difference between the means of two groups, and finding a confidence interval. The difference of the sample means is an unbiased estimator of the difference in population mean. If both have normal distribution, the difference will have normal distribution. Options include pooled-variance t procedure, Welch procedure, Mann-Witney U, bootstrap methods, and permutaiton tests. |
|
|
Term
|
Definition
| A metho of inference for two means. |
|
|
Term
|
Definition
| Individuals that are grouped into pairs according to variables that are likely to affect the response. Treated as single samples. |
|
|
Term
|
Definition
| β0 hat and β1 hat are chosen so that the sum of squred residuals is minimized. |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
| Multiple linear regression |
|
Definition
| There is more than one explanatory variable. |
|
|
Term
|
Definition
aka Distribution-free procedure
aka Sign test
aka Sign-ranked test
Used if distribution is not normal. |
|
|
Term
|
Definition
| The observed sample counts in a category. |
|
|
Term
| Paired difference t procedure |
|
Definition
| If the null hypothesis is true, t will have a distribution of n - 1 degrees of freedom. |
|
|
Term
|
Definition
| A method of inference for two means. |
|
|
Term
| Pooled sample proportion (p bar) |
|
Definition
|
|
Term
| Pooled-variance procedure |
|
Definition
| A method of inference for two means. Assumes equal population variances, and results in an exact method. Consistent with other common statistical procedures such as ANOVA and linear regression. Assumptions are likely to be untrue. Performs poorly if assumptions are violated, and even worse when there is a difference in sample sizes. Large sample sizes doesn't erase this problem. |
|
|
Term
|
Definition
| Either predicting a single value of Y at a given X (confidence interval is a constant width), or estimating the theoretical mean of Y at a given X (confidence interval is narrower than the mean of X). |
|
|
Term
|
Definition
| The mount of individuals in a population showing a certain measured characteristic. |
|
|
Term
|
Definition
| Explores possible relationships between Y and X. |
|
|
Term
|
Definition
| The observed value minus the predicted value. |
|
|
Term
|
Definition
| A scatter plot of residuals. |
|
|
Term
|
Definition
aka Dependent variable
The variable we want to predict. |
|
|
Term
|
Definition
| There is one explanatory variable. |
|
|
Term
|
Definition
|
|
Term
|
Definition
| The sum of squares within groups. |
|
|
Term
|
Definition
| The sum of squares between groups. |
|
|
Term
|
Definition
| A chi-square test for a two-way table. |
|
|
Term
|
Definition
aka Unpooled-variance procedure
A method of inference for two means. Does not assume equal proportion variances, and results in an approximate method. Usable in a wider variety of situations. Does not require assumptions. Works better than pooled-varaince but is less exact. |
|
|
Term
| Welch-Satterthwaite approximation |
|
Definition
| A complex equation that gives the degrees of freedom in a Welch procedure. |
|
|
