Statistics Homework and Lecture Notes: ANOVA and Linear Regression

This flashcard set covers the statistical concepts of ANOVA and Simple Linear Regression as demonstrated in the lecture on lead concentrations, denture adhesive holding force, and healthcare expenditures.

Null Hypothesis ($H_0$) in Lead Concentration ANOVA

The assumption that the mean lead concentration in breast milk is equal across all five age groups: $\mu_1 = \mu_2 = \mu_3 = \mu_4 = \mu_5$.

Alternative Hypothesis ($H_a$) in Lead Concentration ANOVA

The assumption that at least one pair of the age group means for lead concentration is unequal.

Grand Mean ($\bar{x}..$)

The overall average of all observations across all groups, calculated as the total sum ($T..$) divided by the total number of observations ($N$); for the lead study, this was $0.828$.

Sum of Squares Treatment ($SS_{Treat}$)

A measure of the variation between group means; in the lead concentration study, it was calculated as $2.563$.

Sum of Squares Error ($SS_{Error}$)

A measure of the variation within groups, also known as residual sum of squares; in the lead concentration study, it was $7.413$.

F-statistic ($F_{stat}$)

The ratio of Mean Square Treatment ($MS_{Treat}$) to Mean Square Error ($MS_{Error}$), used to determine if group means are significantly different; for the lead study, $F_{stat} = 3.63$.

Lead Accumulation Conclusion

The finding that lead accumulates with age, suggested by the significantly higher lead concentrations in the oldest age group (≥35).

Denture Adhesive ANOVA Result

With $F_{stat} = 3.66$ and $F_{crit} = 3.10$, the null hypothesis was rejected, meaning at least one denture adhesive has a different mean holding force.

Simple Linear Regression Equation

Mathematical model for healthcare expenditure ($Y$) based on age ($X$), given as $\hat{y} = a + bx$, where for the studied data $\hat{y} = -535,279.6 + 8,723.5x$.

Slope ($b$) in Healthcare Regression

The value $8,723.5$, representing that for each year of aging, the mean cumulative healthcare expenditure increases by $\$8,723.5$.

$$SS_{xx}$$

The total sum of squares for the independent variable (age), calculated as $\sum x^2 - \frac{(\sum x)^2}{n}$, which equaled $1,050$ in the healthcare study.

$$S_{xy}$$

The sum of cross-products between age and expenditure, calculated as $\sum xy - \frac{(\sum x)(\sum y)}{n}$, which equaled $9,159,632.5$.

Regression Sum of Squares ($SS_{rear}$)

The portion of total variation in expenditure explained by age, calculated as $(b)(S_{xy})$, which was $79,901,783,482$.

Standard Error of the Slope ($s_b$)

The standard deviation of the slope estimate, calculated as $\sqrt{\frac{MS_{Error}}{SS_{xx}}}$, which was approximately $396.9$.

Prediction Mean for Age 77 ($\hat{\mu}_{y|77}$)

The estimated mean healthcare expenditure for a 77-year-old, calculated as $\$136,429.9$.

Standard Error of Prediction ($s_{\hat{y}_i}$)

The variability in the mean prediction for a specific value of $x$, which increases the further the specific $x$ is from the grand mean of $x$ ($\bar{x}$).

Regression Linear Relationship Conclusion

Because F_{stat} (483) > F_{crit} (5.99), there is significant evidence of a linear relationship between age at death and mean healthcare expenditure.