1. Advantages of Multivariate Analysis Close resemblance to how the researcher thinks. Close resemblance to how the researcher thinks. Easy visualisation and interpretation of data. Easy visualisation and interpretation of data. More information is analysed simultaneously, giving greater power. More information is analysed simultaneously, giving greater power. Relationship between variables is understood better. Relationship between variables is understood better. Focus shifts from individual factors taken singly to relationship among variables. Focus shifts from individual factors taken singly to relationship among variables.

2. Definitions - I Independent (or Explanatory or Predictor) variable always on the X axis. Independent (or Explanatory or Predictor) variable always on the X axis. Dependent (or Outcome or Response) variable always on the Y axis. Dependent (or Outcome or Response) variable always on the Y axis. In OBSERVATIONAL studies researcher observes the effects of explanatory variables on outcome. In OBSERVATIONAL studies researcher observes the effects of explanatory variables on outcome. In INTERVENTION studies researcher manipulates explanatory variable (e.g. dose of drug) to influence outcome In INTERVENTION studies researcher manipulates explanatory variable (e.g. dose of drug) to influence outcome

3. Definitions - II Scatter plot helps to visualise the relationship between two variables. Scatter plot helps to visualise the relationship between two variables. The figure shows a scatter plot with a regression line. For a given value of X there is a spread of Y values. The regression line represents the mean values of Y. The figure shows a scatter plot with a regression line. For a given value of X there is a spread of Y values. The regression line represents the mean values of Y.

4. Definitions - III INTERCEPT is the value of Y for X = 0. It denotes the point where the regression line meets the Y axis INTERCEPT is the value of Y for X = 0. It denotes the point where the regression line meets the Y axis SLOPE is a measure of the change in the value of Y for a unit change in X. SLOPE is a measure of the change in the value of Y for a unit change in X. Intercept Slope X axis Y axis

5. Basic Assumptions Y increases or decreases linearly with increase or decrease in X. Y increases or decreases linearly with increase or decrease in X. For any given value of X the values of Y are distributed Normally. For any given value of X the values of Y are distributed Normally. Variance of Y at any given value of X is the same for all value of X. Variance of Y at any given value of X is the same for all value of X. The deviations in any one value of Y has no effect on other values of Y for any given X The deviations in any one value of Y has no effect on other values of Y for any given X

6. The Residuals The difference between the observed value of Y and the value on the regression line (Fitted value) is the residual. The difference between the observed value of Y and the value on the regression line (Fitted value) is the residual. The statistical programme minimizes the sum of the squares of the residuals. In a Good Fit the data points are all crowded around the regression line. The statistical programme minimizes the sum of the squares of the residuals. In a Good Fit the data points are all crowded around the regression line. Residual

7. Analysis of Variance - I The variation of Y values around the regression line is a measure of how X and Y relate to each other. The variation of Y values around the regression line is a measure of how X and Y relate to each other. Method of quantifying the variation is by Analysis of variance presented as Analysis of Variance table Method of quantifying the variation is by Analysis of variance presented as Analysis of Variance table Total sum of squares represents total variation of Y values around their mean - S yy Total sum of squares represents total variation of Y values around their mean - S yy

8. Analysis of Variance - II Total Sum of Squares ( S yy ) is made up of two parts: (i). Explained by the regression (ii). Residual Sum of Squares Sum of Squares ÷ its degree of freedom = Mean Sum of Squares (MSS) The ratio MSS due to regression ÷ MSS Residual = F ratio

9. Reading the output Regression Equation Regression Equation Residual Sum of Squares (RSS) Residual Sum of Squares (RSS) Values of α and β. Values of α and β. R 2 R 2 S (standard deviation) S (standard deviation) Testing for β ≠ 0 Testing for β ≠ 0 Analysis of Variance Table Analysis of Variance Table F test F test Outliers Outliers Remote from the rest Remote from the rest