Estimates, residuals & LSRL

• What is the least square's regression line?


  A type of line of best fit.
• How is the least square's regression line calculated?


  Since data used is for 2 variables the line is of the for y = a + bx
• Interpreting values

  • Gradient - as x increases by 1, y increases / decreases by b
  • Y-intercept - when x is 0, y = a.
• Exam technique for for least squares regression line 


  • Can be used to find estimates for other x values
  • May be asked to comment on validity of estimate (use interpolation and extrapolation)
• What is a residual


  The directed vertical distance from a point to a line (above = positive, below = negative)
• How are residuals calculated

  Residual =

  Observed

  Minus

  Expected
• Dealing with outliers


  • reference the size of the residual
  • compare residuals
  • suggest possible explanations for outliers