N2025 P2 Q8

N2025 P2 Q8

11 marks
Free
  1. Dilip is investigating how athletes perform in related events. He collects data about the best performances of \(10\) randomly chosen athletes in high jump and long jump. The results, in metres, are shown in the scatter diagram. The correlation coefficient is \(0.39\).

    What do the scatter diagram and correlation coefficient suggest about the relationship between best performance in high jump and long jump amongst athletes taking part in both events? Justify your answer.[2]
  2. \(x\)2.13.96.27.81012.513.815
    \(t\)11265375119165246300
    1. Sketch a scatter diagram of the data. State what the scatter diagram suggests about the relationship between \(x\) and \(t\).[2]
      1. Determine which of the following models is the better fit to the data where \(a\), \(b\), \(c\) and \(d\) are constants.

        \(t=ax+b \qquad t=cx^2+d\)

        [2]
      2. Find the regression equation for the model identified in part (ii)(A).

        [2]
      1. Use the regression equation found in part (ii)(B) to estimate the time Dilip would take for a run of \(11\) km.

        [1]
      2. Give two reasons why your estimate is reliable.[2]

Default solution

  1. The points show a weak positive linear association: athletes with greater high-jump performances tend to have greater long-jump performances, but with substantial scatter. The correlation coefficient 0.39 is positive and far from 1. This does not establish that performance in one event causes performance in the other.
    1. For this part, the original paper defines x as the distance run in kilometres and t as the time taken in minutes.
      The time increases as distance increases, at an increasing rate; the relationship is positive and curved rather than approximately straight.
      1. Compare correlation after each proposed transformation. Let u=x² for the quadratic model.
        \(r_{t,x}=0.957516\ldots,\qquad r_{t,u}=0.990466\ldots\)
        The correlation for t against x² is closer to 1, agreeing with the upward-curving scatter. The quadratic model t=cx²+d is the better fit.
      2. Regress the response t on the predictor u=x². Retain unrounded coefficients for the later prediction.
        \(t=1.2496914048\ldots u+0.8758090330\ldots\)
        \(t\approx1.25x^2+0.876\)
      1. \(t\)\({}=1.2496914048\ldots(11)^2+0.8758090330\ldots\)\(t\)\({}=152.088469\ldots\approx152\ \text{minutes}\)
      2. The distance 11 km is within the observed range 2.1 to 15 km, so this is interpolation. The transformed correlation is close to 1 and the points lie close to the fitted quadratic trend, supporting the estimate under similar running conditions.
Answer:(a) Weak positive linear association, with substantial scatter. (b)(i) Positive curved relationship, increasing at an increasing rate (scatter shown). (ii)(A) Quadratic model. (B) \(t\approx1.25x^2+0.876\). (iii)(A) Approximately 152 minutes. (B) Interpolation and strong transformed linear correlation.

Need help? Join our JC Math tuition classes.

Learn more