130 likes | 268 Views
Predicting Heights. Jimmy L. Colleen M. Brandon D. What we measured. Middle Finger We measured from the from the knuckle to the tip of the middle finger Recorded that measurement to the nearest 8 th inch Possible bias: nail length. Forearm Length
E N D
Predicting Heights Jimmy L. Colleen M. Brandon D.
What we measured • Middle Finger • We measured from the from the knuckle to the tip of the middle finger • Recorded that measurement to the nearest 8th inch • Possible bias: nail length • Forearm Length • We measured from the elbow to the very end of the forearm or the start of the wrist • Recorded that measurement to the nearest 8th inch • Possible bias: misjudging where the forearm ends • Wrist Circumference • We measured the circumference of the wrist from the point just below where your thumb joins your hand • We recorded that measurement to the nearest 8th inch • Possible bias: measuring in the wrong place, bracelets that may have gotten in the way
Middle Finger • Scatter Plot: Positive, linear, weak • R2 -14% of the change in height can determined by the change in middle finger length
Wrist Scatter Plot: Positive, moderately weak, linear R2 - 27% of the change in height can be determined from the change in wrist circumference
Forearm Scatter Plot: Positive, Moderate, linear R2 - 47% of the change of height can be determined by the change is forearm length
Middle Finger: Gender • Females: have a higher y-intercept but a smaller slope • R = .44 • Weak Scattered • Males: have a lower y-intercept but a higher slope • R = .10 • Weak Scattered
Wrist: Gender • Females: Higher y-intercept but lower slope • R = .31 • Weak Linear • Male: Lower y-intercept but higher slope • R = .24 • Weak Scattered
Forearm: Gender • Females: have a lower y-intercept but a higher slope. • R = .16 • Weak linear • Males: have a higher y-intercept but a lower slope. • R = .54 • Weak Linear
And the winner is… • Forearm length was chosen as our best model because… • The residual plot is scattered • With an r = .687, it has the highest correlation • 47% of the change in height can be determined by the change in the length of the forearm
Predicting our heights • Jimmy • Forearm length = 10.5 • 3.62(10.5)+32.1=Height • Height = 70.11 inches • Actual Height = 69.5 inches • Residual = -.61inches • Colleen • Forearm length = 9.5 inches • 3.62(9.5)+32.1=Height • Height = 66.49 inches • Actual Height = 64 inches • Residual = -2.49 inches • Brandon • Forearm length = 10.25 • 3.62(10.25)+32.1=Height • Height = 69.205 inches • Actual Height = 71 inches • Residual = 1.795 inches
Predicting the teachers • Ms. Wilson • Forearm length = 10 inches • 3.62(10)+32.1=Height • Predicted Height = 68.3 inches • Mrs. Robinson • Forearm length = 10 inches • 3.62(10)+32.1=Height • Predicted Height = 68.3 inches • Ms. Arden • Forearm length = 9 inches • 3.62(9)+32.1=Height • Predicted Height = 64.68 inches • Mr. Smith • Forearm length = 11 inches • 3.62(11)+32.1=Height • Predicted Height = 71.92 inches We are somewhat confident in our predictions because even though this model had the highest correlation out of all of the others, its strength is still only moderate
Bias/Error • Not measuring the same for everyone • Not knowing where to call the measurement from
Conclusion • Best body part to predict height would be the forearm (based on our data) • How you measure can change your data