Making a prediction
A prediction can be made by reading from the graph.
A prediction can be made by substituting a suitable value into the trend equation, but with reference to the amount of scatter and the width of the band drawn around the data.
A prediction should be given as an interval. This interval would be determined by the width of the band.
Predictions should only be made for values within the data range.
To predict the percentage body fat of an AIS athlete with a Sum of Skin Folds of 120mm:
Substitute into the trend equation
%BFat = 0.18304 x SSF + 0.87336
%BFat = 0.18304 x 120 + 0.87336
%BFat = 22.8 %
The model predicts that an AIS athlete with an SSF of 120mm will have a percentage body fat of approximately 22.8%.
Looking at the scattergraph we can also say that the percentage body fat of this athlete is likely to be between 20.5% and 25.5%
Note: Round your answer sensibly in context!
Because the relationship is moderately strong (the points are close to the trend) my estimate of 22.8% is reasonably precise.
Use precise, reasonably precise or not very precise to discuss your prediction