An Evaluation of the Predictive Validity and Time Saving Potential of a Roof Covering Quantification Model

Quantification of roof covering often poses a challenge for quantity surveyors/construction estimators and builders due to their time consuming nature. The study was aimed at assessing the predictive efficiency and time saving potential of a regression model in generating the approximate quantities of installed roof covering using the roof covering area on plan as the independent variable. Thirty architectural drawings (roof plans and sections) comprising eleven bungalows, sixteen duplexes and three storey buildings were randomly selected and the quantities of the roof area on plan were measured. The time spent by a group of proficient quantity surveying graduates to apply the model in comparison with the traditional take-off method was also recorded. Pearson Correlation and Student t-test were employed to test the hypotheses. Results reveal that the mean difference between the modeled quantities and installed roof covering quantities are 320.7000 and 322.4667 respectively. The results also show that the p-value of 0.009 is less than the alpha value of 0.05, indicating rejection of the null hypothesis, implying that the model saves time for quantification. The Coefficient of Correlation (R) of 0.990 suggests a very strong relationship between the area of roof covering on plan and the area of installed roof covering, implying that the model accurately predicts approximate quantities of the installed roof covering. The study concludes that the model is a realistic alternative for generating approximate quantities of installed roof covering as long as the pitch criteria of 250 –390 are met. It was advanced that quantity surveyors/construction estimators and builders use the model for estimation during the planning and design stages of projects, especially when time is of the essence.