Develop a framework for residential construction cost estimating in the Australian market

Dixon, Andrew Peter (2021) Develop a framework for residential construction cost estimating in the Australian market. [USQ Project]

Text (Project)
DIXON Peter dissertation_redacted.pdf

Download (15MB) | Preview


Time is a valuable resource in the construction industry and it is critical to the financial success of a project that accurate cost estimates are produced. Several methods of estimating currently exist with varying degrees of accuracy and completion time. First principle estimating is the most time consuming, often taking hours to complete, however is the most accurate. The unit rate method is quick to apply although it suffers from inaccuracy. A need exists for an accurate method of cost estimating that can be quickly applied. This study solves the problem by developing a framework for cost estimating based on the residential construction sector in the Australian market, which has not been done previously.

The approach taken in this study is based on cost modelling, which is a method of statistically predicting construction costs using input variables known as cost drivers. Cost drivers are factors of statistical significance that affect the total cost of a construction project. The literature review found that previous cost modelling studies focused on a broad range of cost drivers which yield a model that is not commercially viable and inaccurate. Therefore, this study has focussed on design related cost drivers only. This will improve accuracy and the commercial viability of the framework. Previous studies used cost data from publicly available or historical sources. This data includes contractor mark-up strategies, risk contingencies, variance in construction methodology and fluctuations in unit costs between localities which skew results. This study will utilise an up-to-date cost estimating database available in the construction industry for uniform data collection. It will also focus on construction cost only rather than final project cost, this removes the influence of mark-up and contingency factors. These steps will ensure the relevance of the developed framework.

A case study using semi-structured interviews was conducted on a cost estimating company in the residential sector of the Australian construction industry. The purpose was to confirm that the first principle estimating method is currently used, it is time consuming and the most accurate method available. It also examined the validity of cost drivers found in the literature review and expanded the design related cost drivers used for the statistical analysis. In addition, the case study findings were used to calculate a first principle estimate on 170 house designs. This method was used to create the cost data samples for the statistical analysis.

A statistical analysis was conducted on the sample data using SPSS which resulted in a model that predicts the construction cost for a project. Linear regression analysis and two neural network models were tested. Models from previous studies range in accuracy from 3.98% to 19.60%, this level of accuracy is not deemed commercially viable. With a focus on design related cost drivers this study found linear regression analysis performed best and improved the accuracy of previous studies to a mean absolute percentage error (MAPE) of 1.70%. The linear regression statistical model was used to develop the framework.

The discoveries of this study benefit cost estimating professionals by offering an estimating method that is accurate, which can be applied faster than traditional first principle methods. The framework can be operated by users with little training compared to fully qualified estimators completing first principle estimates. Further development of this technique, which involves design related cost drivers only, can be applied to other sectors on the construction industry. This has the potential to lower resources for companies tendering for the procurement of work by offering an accurate method that reduces the time and skill to apply.

Statistics for USQ ePrint 51809
Statistics for this ePrint Item
Item Type: USQ Project
Item Status: Live Archive
Faculty/School / Institute/Centre: Historic - Faculty of Health, Engineering and Sciences - School of Civil Engineering and Surveying (1 Jul 2013 - 31 Dec 2021)
Supervisors: Heravi, Amirhossein
Qualification: Bachelor of Construction (Honours) (Construction Management)
Date Deposited: 03 Jan 2023 01:11
Last Modified: 26 Jun 2023 01:10
Uncontrolled Keywords: residential, construction, cost, estimation, Australia, cost modelling, cost drivers, SPSS

Actions (login required)

View Item Archive Repository Staff Only