• Join over 1.2 million students every month
  • Accelerate your learning by 29%
  • Unlimited access from just £6.99 per month

In order to determine the effectiveness of auditing programs in promoting energy savings, the NSW Department of Energy has carried out a project in which they ran a regression model on a small number of households.

Extracts from this document...


1. Introduction

In order to determine the effectiveness of auditing programs in promoting energy savings, the NSW Department of Energy has carried out a project in which they ran a regression model on a small number of households. The households were investigated based on two forms of data: billing data consisting of annual household consumption of electricity and survey data including household demographic characteristics. Based on the econometrics, the Department concluded of a positive effectiveness of the audit program.

In this paper, we are going to review the project in the following parts: Initial data analysis, Critical review of the Department’s report, Proposed econometric approach, Initial result & model evaluation and Final results and main conclusions.

2. Initial data analysis

a. Replication

Using Shazam, we successfully replicated the results presented in the Department of Energy’s report. See appendix 1.1 and 1.2 for details of the regressions. The only difference related to scale is that the coefficient for income is 1000 times smaller than that is reported by the Department. The reason comes from the use of the scale where the Department uses 1 dollar to represent as unit while the data displays thousand dollar unit.

b. Significance of the intercept

Interestingly, the result of the regression of model (3) suggests an insignificance of the intercept, i.e. the intercept is not significantly different from 0. To confirm that, the following test is conducted.

Test hypothesis:

H0: β1=0

HA: β1≠0

LOS: α=0.05

Test statistic:  t = -0.1668 < 2    (t stat is from Appendix 1.2)

Conclusion: Do not reject H0 i.e. we conclude that the intercept is not significantly different from 0 at 0.05 LOS.

The above outcome encourages us to run a regression of no intercept. The detailed results are shown in Appendix 1.3 that derives the following model:

             ELEXP = 69.282 ROOMNO  -99.47DAUD +  u

                   (4.232)      (40.88)

This model seems to produce better results than the previous one (3).

...read more.


Data and their features should be presented in an appendix.

d. The regression model

The advantage of conciseness actually has a counter effect on this part. It is too much concise and in severe lack of technical language and details. The report says very little about building the model and underlying assumptions. In this very important part of the project where is a fundament for every later conclusion, any related reason should be fully presented and any necessary technique should be applied. Particularly, when econometrics has been chosen to be the media of analyzing the data, the work must have been done in logic order under requirements of a model regression building in econometrics.

The paper does not give any reason for choosing the model or omitting variables. No explanation and no evidence are stated to prove that the model is not misspecified. Depending on data and purpose of report, etc, a model can be linear, quadratic or log forms. There should be a discussion about this or at least a review of the similar paper to know what variables should be included in the model.  Indeed, there are a variety of methods to improve the model, e.g. variables can be transformed to log, interaction variables can be added, etc. However, none of them is presented including some basic diagnostic tests such as tests of heteroscedasticity, autocorrelation, etc to give the model statistical credibility.

Interpretation is inadequate and not justified formally by appropriate statistic tests or economic theories. E.g. in the interpretation of “significance” of the model (3), coefficients of ROOMNO (5.4) and DAUD (-2.2) without any explanation that can mathematically results to the conclusion of a better model. In addition, the conclusions of R2 and ROOMNO are not supported by any statistics.

...read more.


There are no improvements in heteroscedastic problems. P-B-G’s p-values of both transformed model 1 and 2 are still 0.00. This shows that the action of dividing both sides of  the models by square root of income does not help to improve the heteroscedastic problems. This suggests the heteroscesdasticity may be resulted from some factor other than incomes. Several further tests and attempts need to be done to solve the problems.

c. Conclusion about impacts of the audit programs:

Due to heteroscedasticity, we can not give any concrete conclusions about the effectiveness of the audit programs. Their effects on electricity expenditure may be varied and inaccurate. However, the large negative values of coefficients of DAUD in all the models suggest that the audit programs have positive effects on energy savings, i.e. audit programs significantly reduce the electricity expenditure.

d. Executive summary:

The Department of Energy, through their small-scaled research, has considered audit programs as one of the possible demand-side management scheme that can improve energy savings.

A full-scaled project has been conducted to measure the effectiveness of such scheme. Data used are expanded to all 296 households whose electricity consumption is monitored.

The purposes of this paper are (1) examine the previous report and (2) analyse the full data in order to conclude about effectiveness of audit programs in promoting energy savings.

Econometric evidence shows that (1) although the previous report is concise and accurate, it is still incomplete and in lack of in dept analysis, (2) a few econometric problems may result in changes in measure of effects of audit programs and (3) the results suggest that the audit programs have strong positive effects on energy savings.

This report, therefore, recommends the audit programs should be implemented; however, some further researches may be required to clearly conclude about the effects of such schemes.

[1]٭  SST = ∑(yi – ў)2 < ∑yi2

SSE > SSE1- SSE < 1- SSE R2 < r2

   SST    ∑yi2          SST        ∑yi2

[2]All the tests are fully showed in the Appendix 3.a-d  

...read more.

This student written piece of work is one of many that can be found in our University Degree Computer Science section.

Found what you're looking for?

  • Start learning 29% faster today
  • 150,000+ documents available
  • Just £6.99 a month

Not the one? Search for your essay title...
  • Join over 1.2 million students every month
  • Accelerate your learning by 29%
  • Unlimited access from just £6.99 per month

See related essaysSee related essays

Related University Degree Computer Science essays

  1. Phong Shading and Gouraud Shading The standard reflection model in computer graphics that ...

    The final polygon colour is the sum of all four light components, each of which is formed by multiplying the glMaterial colour by the glLight colour (modified by the directionality in the case of Diffuse and Specular). Since there is no Emission colour for the glLight, that is added to the final colour without modification.

  2. The project explains various algorithms that are exercised to recognize the characters present on ...

    == 8 ) + ((a & 0x0004) == 4) + ((a & 0x0002) == 2 ) + ((a & 0x0001)== 1)+ result_match ; } match_percentage = ((64 - result_match)*100)/64 ; if (highest_percentage < match_percentage) { highest_percentage = match_percentage; mapping_quadrant[x] = tt; mapping_percentile[x] = highest_percentage; } else { } tt= tt + 1; } if (mapping_percentile[x] < 75 )

  1. Geometric Brownian Motion. The aim of this project is to gain an understanding ...

    2.3 Stochastic process A stochastic process is based on successive draws of a random variable from a defined probability distribution. The values of the random variable cannot be known for certain at the start of the process, the value can only be characterised using the tools of probability such as

  2. An Introduction to the IEEE 802.11p WAVE standard

    With respect to V2V communications, forming an ad-hoc network between two vehicles requires very specific routing mechanisms. To achieve what is set out by the standard, both single and multi-hop routing will be supported. It will do this by utilising a packet forwarding scheme that routes packets through nodes closest to the destination.

  1. Lifecycle Management Of Information Technology Project In Construction

    related to project re?t? with principal inve?tigator?. Re?earch will be conducted in ?uch the way that integrity of re?earch enterpri?e will be maintained? and negative after-effect? which might dimini?h potential for future re?earch were avoided. The choice of re?earch i??ue? will be ba?ed on be?t ?cientific judgment and on an a??e??ment of potential benefit to participant?

  2. Entering Compiling and Running C Programs ...

    { printf("%d is dividable by 5. \n" ,N); } else { printf("%d is NOT dividable by 5. \n",N); } getch(); } Output of the above program Explain the purpose of the symbol '%' in the program? It functions like '/', the symbol divides the numbers,when the reminder is 0 the output will be "Dividable" otherwise" NOT dividable".

  1. So in order to understand what the main areas where organisation should be aware ...

    Can you remind me what it is?" Footprinting contains researching the target website's specific qualities such as open ports, services, security feature, and any other information you can get out of the machine on the network. Internet Footprinting: "Systematic collection of information on an intended target with the goal to

  2. At the end of our audit of service in &amp;quot;U Szwejka&amp;quot; Restaurant we would ...

    the background o Songs in English - Czech songs might be more suitable o A Czech language learning tape was played in the toilet! A fun element! o Lighting pleasant, but good enough to see what was on the table o The table and bench were fixed in place.

  • Over 160,000 pieces
    of student written work
  • Annotated by
    experienced teachers
  • Ideas and feedback to
    improve your own work