Econn-08/Discussion
Jambo
econn-08_-discussion.docxPost –ecomnn-08, followed by two responses.
In your final project, you will be making recommendations on the results of your study. Discuss recommendations and compare these to the recommendations discovered in your original literature review. (Make sure your topic is analysis off demand electricity)
Note; And review and make a corrections milestone –two from the feed back
Instruction for making recommendations.
We continue this week to look at how econometrics is used in the real world. Over the past few weeks we looked at some thorny problems in running regressions. We will continue with some of these issues next week as well. If nothing else, these Modules should provide you with a sense that econometric analysis can be quite complex. There is a temptation in business sometimes to think that all an employee needs to do statistical analysis is to be given a canned package and simply input data.
This week we discuss what a proper model specification looks like. As with many econometric applications there is no simple answer. The quick response is that it is the model that works best. However, what is “best” is a managerial judgement. The manager decides what the model is to be used for and how to evaluate it. A good forecasting model would have a high R square and stable coefficients over time. However such a model might not provide a good estimate of a price elasticity of demand.
There is a poster that says the following:
CHOOSE TWO OUT OF THREE:
FAST
ACCURATE
INEXPENSIVE
The idea, applied to modeling, is that there are tradeoffs. If the model is to be estimated quickly because of a time deadline then it will probably be inaccurate or expensive. An accurate model that is inexpensive to build will not be created quickly.
As many of you have discovered by now each step of the process requires a judgement call. To take one example, data that is easily and freely available tend to be secondary data designed to answer a specific question unrelated to the research. Getting more appropriate data usually requires some monetary expenditure and time. In a fast-paced business world sometimes the data that is readily available might be good enough to use. At the end of the day the model has to be looked at in its totality of technical characteristics, costs, and management’s expectations of it.
and make a corrections milestone –two from the feed back
