I was trying to construct a model to estimate the benefits of an urban redevelopment project using regional and case study specific variables, however I have run into problems with my data collection and there are huge gaps in my data and the data I do have might be questionable, as such I am switching my focus away from empirically estimating and critical analysing the reasons why the urban redevelopment project will be so beneficial for Kathmandu. I want to still write a Methodology section explaining why the focus of my paper changed, is this professional?
Nicholas, I am sorry that you had this problem in your methodology. In all econometric modelling we have this problem, the lack of consistent and believable data. Does this mean that econometrics failed, i do not think so. If you change your empasis or the process it is important to give a run down of why the change took place. It could be in the interest of the client, whoever that is, to take note of the problems you encountered. It might, therefore, lead to suggestions for a better and more comprehensive data collection strategy. I, thus, think that it is very professional to make people aware of the reasons why a specific process failed.
Could I suggest that you persue your new qualitative study using multi-criteria analysis that will help the decission makers the oportunity to input their measures of importance for each of the factors you consider.
Hi Nicholas:
Adequacy of data has been concern in developing countries; however, it should not restrict using quantitative techniques. There are several ways to collect missing data including expert opinion. One can develop different scenarios.
I think shifting from quantitative to qualitative techniques is not a welcome step; later being more subjective.
I will be happy to help you.
I assume that you have rethought your original model and concluded that no econometric analysis will suffice.
In that event, I think you should explain why your model failed, propose suggestions in the data and/or theory that would make it work, then go ahead and present your qualitative analysis. At worsr, some journal referee will tell you to remove the econometric stuff. But in general, it's better to show that you have thought about an issue than to let your reader wonder if you didn't know about it.
It is better to explain the possible reasons why your econometric model is failed. Is it the problem with quality of data or econometric modeling techniques are not sufficient to serve the purpose. In later case, it will guide the researcher particularly econometricians to improve their models.
Dear Nicholas,
I can relate to what happened to your research and at the same time, I am sure that you can't get any effective answer until you share more details. I am offering my assistance to you in case you need. I have been using econometrics with some colleagues since last 10 years and at the same time we are also in qualitative research. If possible, please share more information so that we can find a solution in a collaborative manner. My email is: [email protected]
All of you have provided excellent answers and have came to the conclusion which I basically knew already, I'm just quite new to economic research off my own back and needed some expert guidance. The main problem is the unreliability of data and in a lot of cases a lack of data. This is also coupled with time constraints as I finish my internship in just less than a month so I unfortunately do not have time to conduct further in depth research to find the missing data. However, I will discuss the reasons why I was unable to construct my model and explain in qualitative terms; at least through trying to construct my model I have researched nearly 50 redevelopment projects globally and now have a plethora of knowledge on the subject, with this in mind I think it might actually give a better overview of the situation to explain it in qualitative terms. I can pick up the model later if it is convenient.
Thank you for all your answers and your offers of help, when my paper is finished I will post it here for anyone's reading pleasure (hopefully pleasure and not disgust).
It is unusual to face such problems in quantitative analysis. Probably you should have another look at objectives of the study ,data requirements and your questionnaire. This may require some sort of pretesting exercise to identify gaps in your data and reformulate the questionnaire. Even the you should be careful in formulating model that ought to approximate your theoretical construct. So please do not give up so soon. Nevertheless qualitative analysis sometimes more useful as it helps to identify logic of the situation for rational action.This might help you to identify a better model. Would be interest to know how you progress.
Welcome to research in economics, lol!! Your encounter is common. Before you throw arms in the air and with them many hours of work, take a look at the articles by Kevin D. Hoover andStephen J. Perez "three attitudes towards data mining" in journal of economic methodology, and Ad Feelders "data mining in economic science". Both articles are available online in PDF. They should give you ideas on how to refocus your paper while acknowledge it's limitations.
Research requires a lot of creativity in both quantitative and qualitative methodologies....the are just different levels of data....bottom line is that research is an art...the insight of inquiry ends up being a fairly personal pursuit....the outcomes may be replicable....but you are the method...
If you are writing for a journal, go straight into your methodology that results in the finding that your are reporting. A sentence on why you did not use econometric approach should suffice. It will be a waste of space trying to explain details of the econometric failure for the purpose of your study, but if you are reporting in a degree thesis, you have ample opportunity to show what you tried, and what failed, why you failed etc. It is a way to show depth in the mastery of methodologies. In short, it depends on the purpose of your reporting.
I agree with Amavilah and Adejumo!
But: a good qualitative analysis is extremely worthwhile as well. Nobel prize winner Ronald Coase has restated in an interview that what economists should do is to explain "how things actually work" and he was weary of econometrics. You can find this on
http://www.econtalk.org/archives/2012/05/coase_on_extern.html
If the datagaps are not too large, you might always try methods such as Amelia http://gking.harvard.edu/gking/amelia.
Very nice to know that you have found constructive suggestions of fellow researchers meaningful. i am sure you would like to come back to it whenever you get more time..
Good luck1
Great find, Gerdien! Now let's send Nic Walton our consulting bill, lol. Love it! In some ways econometrics has become the tail that wags the dog -- a place to play hide-n-seek when we don't understand 'how things actually work'. What Coase is saying is that if research is advancing understanding of some aspect of life/science, it shouldn't be abandoned just because it is difficult or if impossible to measure.
Dear all, I recall that at some time in the past there was an article published with the title: "How to take the con out of econometrics." I can not remember when or in which journal or who the author is but this always remain my concern when working with economic data. We can at best explian 30 to 40% a the actual events and then we feel we can use that to explain the intricate interactions in the economy. Not very good science, do you agree.
@Robert Bradfield, the article you refer to is "Let's take the con out of econometrics." It is by Edward E. Leamer of UCLA, and was published in the American Economic Review, Vol.73, pp. 31-43, 1983. Not only is 60-70% error too large to minimize successfully, denying its existence is just unscientific.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series