After collecting my data, I decide to test my hypothesis using regression analysis. I also recognized the fact that my data must meet the assumptions of the tool before I can use it. Therefore I will like to know if after testing two assumptions and realizing the data had met those assumptions, can I just run the analysis or I must test all the assumptions before?
It might be good to know what the assumptions are that have to be met by your software.
However, a simple regression can be run on more or less any type data. And regression analyseis will return a result. But it does not necessarily make sense to interpret the result if some initial errors were made.
The most important of course is: Are your data linear? That means, can your data be modelled by a simple model: y=mx+b. In other words a straight line, that depends on the measured value x. The slope of the straight line model is the m part, the b is the intercept.
In nearly every case it is advisable to plot the data prior to analysis. You can get a pretty good impression of that what can/should be done by visual inspection with the bare eye.
To get an idea about suitable use of linear models you may read about the Anscombe quartet, a data set constructed in 1973 to demonstrate how a violation of the linear assuption avoids a proper interpretation of the linear regression results:
https://en.wikipedia.org/wiki/Anscombe%27s_quartet
Jan
Whether regression analysis can be conducted depends on whether they are correlated, that is, whether the changes in a certain dimension of data change with changes in other dimensions of data, without conditional assumptions.
Enoch KABINAA Suglo -
Even a curved line may be modeled by a quadratic 'linear' regression. You would like your estimated residuals to be approximately 'normally' distributed. (Your data may be skewed, or otherwise nonnormal.) One often tests for heteroscedasticity, but you should look into weighted least squares (WLS) regression, as heteroscedasticity is to be expected with larger predicted values associated with larger expected variance of residuals. For time series, or even some spatial considerations in finite probability cases, you may have autocorrelation, and need to consider general least squares (GLS).
I suggest you first graph your data. A scatterplot can tell you a great deal. Then if you have further questions, you could show a graph along with your questions, and include the context.
Cheers - Jim
Can you be more specific please, and tell us something about your data and expectations?
Provide your data. In order to maintain confidentiality, you may not provide the event process, but only provide samples to help you analyze.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators