I have FDI data series of more than 40 years. But the series contains negative values of 5 different years. Since, I am using FDI as independent variable and necessary to transform those values in log form. What are the methods that I can transform negative data set in to log form?
Hello Nabaraj,
Assuming that transforming or re-expressing your scores makes sense in the context of your analysis:
If the zero point of your measure is arbitrary (as is the case in temperature quantified as degrees Fahrenheit), then you could first add a constant (of value MIN + 1 or larger, where MIN is the lowest observed score in the data set), then apply the log transform. Similar examples would include scales such as z-scores, where negative numbers simply indicate scores below the mean, or Likert-type response scales, where the so-called neutral category was scored as zero, and scores to the left were negative, while scores to the right were positive.
Good luck with your work.
If values can be negative, log-transformation makes no sense. Please make yourself clear about the definition of the variable (FDI?) and what a log-transformation means.
If the negative values are the result of some subtraction, you should either consider using ratios (rather than differences) or use a method that does not allow producing negative differences (try to de-convolute additive noise from a strictly positive signal).
Hello Nabaraj,
Assuming that transforming or re-expressing your scores makes sense in the context of your analysis:
If the zero point of your measure is arbitrary (as is the case in temperature quantified as degrees Fahrenheit), then you could first add a constant (of value MIN + 1 or larger, where MIN is the lowest observed score in the data set), then apply the log transform. Similar examples would include scales such as z-scores, where negative numbers simply indicate scores below the mean, or Likert-type response scales, where the so-called neutral category was scored as zero, and scores to the left were negative, while scores to the right were positive.
Good luck with your work.
- Badges
- Science topic