I have undertaken a meta-regression as part of a meta-analysis.
However, the output for one of the variables (sample size) the regression coefficient and standard error include an 'e' and I am not sure what this means.
The coefficient is: -4.90e-07
The standard error is: 1.70e-06.
I don't know if this is to do with the distribution of the variable? There is a wide variation in the sample sizes (18-71,303) and the sktest was significant. Though no ladder transformation improved the distribution. I truncated the variable (as one study with the n of 71,303 was far larger than the other studies). Though the sktest is still significant.
Overall the variable appears to explain less heterogeneity than would be expected by chance (adjusted r-square = -9.02%). So I don't plan to include it in my model with multiple covariates.
However, if anyone had any advice on how to interpret an e in the coefficient that would be much appreciated.
However, the output for one of the variables (sample size) the regression coefficient and standard error include an 'e' and I am not sure what this means.
The coefficient is: -4.90e-07
The standard error is: 1.70e-06.
I don't know if this is to do with the distribution of the variable? There is a wide variation in the sample sizes (18-71,303) and the sktest was significant. Though no ladder transformation improved the distribution. I truncated the variable (as one study with the n of 71,303 was far larger than the other studies). Though the sktest is still significant.
Overall the variable appears to explain less heterogeneity than would be expected by chance (adjusted r-square = -9.02%). So I don't plan to include it in my model with multiple covariates.
However, if anyone had any advice on how to interpret an e in the coefficient that would be much appreciated.
Comment