What is the best statistical test and what parameter in that test, is used to remove the effect of multiple factors on 1 variable and maintain just 1 factor?
Regression. The regression coefficient (b) for a given parameter is an estimate of what effect can be uniquely attributed to that factor when effects of other factors are also being considered.
Thank you Prof. Stephen
Do you mean that only b value is sufficient to conclude if the variable is significantly affecting the dependent variable? what is the mean of other parameters in regression analysis? what is the importance of P value here?
No, the coefficient alone is not enough to make any conclusion about statistical significance. A regression coefficient is an effect size; effect size and statistical significance are different things. In your original post you did not ask about significance testing, you just said you wanted to know what the parameter's effect was; to know what the parameter's effect is, the coefficient (b) is what you look at. To know whether that effect is statistically significant, you need to look at the p-value or confidence interval associated with that coefficient.
By way of comparison: if you were interested in the weight difference between dogs and cats, you could look at the raw difference (the mean of the dogs' weights, minus the mean of the cats' weights) to know how big the difference is. That's like looking at a regression coefficient. To know if the weight difference is statistically significant, you need to look at the p-value associated with that difference; that's like looking at a regression coefficient's p-value.
The coefficient and statistical significance are two different things.as said by prof.Stephen. normally there are various parameters on which we have to conclude from a statiscal analysis. To commonalise any thing we need to look at the p vlaue., to know whether it is staticticall
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics