In interpreting indirect effect in serial multiple mediation model the bootstrap confidence interval is used.
When bootstrap confidence interval includes zero we cannot reject the null hypothesis e.g. -0.07 to 0.00.
On the other hand, when bootstrap confidence interval is greater than zero let say 0.01 to 0.12 we assumed that the mediation is significant (we can reject the null hypothesis).
My question is how to interpret when bootstrap confidence interval is less than zero e.g. -0.05 to -0.00.
It also confirms the mediation hypothesis, in general the intervals should not pass 0 but can be positive or negative.
Dear Cristian Ramos-Vera, thank you for your answer. If you much familiar with statistics can you guide us why there is minus value with both bounds of class interval in some cases?
Hello Sadiq,
In reply to your second query, the reason is that the "-0.00" represents a negative value, that rounds to zero when only two decimal places are given. For example, -0.003 would be reported by that software as -0.00, whereas -0.007 would be reported as -0.01.
Good luck with your work.
Hello Morse, thank you for your response probably I couldn't make you understand my question. My question is why minus values in both bounds of class interval in some cases? Is it related to sampling or directionality like in correlation coefficient?
Hello again, Sadiq,
The fact that both ends of the confidence interval are negative means that the best guess concerning the true value of the parameter of interest is that it is negative. (And yes, negative values can occur in linear models.)
Good luck with your work.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling