Confidence Intervals vs. SEm

[edieb]

---

Members do not see this ad.

I want to make sure I understand something: a confidence interval is the range in whic a true score (i.e., score without error variance component) may lie whereas a Std Error of measurement still includes the error variance but is where a person's score would lie if he/she took parallel forms over and over.

What is the point of calculating both? It seems that a confidence interval, since it excludes error variance, would suffice...


Thanks

 

[50960]

I am pretty sure a CI includes error variance. It means that 98% (or whatever) of the time you sampled the same population you would get numbers within this range..it is predictive. SEM is descriptive.

 

[edieb]

Thanks for the response, but a CI does not include error variance; rather it the interval in which the true score (i.e., score without error variance) lies.

 

[50960]

If by error variance you mean error of measurement then you are right. It does include random error.