As others have mentioned, the term "effect" might refer either of the
following:

Size of change (in the case of experiments, where we might infer a causal
relationship, such as *If a student studies new content using strategy x
for at least x amount of time, he/se will recall x% more material than a
student who used rote repetition for the same amount of time*). Statistical
procedures exist for estimating these effect sizes (e.g., Cohen's *d*).

or

Strength of relationship - in the case of correlational findings, where we
might argue that we can explain X% of the variation in scores on some
variable based on its relation with (or an individual or group's score on)
some predictor variable (or variables, in the case of a multivariate
model). This can also be estimated statistically, using r-squared (percent
of variablity explained).



_____________________________________________

Claudia J. Stanny, Ph.D.
Director
Center for University Teaching, Learning, and Assessment
BLDG 53 Suite 201
University of West Florida
Pensacola, FL  32514

Phone:   (850) 857-6355 (direct) or  473-7435 (CUTLA)

[email protected]

CUTLA Web Site: http://uwf.edu/offices/cutla/ <http://uwf.edu/cutla/>

---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here: 
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=51057
or send a blank email to 
leave-51057-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu

Reply via email to