one way is residual scatter plots (residual vs. independent variable) which kind of the pattern from the plot we could conclude that there is heteroskedasticity? Thanks.
What your equation predicts and what actual data look like make up the serial correlation case. So if the plot shows actual data points clusterd above or below the predicted line from the regression, with or without patter, it indicates serial correlation. Heroskedasticity is the variance of the error term, that is how the errors (predicted minus actual) vary around their own mean. They should show consistent yet random variation. Residual vs. independent variable, as you say, helps with serial correlation, but not directly helpful for heteroskedasticity…however, check with some quant heads, they may know better.
The scatterplot will show more variance as the value of the independent variable changes. Check out this graph http://www.unc.edu/courses/2007spring/biol/145/001/gifs/lab12/lab12scatterplot1.png
Page 122 Schweser Slidebook/video has a great graph. If you don’t have the book. Imagine the plot against the line. The left side vs the right side of graph look different. One is closer to the line, the other is further from the line. Schweser drew two circle around the error terms, one circle was significantly bigger than other. Hence the variance of the error terms are not consistent.
^ wtf? How much time you invested in this story?
^^ amazing story 
its easier to remember when you add sex to such dry topics 
I am both disturbed and entertained by your story.
According to my firm we are about to celebrate PRIDE month, what a great way to incorporate the PRIDE! BRAVO… If I had the balls I would ask your permission to have your example posted on our intranet as proof that we as a financial society are closer to acceptance. Is there a better way to show tolerability than to use homosexuality in a CFA example? I think not!
I’ve got stories for all the silly things we have to memorize on this stupid test.
Joking aside, if you get no regression chart then you can detect heteroskedasticity with breusch pagan test.
The story works better for me if the person is a bi-sexual woman.
Sorry for the above post. I am not intending to insult heterosexuals, homosexuals, or bi-sexuals. I cannot edit it so Moderator please delete prior post. Atomicrooster… you have permission to post the post as evidence why you never should post anything that involves, politics, religion, sexual orientation, etc ever any time. A quick submit is too easy to say something stupid… to think I thought avoiding twitter was enough.
Please delete mine as well with my apologies.