SoylentNews is people
SoylentNews
Scott Jaschik writes at Inside Higher Education that although most faculty members would deny that physical appearance is a legitimate criterion in grading, a study finds that among similarly qualified female students, those who are physically attractive earn better grades than less attractive female students. For male students, there is no significant relationship between attractiveness and grades. The results hold true whether the faculty member is a man or a woman.
The researchers obtained student identification photographs for students at Metropolitan State University of Denver and had the attractiveness rated, on a scale of 1-10, of all the students. Then they examined 168,092 course grades awarded to the students, using factors such as ACT scores to control for student academic ability. For female students, an increase of one standard deviation in attractiveness was associated with a 0.024 increase in grade (on a 4.0 scale).
The results mirror a similar study that found that those who are attractive in high school are more likely to go on to earn a four-year college degree. Hernández-Julián says that he found the results of the Metro State study “troubling” and says that there are two possible explanations: “Is it that professors invest more time and energy into the better-looking students, helping them learn more and earn the higher grades? Or do professors simply reward the appearance with higher grades given identical performance? The likely answer, given our growing understanding of the prevalence of implicit biases, is that professors make small adjustments on both of these margins."
(Score: 1, Interesting) by Anonymous Coward on Thursday January 07 2016, @03:58PM
For female students, an increase of one standard deviation in attractiveness was associated with a 0.024 increase in grade (on a 4.0 scale).
So a whopping 6/10ths of a percent? That seems to beggar belief that it's statistically significant.
(Score: -1, Insightful) by arcz on Thursday January 07 2016, @04:34PM
0.024 is not 0.024%
0.024 = 2.4%
Per cent = per 100
n% = n/100
(Score: 1, Troll) by arcz on Thursday January 07 2016, @04:37PM
misread your comment, ignore that
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @04:37PM
What. The. Fuck?
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @04:44PM
measured note deviation / note range = 0.024 / 4 = 0.006 = 0.6% = 6/10 of a percent.
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @06:15PM
3/5 of a percent.
(Score: 0) by Anonymous Coward on Friday January 08 2016, @09:38AM
Err … you know that 6/10 and 3/5 are exactly the same number? And you did notice that the OP did use 6/10, not 3/5?
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @04:36PM
They examined 168,092 course grades. 1/sqrt(168,092) is approximately 0.00244. I'm pretty sure the standard deviation of the grades is at least 1, probably larger, making their measured deviation at most one standard deviation, probably less. So I agree, that's most probably not statistically significant.
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @04:42PM
Err … I just notice that their number was 0.024, not 0.0024 — so I take that conclusion back and state the opposite: At a factor 10 it almost certainly is significant!
(Score: 1) by CHK6 on Thursday January 07 2016, @04:59PM
I was wondering, so what value would have been acceptable by those conducting the study to find there was no difference? If significant digits were used, then a 0.0 would mean there was no difference.
Sigh....
(Score: 0) by Anonymous Coward on Friday January 08 2016, @09:50AM
Usually a threshold of p=0.05 is used. Without knowing the distribution, I cannot tell for sure, but given that the standard deviation cannot be larger than sqrt(8)~2.8 (because of the scale size of 4.0; the worst case is if half the students got the best mark, the other half the worst — the real standard deviation is almost certainly much smaller; I'd guess something in the order of 1), and the sample size of 168092, I figure that everything larger than 0.014 should be definitely significant. The deviation they found, 0.024, is clearly larger.
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @06:23PM
You do realize that this is not referring to % at all, but to GPA... right?
(Score: 0) by Anonymous Coward on Thursday January 07 2016, @06:32PM
.024 / 4.0 = .006 or .6%. So, yes, I do know how to read.