Undermatch vs Overmatch: the role of Cognitive & Non-Cognitive Skills in application process

College Search & Lists
1

“Gritting It Out” is a a recently revised study was recently released wherein the author attempts to discover the impact non-cognitive skills had on the college application process and graduation rate.

An undermatch occurs when a student applies to a college that her cognitive stats (GPA, standardized test scores) indicate is a less selective college for her academic abilities. An overmatch occurs when a student applies to a college that his stats indicate is a more selective college than warranted. Instead of relying solely on stats to determine undermatch/overmatch, this study also takes into account non-cognitive skills to determine undermatch/overmatch and graduation likelihood

Utilizing his methodology, the author concludes far fewer students undermatch than previously thought. And far fewer students overmatch when accounting for non-cognitive skills.

Here is the link: https://www.sciencedirect.com/science/article/pii/S0272775720305197?casa_token=F1HFCk1i_GoAAAAA:X0I8YhocHhfNmJOTNPmR_2YuS7SUl33nQmDveBt65TF264MtndiGWv9h9xPwoS5DBg8zUZOIMJM
If CC removes the link, the article can be found on the ScienceDirect dot com site. Just search for “ScienceDirect” and “Gritting It Out”.

Here is but one insightful quote from the introduction.

There are lots of wonky stats and stuff going on here, but it is also very interested … at least to me.

1 Like
2

I tried to access this but couldn’t without paying $20. I’m curious what skills the author took into account for “non-cognitive skills”.

3

@melvin123 , that part interested me too. Others sources list skills like Motivation, Effort, Interpersonal and Communication, and self-confidence. I’ll try to go back and see if the author here specifically defined the term for his purposes.

I haven’t paid and was able to access the study. There was a tab on the far right, slightly upper side of the screen that said “View Full Text” or maybe just “Full Text”. When I clicked that, the entire study appeared.

1 Like
4

I went back and extracted this quote. It’s not everything he said about Non-Cog Skills:

  • In particular, four questions measured students’ locus of control, which captures their belief in the control they have over various events in their lives. An additional four questions measure students’ self-concept, which captures students belief in themselves. I consider students’ answers in each measure and also construct an index of non-cognitive skills by averaging across the two — the results are robust to alternative definitions. *
5

It’s not opening for me either :frowning:

Does the author control for financial constraints?

6

That’s disappointing. I’m sorry the link doesn’t appear to be opening for all.

The author was very cognizant for several factors, including SES. A couple of quotes that stood out in this regard were:
“Across both cohorts, students from high-SES families are less likely to enroll in inclusive colleges.”
…and…

  • “Previous work has found that coming from lower SES families (Bailey, Dynarski, 2011, Stinebrickner, Stinebrickner, 2003), being an underrepresented minority student (NCES, 2018) and lower pre-college academic preparation (Agan, 2013, Archibald, Feldman, McHenry, 2015, Bound, Lovenheim, Turner, 2010) are strong predictors of degree non-completion. In this paper, I contribute to this literature by showing that higher non-cognitive skills among students enrolled in less selective four-year colleges are strongly associated with higher degree completion rates.” *
    … and …
  • “HSB includes detailed information on students’ background characteristics, including their gender, race, household composition, parental education and socioeconomic status (SES).” *

I’ll keep looking for a link to access a complete version of the study. I’m stumped why this one doesn’t work for all. I know I haven’t paid to join. Maybe there’s some sort of cross-site affiliation going on that allows me to access it?

7

Thanks @EconPop !

8

Perhaps you went through an institutional account or library at an institution that has a subscription?

9

Wouldn’t parental financial limitations be a strong contributor to “undermatching”?

For example, it is common on these forums for students to have to chase merit at less selective colleges than they can get admitted to, because they and their parents do not have the money to afford the most selective colleges that they can get admitted to.

1 Like
10

Good question that I’m not sure I can yet answer to your satisfaction. For one thing, it is my understanding that schools that some CC users refer to as “lesser” schools are, in this study, included in the top grouping of schools.

The authors states that for the purposes of this study, the highest level of universities is the top quintile. It is not some artificial delineation such as is used in rankings for marketing and sales purposes.

While I know some on CC would argue that is too broad a grouping, I agree with the author that it makes sense to not follow the line of thinking that separates the T-20 from the greater T80. In reality, there is little, if anything, that, for the vast majority of student situations, separates an education from the T10 from an education at the T20-30 from an education at the T40-50, and so one. Whether one attends Hopkins or Rice or Lehigh or CMU or WPI, one has attended one of the absolute best universities in the country. IMO, the study would be doing the research a disservice to include only the T20 in the “most selective” category.

To clarify, the study breaks everything down into quintiles … mostly. But has top 10% then next 10% then quintiles for the rest… Therefore, the study compares students with GPAs and Test Scores in the top 10% and aligns them with universities that rate as the top 10% of all American universities – then the next quintiles, and so forth. Although, at times the study refers to the top 10%, and at other times the top 20%. So it sometimes uses both. I had to drift away from studying the numbers once my girls awakened.

The study isn’t trying to suss out the tiny sliver of a difference between attending Yale or CMU. It is trying to compare attending Emory or attending U of SW MO; Claremont McKenna or Lake Forest. It’s trying to find truth in numbers, not marketing numbers. :smiley:

While I’m not entirely certain because I haven’t tried to determine exactly which universities are in the top quintile yet, I would assume that top-quintile grouping includes quite a few schools that offer both need-based aid and merit-based aid. Such as, some of the schools often mentioned on CC as very good both in educational quality and merit-based aid.

The study refers to schools on a scale that ranks from

11

Some of the situations based on parental financial limitations mentioned here may be analogous to the latter. For example:

  • Student has a reasonable chance to get admitted to top-whatever universities and LACs, but instead has to aim for University of Alabama / Arizona / Mississippi / New Mexico for large merit scholarships.
  • Student could get admitted to the state flagship, but because the state gives poor financial aid to in-state residents, has to commute to the nearby minimally selective non-flagship state university because that is the only affordable choice.
12

Google is your friend ;).

http://econweb.umd.edu/~saltiel/files/noncognitive_mostrecent.pdf

13

The study used two different data sets (HSB and ELS - note that some of the data starts in 1980)

Post 3 has the definition of non-cognitive skills used for the HSB study.

For ELS :

Details on the first two:
Students are prompted with the following statements: ”I can learn something really hard,” ”I can get no bad grades if I decide to,” ”I can get no problems wrong if I decide to,” and ”I can learn something well if I want to.”

The statements are as follows: ”I remember most important things when I study,” ”I work as hard as possible when I study,” ”I keep studying even if the material is difficult,” ”I do my best to learn,” ”I put forth my best effort when studying.”

Their responses are graded on a four-point Likert scale.

14

Does the study take into account cultural issues or regional preferences?

I know girls from various religious/ethnic/cultural groups who almost always “Undermatch” because their families state “You live at home and commute or you don’t go”, or some variation thereof. I know kids of both genders from religious backgrounds where the choices are one of three Bible colleges/denominational groups where their stats put them in the 90th or 95th percentile- but their options are to go to one of these schools or get a job at the local car wash

Doesn’t matter how much grit or tenacity or spunk you have. Not every HS kid has parents pushing them towards CMU or Penn State or Pitt even when the “stats” and personal qualities suggest that these schools would be terrific options academically.

15

That is correct. I’ll speak from my own experience working with kids in rural areas. Unless they have exceptional parents, they prefer to stay with known schools that past graduates attended. Suggesting they apply to elite colleges doesn’t always go over well with the parents, who prefer their child attend a community college for two years and live at home rather than apply to a “fancy” college far away. Some parents might even get angry with an adult who suggests their child apply to a more elite school. I’ve been there.

1 Like
16

How common is overmatching? Seems like it may occur in these situations:

A. Student admitted with academic credentials far below the norm for the college or major. Could include some development/VIP admits and some recruited athletes. Could also include (at colleges that admit by major) some students who choose an unpopular major that is easy to get admitted to, but hard in terms of academic rigor (e.g. philosophy or physics at some colleges).

B. Student achieves highly in high school under close parental supervision, lots of after-school tutoring, etc. and achieved highly on standardized tests with lots of test prep. On their own in college, student struggles academically due to weak self-motivation and time management. New social distractions in college without as much in the way of parental limits can also have an effect.

1 Like