It is a “target”…. that they will most likely miss!
1 Like
I note that even people with some statistics training usually have spent little or no time learning about dependent probabilities, they have mostly just worked with independent probabilities. In my opinion this alone is a real problem with college admissions discussions where the probabilities are not fully independent.
Then there is a lot of behavioral research on probabilities and the problems people have really understanding what they mean in practical contexts. Low-probability/high-impact events in particular are notoriously hard for people to really process well when making decisions. Here is a recent summary if people are interested:
https://www.nature.com/articles/s41599-024-03403-9
The basic observation is “people tend to either overestimate the probability of low-probability high-impact events or discount them entirely.”
That article goes on to explain how biased information searches can play a role in this effect. I find it “interesting” to think about paragraphs like this in connection to what I have seen during my now years of participating in online conversations about college admissions:
First and foremost, information search apparently serves a hedonic purpose as well as an instrumental one, meaning that we tend to seek out information that makes us feel good and avoid information that makes us feel bad (e.g., Golman et al., 2017; Sharot and Sunstein, 2020; Charpentier et al., 2018), leading to, for example, rejection of medical screenings (e.g., Dwyer et al., 2015; Persoskie et al., 2014). This implies an unwillingness to seek out information on events with a substantial negative impact, simply due to their unpleasant nature. Similarly, there is some indication that people prefer to avoid uncertainty (ambiguity aversion, e.g., Ellsberg, 1961; Keren and Gerritsen, 1999; but see Kocher et al., 2018). This is potentially yet another reason why people might avoid seeking out information on events with low probability, since such events are, per definition, associated with substantial uncertainty. Lastly, people tend to seek out information that confirms existing beliefs or favored hypotheses while avoiding information that does not (e.g., Johnston, 1996; Jones and Sugden, 2001; Jones and Sugden, 2001; Koriat et al., 1980), meaning that beliefs shaped by already biased thinking are unlikely to change just because the information is (in theory) available.
The article also explains the role environmental factors can play in these issues:
The efficacy of information search is highly dependent on the degree to which the information is available in the first place. Availability of information (or the lack thereof) might of course be due to intentional promotion or suppression of certain information but can also be a consequence of more organically emergent environments where information consistent with one’s previously held convictions is premiered, either due to homogenized communities or algorithmic curation, generally referred to as echo chambers and filter bubbles respectively. The primary implications of such environments on low-probability high-impact events specifically is presumably to strengthen the tendencies to either discount or overestimate the probability of events that are already present in an individual. Research on such environments are generally situated in the context of ideological polarization and the reduction thereof, indicating that such polarization can be exacerbated by the structure of, for example, social media platforms (e.g., Bakshy et al., 2015; Barberá, 2015; Levy, 2021). Nevertheless, the structure of the social media platforms can equally be used to reduce polarization; a study by Combs et al. (2023) demonstrated that anonymous cross-party conversations through a mobile chat platform were effective in reducing ideological polarization and that such effects were correlated by the civility of the dialogue. This implies that social media platforms that are, on the one hand, structured in such a way as to promote a plurality of perspectives and, on the other hand, curated to ensure civility, would reduce homogeneity and thereby improve availability of information.
This paragraph outlines the dangers of certain sorts of online communities, but also ends more promisingly about other sorts of online communities. I for one do my best to contribute to College Confidential being the more helpful sort of online community for kids and parents engaging in information searches about college admissions.
Also of possible relevance to college admissions is what it calls the description-experience gap (“The description-experience gap denotes that, in otherwise comparable situations, people tend to act as if they overestimate low probabilities when making decisions from description (i.e., when probability information is explicitly stated) but act as if they underestimate low probabilities when making decisions from experience (i.e., when the outcomes of choices are experienced; e.g., Hertwig et al., 2004; Hertwig and Erev, 2009).”). This is a challenging one because a lot of the issues may come from individuals having at most a small sample of relevant experiences, but as a result I do think the collective sharing of experiences can possibly help.
There is also a section on what are sometimes called the availability and affect heuristics, which can actually interfere with the application of formal statistical training, leading to the idea that perhaps “debiasing” is sometimes most useful:
Predicting or controlling the impact of either availability or affect would be a very complicated matter, and it is arguably more productive to aim for interventions to “debias” thinking. There have been some promising results in domain-general debiasing through training interventions (e.g., Fischoff, 1982; Hirt and Markman, 1995; Morewedge et al., 2015), particularly for formal training in inferential rules (Nisbet et al., 1987), although more research is needed to confirm their effectiveness in the context of low-probability high-impact events specifically. However, it is arguably equally important not to communicate risk information in ways that lean into biases in the first place; information on low-probability high-impact events should be communicated in a clear and concise manner that avoids reliance on popular narratives (availability) or emotionally provocative content (affect). This might simultaneously be one of the most important and most challenging endeavors, since popular narratives and emotionally charged content are common tools for increasing engagement in the current media landscape.
Again I think there are some possible tips for us in these insights.
Finally there is a section on how to more effectively communicate risk. I won’t even try to summarize that but there are more things I see as tips in that section.
Anyway, point is the sorts of issues we encounter in these conversations are often well-known challenges for humans trying to gather and process information in the modern world. And while there are not necessarily any silver bullets, there are things we can think about when participating in online communities that may help.
6 Likes
Perhaps the most obvious example of discounting low probability high impact events is risky car driving behavior. For any given drive, the likelihood of crashing is small, so many drivers discount the risk.
But over many drives, added risk from risky driving increases the probability of a crash during one of the many drives to where it may not be low probability any more.
3 Likes
However, people don’t actually understand what simple terms, like “10% acceptance rate” mean in real life contexts, like college admissions. I think that people don’t really have a sense of how percentiles relate to what happens to an individual. Just to begin with, they don’t fully internalize that “probability” doesn’t change the fact that there are really only two outcomes for any individuals - accept or reject (defer is not an outcome, it’s a deferral of the outcome).
You can see that when a kid or a parent complains that “it’s unfair” that the kid was rejected from a college that the kid was “supposed to have” a 75% chance of acceptance to.
Also importantly, people have different perceptions of “chances” depending on how these are expressed.
Saying “5% acceptance rate” seems a much larger chance than “a 95% rejection rate”, while “a 40% acceptance rate” seems more like a toss up that “a 60% rejection rate”.
Similarly, telling a person that there is a 15% chance that they will be rejected provides a different perspective from “an 85% chance of acceptance”.
All of these are ways in which people don’t really understand probability.
It seems to me, in the context of the population of the typical type of student vying for the top tier schools , most DO understand what the acceptance rate means. If not, they certainly know how to find it out. I wouldn’t sell this population short.
I tend to agree. The challenge, often times, is that they think they are better than the average applicant - when they are actually worse.
7 Likes
One last post in the topic. People are naturally bad at understanding probabilities. They need to do more than take a lecture or read one article to start understanding probabilities. It is not opinion, it is not imagination, it is a fact.
4 Likes
“Chance me”s drive me nuts. 90 percent of them are kids who think they are studs but are pretending to be humble/self-deprecating, and they’re applying to schools that are so selective that it’s impossible to chance them. I suppose there is no way to prevent posts like this from kids who are hungry for affirmation, but I wonder if it would be better if the adults on this board didn’t engage on such posts.
11 Likes
Terrific suggestion. Maybe the stock answer- “These colleges are competitive for every single applicant, no matter how talented and special, please make sure you have a robust list of more likely schools as well” and then Lather Rinse Repeat is the way to go.
Frankly, the kids don’t listen anyway (or so it seems). And then we get into a pointless and argumentative cycle of the kid insisting that if we knew how rigorous their HS was we’d understand how special they are (hey kid- then go talk to your “rigorous” HS guidance counselor) and the “hive” telling the kid that 24 valedictorians, all with perfect GPA’s, and the kid listing a bunch of 3’s and 4’s on the AP’s suggests “not quite enough rigor” for Cal Tech….
It really helps nobody.
2 Likes
I’m not sure if you’re being tongue in cheek or not.
Spend ten minutes on reddit a2c. I think we can all agree that in general, 90% of people here are trying to be helpful. Not the case on reddit.
It’s the chances forum. We have some kind of knowledge, but we aren’t AO’s. I guess generalizing is the name of the game.
3 Likes
I think I agree with this. Where they are clearly just wanting affirmation, we should give them a stock answer like this and not indulge.
1 Like
maybe 80%
1 Like
Okay, let’s compromise and say 85%
1 Like
Where people are helpful is raising questions for the applicants to consider, such as affordability, fit based on size and location, alternative/lookalikes that are not so selective and culling the list of final applications.
5 Likes
Perhaps otherwise known as “issue spotting,” which often makes sense from a perspective of wisdom, but may or may not be received with appreciation.
7 Likes
I think the stock answer (for the high stats kids looking for affirmation) might be the way to go. It always makes me sad when people start pulling apart their profile, commenting on “just how high did they go in math” or “2 years of Chinese, 2 years of Latin just won’t cut it.” We’re an anonymous forum with little true context for the individual and we don’t read the minds of AOs. Damaging the self-esteem of 17 year olds with our “reality checks” just doesn’t seem like a good use of our time. I’m not saying we all do this—but I’ve seen enough of it that I tend to avoid all Chance Me posts now. I mostly want to tell those kids: You are great. You are special. You will be ok. And I’m so sorry that the college application process has become so fraught that you feel compelled to bring your case here in an attempt to feel some sense of control.
14 Likes
My concern, as many know, is finances. I’m dealing with a youngster who never talked with his parents. He thought he could live home, attend the state school, and afford college.
Turns out, while his parents can pay, they’re not - so he’s kind of in a pickle and wants to go to med school.
I suggested talking to them of course, but maybe having to go to CC but they look down on that - even though they’re not willing to fund him.
I hate that so many families don’t have this discussion up front. If you don’t have funding, it doesn’t matter how great a candidate one is.
People need to have this discussion up front vs. later.
8 Likes
Awesome.
3 Likes
Agree 100%. I can’t shake the feeling that some commenters relish the “reality check” role. Some things need to be said, but we need to be self-aware about our motivations.
Certain types of reality checks are needed. (The kid who thinks OOS publics are going to give her need-based aid… the one who thinks his outside scholarship will cover his EFC… the one who will be disqualified for the UC’s because they’ve never taken a visual/performing arts class.) Knowledge is power, and if there are facts a kid needs to know that will impact their process, it would be irresponsible not to tell them - especially if it’s something they can still do something about.
For sure, the kid with no safety/likely schools needs to be urged to find some and offered help with that, even though it may not be what they asked and they may decline the help. The financial questions should be raised, although if the kid or parent assures us that they’ve got that part handled, we needn’t continue to harp on it.
The proposed “stock answer” is excellent. Sometimes we have more thoughts to add that could be helpful, but normalizing a stock answer that declines to read tea leaves seems like a great idea. And posts that emphasize the importance of a balanced list, and help stressed-out kids (and parents) to see that not only are they special, but that all sorts of colleges are special, too, can make a difference even if the OP isn’t the reader who takes that message to heart.
16 Likes
agree, 100%. I am always sad when kids haven’t had any communications with their parents, or parents have been super vague.
Though also on these boards, people seemingly ignore the fact that at the top private a large percentage of students are full pay. It isn’t rare (Colby, it is over 50%, Princeton over 40%). Sometimes, people on CC seem not to believe students when they say their parents can/will pay. 
It may be that fewer of these people are on CC. If you are a billionaire ,you have private counselors, but if you are a ‘normal’ full pay ,you may be relying on your normal private school counselor only and may well be on here.
2 Likes