My totally biased guess about the correlation with political viewpoints:
1) a large chunk of people who also tend right wing will consider it their duty to "fix" their straying friend or family member. They'd much rather preach at their sinning family member, or punish them, than cut them off. (The well publicized tendency to expel and shun gay offspring - especially prominent some decades ago -would seem to directly contradict this, but I'm guessing the numbers were smaller than the chronic attempts at re-conversion.)
2) the idea of cutting off "toxic" people and "toxic" relationships seems to be encouraged in therapy culture (not necc. the same as by actual therapists). It's especially supported in all kinds of support groups for people with various problems. Possibly this comes originally from advice commonly given to those dealing with addicts - they are commonly advised that you can't help the addict in your life; all you can do is farther enable their addiction.
So when the "relationship" degenerates towards an end state of fighting about politics - or worse, fighting about whether or not the person on the left is some kind of sinner, needing to repent - the one on the left is more likely to pull the plug.
No data to support this except anecdata, from observing the large number of queers and ex-Christians among my acquaintances.
Mmmmmm, I kinda vibe with this. I definitely buy that, as a relationship deteriorates that each side has...preferred strategies that are highly correlated with political affiliation.
I'm just struggling with how I'd dive deeper into this. For example...
mmm
would we predict that extreme liberals who are religious would be significantly less likely to cancel someone than extreme leftists who aren't religious?
I'm not sure there's actually enough cancellation cases to support filtering that tightly but I'm struggling to think of other ways to dive into this.
>would we predict that extreme liberals who are religious would be significantly less likely to cancel someone than extreme leftists who aren't religious?
"extreme liberals who are [very Christian]" is a pretty narrow and specific cross-section that, from my experience of knowing several, tend to be those who've doubled down on the "Jesus said love thy neighbor, yes that means queer people too" part of the doctrine, and are often committed to spreading that message to conservative religious types, so that matches my anecdotal experience.
Didn't Scott include questions about religion and seriousness-about-it on the survey? Could you add that axis to some of the comparisons? Or is Scott's audience so atheistic that the sample size would be too low?
It's the Marxism and the cancelation, not the religious aspect. There's only ~130 Marxist respondents in the dataset out of ~6800. So, just spitballing, if 20% of Marxists are involved in a cancellation in some way, we're already down to 26 respondents. That's already below 30, which is kind of a magic number in stats for sample size.
So, like, I could break it down further but I'd be trying to draw conclusions by comparing, like, 9 religious Marxists involved in cancellations to 17 atheist Marxists or something. There's just not enough Marxists and not enough cancellations to dig in that deeply.
I don't think people call it "cancelling" when its just one person deciding to cut contact with another, cancelling has always implied a mob to me. Strange wording choice.
Anyway, I do question a little how much can really be learned from dividing this into cutting off vs. being cut off. For example, take a scenario where a queer person comes out or begins to make queer lifestyle decisions (marriage etc), and then conservative family members react with sustained disapproval and hostility. Regardless of whether the final action of cutting off was initiated by the queer person ("I cant take this anymore") or by the other family members ("You're no longer allowed around my children"), it seems that the arrow of causation still points the same way. Maybe both parties would disagree about which cut the other off, or whether an example like this counts as being cut off "over politics".
I think when it comes to personal values, we could guess that for example liberal people are more likely to see cutting off a family member as the appropriate response to the tensions in a relationship, but that doesn't necessarily tell us much about the interpersonal family dynamics that actually lead to people cutting one another off, and by extension your likelihood of being cut off by or cutting off a family member in the future.
I don't think there's any disagreement that in either case there's definitely a deterioration in the relationship that all parties have some responsibility for. To paraphrase an old Bill Burr bit, it's not like someone is just pleasantly chatting about the weather and a big family fight jumps out of nowhere. And I tried to touch on this a few times with bits about "risk of being in a cancellation". So the point that both parties share some responsibility is well taken.
But I don't think that responsibility is even. Ending the relationship, cutting off contact, is a big deal and feels like a big escalation and a big step. Barring any other evidence, my guess is the responsibility is 70% on the side of the canceler and 30% on the side of the canceled. Maybe you feel that it's 60-40 or 80-20 and I think we're on the same page that it's not 100-0 or 50-50.
Why is 50-50 the extreme? As I read it, the main point of the hypothetical above was that the family was overwhelmingly responsible for the irreparable relationship regardless of who's wearing the "canceler" hat when the music stops, but this analysis (and tbf the dataset) can't distinguish that. My prior is that there are lots of these cases among family rifts - the same OoM as ones where majority responsibility factually tends toward the canceler, at least - so we're working in a pretty dense fog if we want to read something about individual propensities from it.
Mostly because knowing who initiated the cancellation is some kind of knowledge, it must affect our analysis somehow, it's just unclear how much.
So, to draw an example, let's say that we had a case where a relationship ends; it could be a family cancellation here, it could be a business partnership, it could be a military alliance, it could be a glorpnorp agreement between two silicon aliens. Absent any other information than the relationship between the two parties ended, we would assign responsibility at 50-50, right? Now add one additional piece of information: that one party decided to end the relationship. Absent any and all other information than that, it seems fair to attribute some additional share of responsibility to the party that ended it.
Given sufficient priors, you may believe something different but, just in the absence of any priors, that additional piece of information should shift responsibility in a certain direction.
On a extremely tangential note, is lifetime risk of getting shot (I assume we're talking US here, so this doesn't really affect me- I'm pretty sure my risk is a great deal lower) as high as 1%? That seems intuitively way too high to me, and I would kind of expect to overestimate this sort of thing rather than underestimate.
I like the write-up, and great stuff for posting your R-code! Hope it's okay to suggest, but I like using 'case_when' now instead of multiple if-else statements (https://dplyr.tidyverse.org/reference/case_when.html), makes my life a little bit easier with multiple conditions!
Oh, not at all, I love anything that helps me write better code. And that looks very SQL-ish, which is good, everyone gets SQL.
I'm just getting much more hesitant about libraries in my old age. External dependencies are bad but...man, that is a lot cleaner....Thanks for the suggestion, I think I'll try that next time.
Wild guess on the weird age buckets: the 16-20ish age range is a time of relatively high risk (*per year*-- as you note lifetime risk still looks low, b/c less lifetime) of cancelling, or being cancelled by, *parents specifically*. That would be teenage rebellion and/or big ideological changes in college. And then the parent's side of that would be the spike in middle age.
The section on income and poverty makes me wonder if, say, Marxists tend to report themselves as being in a lower socioeconomic class for a given income level.
Yeah, that's very plausible. God, the wording is tough here to, because there's a big difference between "canceling" someone at 16-18 vs 18-20.
The Marxist thing is checkable though, one sec...
Oooh, ok, code at bottom,
Marxist Average Income=~$71k
Non-Marxist Average Income=~$131k
Marxist Median Income=$67.5k
Non-Marxist Median Income=$95k
So Marxists are poorer, now self perception
Marxist:
Poor 5%
Working 15%
Middle 43%
Upper 31%
Rich 2%
Other 3%
Non-Marxist:
Poor 1%
Working 7%
Middle 36%
Upper 47%
Rich 7%
Other 1%
Sooo, maybe we'd find something if we dug in more but at a high level, it looks like Marxists both make less and self-perceive as being of a lower social class. Sorry Marxists.
My totally biased guess about the correlation with political viewpoints:
1) a large chunk of people who also tend right wing will consider it their duty to "fix" their straying friend or family member. They'd much rather preach at their sinning family member, or punish them, than cut them off. (The well publicized tendency to expel and shun gay offspring - especially prominent some decades ago -would seem to directly contradict this, but I'm guessing the numbers were smaller than the chronic attempts at re-conversion.)
2) the idea of cutting off "toxic" people and "toxic" relationships seems to be encouraged in therapy culture (not necc. the same as by actual therapists). It's especially supported in all kinds of support groups for people with various problems. Possibly this comes originally from advice commonly given to those dealing with addicts - they are commonly advised that you can't help the addict in your life; all you can do is farther enable their addiction.
So when the "relationship" degenerates towards an end state of fighting about politics - or worse, fighting about whether or not the person on the left is some kind of sinner, needing to repent - the one on the left is more likely to pull the plug.
No data to support this except anecdata, from observing the large number of queers and ex-Christians among my acquaintances.
Mmmmmm, I kinda vibe with this. I definitely buy that, as a relationship deteriorates that each side has...preferred strategies that are highly correlated with political affiliation.
I'm just struggling with how I'd dive deeper into this. For example...
mmm
would we predict that extreme liberals who are religious would be significantly less likely to cancel someone than extreme leftists who aren't religious?
I'm not sure there's actually enough cancellation cases to support filtering that tightly but I'm struggling to think of other ways to dive into this.
>would we predict that extreme liberals who are religious would be significantly less likely to cancel someone than extreme leftists who aren't religious?
"extreme liberals who are [very Christian]" is a pretty narrow and specific cross-section that, from my experience of knowing several, tend to be those who've doubled down on the "Jesus said love thy neighbor, yes that means queer people too" part of the doctrine, and are often committed to spreading that message to conservative religious types, so that matches my anecdotal experience.
Didn't Scott include questions about religion and seriousness-about-it on the survey? Could you add that axis to some of the comparisons? Or is Scott's audience so atheistic that the sample size would be too low?
It's the Marxism and the cancelation, not the religious aspect. There's only ~130 Marxist respondents in the dataset out of ~6800. So, just spitballing, if 20% of Marxists are involved in a cancellation in some way, we're already down to 26 respondents. That's already below 30, which is kind of a magic number in stats for sample size.
So, like, I could break it down further but I'd be trying to draw conclusions by comparing, like, 9 religious Marxists involved in cancellations to 17 atheist Marxists or something. There's just not enough Marxists and not enough cancellations to dig in that deeply.
I don't think people call it "cancelling" when its just one person deciding to cut contact with another, cancelling has always implied a mob to me. Strange wording choice.
Anyway, I do question a little how much can really be learned from dividing this into cutting off vs. being cut off. For example, take a scenario where a queer person comes out or begins to make queer lifestyle decisions (marriage etc), and then conservative family members react with sustained disapproval and hostility. Regardless of whether the final action of cutting off was initiated by the queer person ("I cant take this anymore") or by the other family members ("You're no longer allowed around my children"), it seems that the arrow of causation still points the same way. Maybe both parties would disagree about which cut the other off, or whether an example like this counts as being cut off "over politics".
I think when it comes to personal values, we could guess that for example liberal people are more likely to see cutting off a family member as the appropriate response to the tensions in a relationship, but that doesn't necessarily tell us much about the interpersonal family dynamics that actually lead to people cutting one another off, and by extension your likelihood of being cut off by or cutting off a family member in the future.
I don't think there's any disagreement that in either case there's definitely a deterioration in the relationship that all parties have some responsibility for. To paraphrase an old Bill Burr bit, it's not like someone is just pleasantly chatting about the weather and a big family fight jumps out of nowhere. And I tried to touch on this a few times with bits about "risk of being in a cancellation". So the point that both parties share some responsibility is well taken.
But I don't think that responsibility is even. Ending the relationship, cutting off contact, is a big deal and feels like a big escalation and a big step. Barring any other evidence, my guess is the responsibility is 70% on the side of the canceler and 30% on the side of the canceled. Maybe you feel that it's 60-40 or 80-20 and I think we're on the same page that it's not 100-0 or 50-50.
Why is 50-50 the extreme? As I read it, the main point of the hypothetical above was that the family was overwhelmingly responsible for the irreparable relationship regardless of who's wearing the "canceler" hat when the music stops, but this analysis (and tbf the dataset) can't distinguish that. My prior is that there are lots of these cases among family rifts - the same OoM as ones where majority responsibility factually tends toward the canceler, at least - so we're working in a pretty dense fog if we want to read something about individual propensities from it.
Mostly because knowing who initiated the cancellation is some kind of knowledge, it must affect our analysis somehow, it's just unclear how much.
So, to draw an example, let's say that we had a case where a relationship ends; it could be a family cancellation here, it could be a business partnership, it could be a military alliance, it could be a glorpnorp agreement between two silicon aliens. Absent any other information than the relationship between the two parties ended, we would assign responsibility at 50-50, right? Now add one additional piece of information: that one party decided to end the relationship. Absent any and all other information than that, it seems fair to attribute some additional share of responsibility to the party that ended it.
Given sufficient priors, you may believe something different but, just in the absence of any priors, that additional piece of information should shift responsibility in a certain direction.
On a extremely tangential note, is lifetime risk of getting shot (I assume we're talking US here, so this doesn't really affect me- I'm pretty sure my risk is a great deal lower) as high as 1%? That seems intuitively way too high to me, and I would kind of expect to overestimate this sort of thing rather than underestimate.
I like the write-up, and great stuff for posting your R-code! Hope it's okay to suggest, but I like using 'case_when' now instead of multiple if-else statements (https://dplyr.tidyverse.org/reference/case_when.html), makes my life a little bit easier with multiple conditions!
Oh, not at all, I love anything that helps me write better code. And that looks very SQL-ish, which is good, everyone gets SQL.
I'm just getting much more hesitant about libraries in my old age. External dependencies are bad but...man, that is a lot cleaner....Thanks for the suggestion, I think I'll try that next time.
Wild guess on the weird age buckets: the 16-20ish age range is a time of relatively high risk (*per year*-- as you note lifetime risk still looks low, b/c less lifetime) of cancelling, or being cancelled by, *parents specifically*. That would be teenage rebellion and/or big ideological changes in college. And then the parent's side of that would be the spike in middle age.
The section on income and poverty makes me wonder if, say, Marxists tend to report themselves as being in a lower socioeconomic class for a given income level.
Yeah, that's very plausible. God, the wording is tough here to, because there's a big difference between "canceling" someone at 16-18 vs 18-20.
The Marxist thing is checkable though, one sec...
Oooh, ok, code at bottom,
Marxist Average Income=~$71k
Non-Marxist Average Income=~$131k
Marxist Median Income=$67.5k
Non-Marxist Median Income=$95k
So Marxists are poorer, now self perception
Marxist:
Poor 5%
Working 15%
Middle 43%
Upper 31%
Rich 2%
Other 3%
Non-Marxist:
Poor 1%
Working 7%
Middle 36%
Upper 47%
Rich 7%
Other 1%
Sooo, maybe we'd find something if we dug in more but at a high level, it looks like Marxists both make less and self-perceive as being of a lower social class. Sorry Marxists.
q_df<-cleaned_df[cleaned_df$Political_Affiliation=="Marxist", ]
q2_df<-cleaned_df[!(cleaned_df$Political_Affiliation=="Marxist"), ]
mean(q_df$Income)
mean(q2_df$Income)
median(q_df$Income)
median(q2_df$Income)
table(q_df$Social_Class)
table(q2_df$Social_Class)
Just wanted to say thanks for doing this!
Thanks!