View Tweet
this data is totally incorrect.
interestingly, the estimates have grown over time. here I compare the 2023 results to the 2024 results.
so pope is now saying that irregular church attendance is WAY higher than survey estimates say it is
but regular attendance is WAY lower than surveys say
this is clearly just a case of measurement error.
but regular attendance is WAY lower than surveys say
this is clearly just a case of measurement error.
most surveys ask something like "About how often do you usually attend services?"
if somebody says "every week" but they actually only attended 48 sundays because they were sick once and on vacation a couple times, they correctly chose "every week"
if somebody says "every week" but they actually only attended 48 sundays because they were sick once and on vacation a couple times, they correctly chose "every week"
now, the definition used for "every week" is "36 out of 47 weeks of data" it would be surprising to me if incidental absences would explain all of that!
and I commend the author for following up the 2023 version with a n~5k sample asking people religion + cell phone behaviors.
he found almost a third of Jews don't take their phones to church...
... and that's almost a third of Jews who take online surveys!
he found almost a third of Jews don't take their phones to church...
... and that's almost a third of Jews who take online surveys!
without getting too into the weeds I want to note that Prolific's sample is infamously Extremely Online. Their initial sample growth was largely from viral TikTok videos. They've worked to correct that some since, but alas, their respondents remain Very Online.
so when we see this, we should read it carefully.
it shows us that people are a LOT less likely to bring their phone to church than elsewhere (bottom 3 rows).
and we can assume some who bring it turn it off.
it shows us that people are a LOT less likely to bring their phone to church than elsewhere (bottom 3 rows).
and we can assume some who bring it turn it off.
so far from this validating the results, it problematizes them. people really do selectively leave behind phones when they go to church, and turning off behavior may be even higher.
even in other cases, fitness is, uh. well. this graph is presented as evidence the data is reliable.
notably, the cell data actually does okay for predicting baseball game attendance. still not that great.
i will note there's an obvious reason cell phones predict baseball game attendance better than basketball game attendance.
cell signal may be more obstructed inside large concrete structures than in outdoor arenas.
cell signal may be more obstructed inside large concrete structures than in outdoor arenas.
in fact, physical barriers to cell signal are so well-known to have social effects we actually use them as variables for identifying random variation in cell phone use behavior!
so i ask you
is church usually
inside
or
outside
is church usually
inside
or
outside
church is usually inside.
therefore, measurement of church attendance, especially at older churches made out of stone, will face the problem of diminished signal quality.
the other issue of course is many churches are in areas with literally no signal.
therefore, measurement of church attendance, especially at older churches made out of stone, will face the problem of diminished signal quality.
the other issue of course is many churches are in areas with literally no signal.
but like i lose signal when i'm in costco and home depot, it's just wild to me to use percentage of pings coming from in a church as the measure given that buildings sometimes obstruct signal.
so, using cell phones to estimate population parameters--- bad idea!
now what I TOTALLY buy is the paper segment following people to other stuff!
cell phones may be pretty bad at precisely estimating a population-wide parameter, but I agree a phone appearing at church is WAY likelier to be that religion than some other phone!
cell phones may be pretty bad at precisely estimating a population-wide parameter, but I agree a phone appearing at church is WAY likelier to be that religion than some other phone!
and so we CAN use "phones observed at church X" to say "where else do people at church X tend to go".
THAT is very cool work. And I think totally credible.
THAT is very cool work. And I think totally credible.
Generated by Thread Navigator
Press ⌘ + S to quick-export