This model seems like it has a lot of political implications! (Also, I'm curious whether taking the GWWC pledge actually correlates more with high Altruism or Dutifulness, or even Openness to Ideas.)
"Psychology has a replication crisis and its findings should not be treated as necessarily reliable."
My heuristic is that social psychology is unreliable whereas differential psychology (i.e. personality and ability psychology) is relatively reliable.
Differential psychology often has good sample sizes, with hundreds of participants per study. Furthermore, personality psychology has the advantage that there are many personality traits, and usually researchers report the associations for all the personality traits, so you don't get much file drawer effects because there's likely to be some interesting results somewhere when you take more variables into consideration.
That said, differential psychology does have its flaws and it always pays to be vigilant. For instance, recently-ish, a leading IQ research claimed to have proven that IQ test scores don't really depend on how much effort you put in to the test. Sounds absurd? Yes, and it turns out his analysis was flawed and actually his data showed that IQ test performance does depend on effort: https://twitter.com/tailcalled/status/1520314311285030912
"The Big Five itself has many degrees of researcher freedom (note that all the factors were given six facets!). All models are fake; some models are useful."
As I understand things, the broad Big Five factors themselves were found empirically, but the facets of the NEO PI-R were invented theoretically. To me, this raises the question of whether one could instead derive the facets empirically.
But beware, it's long and technical. Page 118 literally goes "Are you still here? It’s over. I can’t believe you made it this far. The only explanation is that you’re an incurable insomniac with a strong interest in personality structure." I really enjoyed reading it but also I might have a severe psychometrics addiction, so let me try to summarize what they did:
A central tool in psychometrics is factor analysis (or principal component analysis with sparsity-inducing rotations - for these purposes the tools are basically equivalent). If you are not familiar with FA/PCA, basically they look for correlations between different variables and come up with variables corresponding to those correlations. So for instance if "I am concerned about others", "I am sensitive to the needs of others", "I sympathize with others feelings" and other questions like that correlate with each other, factor analysis might come up with a variable based on those correlations, and then a research might look at the items at that variable and go "Aha! I will name that 'Compassion' because that seems to be what it represents!".
Anyway, in this study, factor analysis was applied at an unprecedented scale. Condon took items from 696 personality items fully covering 19 personality inventories and mostly covering an additional 10 personality inventories, and submitted them all to factor analysis with an enormous sample size. Though that raises a question, how did he get that much sample size when he had that many items? He used a special technique called "Synthetic Aperture Personality Assessment", where people were presented with a small random subset of the personality items. Because factor analysis is linear, it is still possible to apply factor analysis with this sort of massive data "missingness".
Using this data, he uncovered 27 factors, which correlated with 5 domains:
* Conservativism was a lone factor with no simple relationship to the domains
Openness domain:
* Intellect
* Creativity
* Art appreciation
* Introspection
* (Non-)conformity
Extraversion domain:
* Sociability
* Sensation seeking
* Attention seeking
* Humor
* Charisma (interstitial for +Openness, -Neuroticism)
* Well-being (interstitial for -Neuroticism, +Agreeableness, +Conscientiousness)
* Emotional Expressiveness (interstitial for +Agreeableness)
Neuroticism:
* (Un-)Adaptability (interstitial for -Openness, -Extraversion)
* Irritability
* Anxiety
* Emotional (In)stability
* (Lack of) self control (interstitial for -Conscientiousness)
Conscientiousness:
* Industry (interstitial for -Neuroticism)
* Order
* Perfectionism
* (Un-)easy-goingness
* Authoritarianism (interstitial for +Agreeableness)
Does this mean that agreeable people tend to be *more* straightforward? Most popular discussion of Big 5 portrays low-agreeableness people as blunt/honest free thinkers, so if the actual psychometric is the opposite, it's good to get the word out!
This is great, but it would be even better if each facet also had a description of what people that are *low* in this facet are like. I know it kinda just means "the opposite of this", but it would be nice to cash it out explicitly with the same sort of examples.
How were these facets grouped into traits? Some of them intuitively seem like they conceptually belong more in other clusters, e.g. positive emotionality in neuroticism, or impulsiveness in conscientiousness (both with the sign flipped).
One thing I've been trying to understand is how very high Neuroticism differs from Borderline Personality Disorder.
On the one hand, they seemed to be defined and characterized very similarly. And even when people do emphasize differences, such as BPD being characterized by a fear of abandonment, it seems to me that these are not "real" differences i.e. people high in Neuroticism also fear abandonment.
But on the other hand, intuitively they "seem" somewhat distinct to me, and also empirically they seem to be very different (e.g. this study found the genetic correlation between Neuroticism and BPD to be a mere 0.34 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9001677/ ).
I would study it empirically by giving people a BPD test, except I don't feel convinced psychometric tests of this are valid unless I understand why I should believe them to be valid.
"The review outlines the major reasons for the dominance of the Big Five model, and suggests it is a good descriptive framework for studying personality in general. However we suggest that researchers and practitioners also consider alternative taxonomic personality representations such as the HEXACO. We provide a range of scenarios whereby alternative frameworks will be more appropriate than the Big Five and offer recommendations both for choosing measures in general and for implementing studies examining personality facets."
This model seems like it has a lot of political implications! (Also, I'm curious whether taking the GWWC pledge actually correlates more with high Altruism or Dutifulness, or even Openness to Ideas.)
"Psychology has a replication crisis and its findings should not be treated as necessarily reliable."
My heuristic is that social psychology is unreliable whereas differential psychology (i.e. personality and ability psychology) is relatively reliable.
Differential psychology often has good sample sizes, with hundreds of participants per study. Furthermore, personality psychology has the advantage that there are many personality traits, and usually researchers report the associations for all the personality traits, so you don't get much file drawer effects because there's likely to be some interesting results somewhere when you take more variables into consideration.
That said, differential psychology does have its flaws and it always pays to be vigilant. For instance, recently-ish, a leading IQ research claimed to have proven that IQ test scores don't really depend on how much effort you put in to the test. Sounds absurd? Yes, and it turns out his analysis was flawed and actually his data showed that IQ test performance does depend on effort: https://twitter.com/tailcalled/status/1520314311285030912
"The Big Five itself has many degrees of researcher freedom (note that all the factors were given six facets!). All models are fake; some models are useful."
As I understand things, the broad Big Five factors themselves were found empirically, but the facets of the NEO PI-R were invented theoretically. To me, this raises the question of whether one could instead derive the facets empirically.
There are various attempts to do that, with perhaps the best current attempt being the SAPA Personality Inventory. An explanation of its development is available here: https://www.sapa-project.org/research/SPI/SPIdevelopment.pdf
But beware, it's long and technical. Page 118 literally goes "Are you still here? It’s over. I can’t believe you made it this far. The only explanation is that you’re an incurable insomniac with a strong interest in personality structure." I really enjoyed reading it but also I might have a severe psychometrics addiction, so let me try to summarize what they did:
A central tool in psychometrics is factor analysis (or principal component analysis with sparsity-inducing rotations - for these purposes the tools are basically equivalent). If you are not familiar with FA/PCA, basically they look for correlations between different variables and come up with variables corresponding to those correlations. So for instance if "I am concerned about others", "I am sensitive to the needs of others", "I sympathize with others feelings" and other questions like that correlate with each other, factor analysis might come up with a variable based on those correlations, and then a research might look at the items at that variable and go "Aha! I will name that 'Compassion' because that seems to be what it represents!".
Anyway, in this study, factor analysis was applied at an unprecedented scale. Condon took items from 696 personality items fully covering 19 personality inventories and mostly covering an additional 10 personality inventories, and submitted them all to factor analysis with an enormous sample size. Though that raises a question, how did he get that much sample size when he had that many items? He used a special technique called "Synthetic Aperture Personality Assessment", where people were presented with a small random subset of the personality items. Because factor analysis is linear, it is still possible to apply factor analysis with this sort of massive data "missingness".
Using this data, he uncovered 27 factors, which correlated with 5 domains:
* Conservativism was a lone factor with no simple relationship to the domains
Openness domain:
* Intellect
* Creativity
* Art appreciation
* Introspection
* (Non-)conformity
Extraversion domain:
* Sociability
* Sensation seeking
* Attention seeking
* Humor
* Charisma (interstitial for +Openness, -Neuroticism)
* Well-being (interstitial for -Neuroticism, +Agreeableness, +Conscientiousness)
* Emotional Expressiveness (interstitial for +Agreeableness)
Neuroticism:
* (Un-)Adaptability (interstitial for -Openness, -Extraversion)
* Irritability
* Anxiety
* Emotional (In)stability
* (Lack of) self control (interstitial for -Conscientiousness)
Conscientiousness:
* Industry (interstitial for -Neuroticism)
* Order
* Perfectionism
* (Un-)easy-goingness
* Authoritarianism (interstitial for +Agreeableness)
Agreeableness:
* Honesty (interstitial for +Conscientiousness)
* Trust
* Compassion
Does this mean that agreeable people tend to be *more* straightforward? Most popular discussion of Big 5 portrays low-agreeableness people as blunt/honest free thinkers, so if the actual psychometric is the opposite, it's good to get the word out!
ok this is golden "random npc personality table" material!
This is great, but it would be even better if each facet also had a description of what people that are *low* in this facet are like. I know it kinda just means "the opposite of this", but it would be nice to cash it out explicitly with the same sort of examples.
I think one low-hanging fruit that it would probably be beneficial for personality psychologists to engage in would probably be something like this: https://twitter.com/tailcalled/status/1577672258767753217
How were these facets grouped into traits? Some of them intuitively seem like they conceptually belong more in other clusters, e.g. positive emotionality in neuroticism, or impulsiveness in conscientiousness (both with the sign flipped).
One thing I've been trying to understand is how very high Neuroticism differs from Borderline Personality Disorder.
On the one hand, they seemed to be defined and characterized very similarly. And even when people do emphasize differences, such as BPD being characterized by a fear of abandonment, it seems to me that these are not "real" differences i.e. people high in Neuroticism also fear abandonment.
But on the other hand, intuitively they "seem" somewhat distinct to me, and also empirically they seem to be very different (e.g. this study found the genetic correlation between Neuroticism and BPD to be a mere 0.34 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9001677/ ).
I would study it empirically by giving people a BPD test, except I don't feel convinced psychometric tests of this are valid unless I understand why I should believe them to be valid.
...Or you could let your readers know about the HEXACO:
https://www.tandfonline.com/doi/pdf/10.1111/ajpy.12202
"The review outlines the major reasons for the dominance of the Big Five model, and suggests it is a good descriptive framework for studying personality in general. However we suggest that researchers and practitioners also consider alternative taxonomic personality representations such as the HEXACO. We provide a range of scenarios whereby alternative frameworks will be more appropriate than the Big Five and offer recommendations both for choosing measures in general and for implementing studies examining personality facets."
This was very useful. Thank you.
Fascinating, which version of the big five are you referencing here?