I’ve only skimmed the post, but I couldn’t really understand where it might be related to EA work. You have listed lots of possible uses, but would you mind expanding on one example which you find most promising?
Hi EdoArad, thanks for the question. Apologies for the lengthy response.
I guess there are 2 separate issues that I’m trying to address:
The significance of personality psychology and psychometrics in EA. In particular, the utility in developing and applying a model of temperament (let’s call it model X) which has substantial predictive power, whilst also being practical and adoptable. (Your question highlights that I haven’t properly addressed this – I will try and write something to rectify that and link it here)
Is the OPS model X? And if so, what bottlenecks exist in its development and application that might be worth EAs addressing?
For the OPS the big bottleneck is the fact that its predictive validity has not been formally evaluated. I am, at this stage, suggesting that it may be worth the time of a few EAs to get this bottleneck removed sooner than it otherwise would be. To be clear, I am not suggesting that were the OPS to be proven valid, it would only be useful to the EA community to the extent that EAs and EA organisations would use it. If the model were proven valid it would presumably receive a lot of investment from other individuals and organisations (businesses, research organisations etc.). The potential returns (e.g. research findings) from this investment could provide a big boost for certain EA cause areas.
Regarding a specific cause area. Let’s take mental health. I do not pretend to be an expert, of any sort, in this area. But here is my basic thinking.
For mental health, many of the serious mental disorders that have been identified, such as those in the DSM-5 and ICD-10, are labelled personality disorders. I believe this is because individual differences in predisposition to these conditions is related to individual differences in temperament. Many of those who end up with a personality disorder seem have lifestyles (or trauma) that accentuate a biological predisposition. A model like the OPS could help individuals identify which personality disorders they are prone to. Also, by applying the ideas of the model, more effective interventions could be designed to help treat these conditions and to avoid people getting them in the first place.
This idea extends to other mental disorders as well. It is suggested that the OPS can predict what sort of things an individual is likely to be afraid of and distressed by. In certain cases, it may be able to predict what is causing a person’s depression and/or anxiety. But it is not yet clear the extent to which it might be able to do this. More importantly, it may help identify the right treatment for an individual. Treatment for depression and anxiety is rarely tailored to the individual. But if there are significant differences in how our minds work, then it makes sense that treatments should be ‘psychometrically tailored’. Dave & Shannon suggest they themselves have been using their model to help clients overcome their mental health struggles.
It’s possible that OPS could be useful to EA, but as stated in the post, the validity is not established. It’s hard for me to see how OPS has more predictive ability for mental illness (and subsequent treatment) than any other model of personality. The key feature that makes OPS unique seems to be that it tracks changing personality throughout the day—but what is it about that feature that makes you believe that it could be a better model with more predictive power? Just more granularity?
What are the key first steps that an EA could take? Are you looking for funding? Looking to connect with an established researcher in psychology, or an established institution?
“While they [Dave & Shannon] have taken steps to move to a more scientific approach, some have argued that they fall short of a truly scientific methodology. Nevertheless, that does not make their system invalid.”
This is probably the biggest bottleneck to convince an EA to get involved here. Have Dave & Shannon published peer-reviewed papers that have results that can be replicated? Have they tried to come into contact with established institutions? What if the best next step is for Dave & Shannon to get into graduate school and go for a PhD doing this as their research?
Hi jkmh, thanks for all your questions, it gives me the opportunity to layout my thinking without having to put it in the structure of an essay. I hope you’ll forgive me for answering in bullet points rather than prose.
Are you looking for funding? Looking to connect with an established researcher in psychology, or an established institution?
Awareness: Firstly, I just want to make sure the EA community is aware that the OPS exists (in case I drop dead or something).
Feedback: Secondly, I’m hoping to get some feedback to see if others share my opinion that the OPS shows enough promise to suggest the model may have utility above and beyond existing systems - particularly the Five Factor Model. (based on their own subjective impressions – as that’s all we really have to go on at this point)
Build a team: If the feedback is positive, a next step could be to set up a research team to conduct the necessary research to test the validity of the model – if I can build that team with non-EA people to avoid wasting EA career hours that would probably be ideal.
Funding: Likewise, it would be a last resort to look to the EA community for any funding that would be needed.
Contacts: Contacts would be useful. I lack any connections with any researchers or institutions in the field. I can obviously try to contact these people independently, but if there is mutual contact within the EA community that would probably aid my cause. But again, at this stage, I am just looking for people’s thoughts on the model.
What are the key first steps that an EA could take?
Learn the system: - Learn how the system works. One might find the system useful in their daily life. Of course, in using the model, one might be making some assumptions about the model’s validity. Fair warning: it can take some time before it clicks. There is a limited amount of material to learn from. Though, it is something than you can, to some extent, learn passively – once you understand the basic components of the system you can cross-reference them with observations of yourself and others as you go about daily life. If anyone is interested in learning the system, please reach out to me so I can try and speed up the process (by finding the best resources etc.).
Share their impressions: It would be better if we could have multiple opinions on how much promise the OPS shows from people who have learned the system – rather than just mine.
Get hired: Anyone based around Portland, Oregon can try and get hired. It seems Dave & Shannon are hiring part-time assistants as a trial run for operators. They are only just starting to build the team, so any new hires could have a significant influence on the direction of the organisation. They would probably benefit from having someone who has formal research training as this is a weak spot. Again, if anyone is interested, please reach out to me first so I can advise on what Dave & Shannon are looking for.
Accelerating testing (indirectly): EAs can help accelerate testing the predictive validity of the OPS, indirectly as well as directly, They could, for example, share the model with people who might be interested in exploring the model (e.g. psychology researchers they might know). Offering any relevant expertise would be useful, I don’t currently have any contacts in the field.
Accelerating testing (directly): If an EA wants to get a bit more involved, they can express an interest in joining an interest group. If there is sufficient interest, I could set up an EA personality psychology & psychometrics social media group where updates can be shared.
It’s possible that OPS could be useful to EA, but as stated in the post, the validity is not established. It’s hard for me to see how OPS has more predictive ability for mental illness (and subsequent treatment) than any other model of personality.
Imbalanced: The concept of the savior functions, particularly the dominant function, seems especially important when it comes to mental health. The model suggests that we are all significantly imbalanced, with these savior functions dictating a lot of our inclinations. It is this imbalance that seems to lead to a lot of mental disorders. While it is possible that models like the FFM could still capture this theorised imbalance in a different way, the self-report questionnaires might not pick it up. If this theorised imbalance is true, then we would want to ensure we identify it accurately in an individual before making a prediction about their likelihood of getting certain mental health conditions.
Usability: Predictive utility is more than just about predictive ability / power. It is also about how easy it for users to get at these predictions. The OPS is a model which individuals can learn about themselves and integrate into their everyday decision-making. The modular nature of the model means that individuals can gain a conceptualisation of how each component works and apply this understanding in their reasoning. This is not sort of model that can only be applied in a research setting.
An illustration of the mind: When one is trying to fix something, often the first thing to do is try to identify what specific part isn’t working properly. For therapists, doctors and others trying to tackle mental health issues, this isn’t so easy. There aren’t that many (as fair I’m aware) mental models which they can use to identify the source of a problem. The OPS could provide such a model, it illustrates parts of the mind which seem to be the source of a lot of mental disorders (particularly personality models). The model is also simple that mental health professionals can use it to communicate their ideas to their patients.
Tailoring new interventions: Ideally, new treatments and interventions will be trialled in advance. When conducting these trials, it seems like it would be useful to use personality as a control (to see if the intervention would have a different impact based on personality). This would involve putting each member of the population sample ‘in a box’ based on their personality. As fair as I’m aware (though I don’t claim to have done thorough research) this doesn’t normally happen. The continuous format of the FFM isn’t exactly conducive to this sort of thing. Moreover, new treatments could also be designed to target specific temperament types, which would presumably improve their effectiveness. I might be wrong but It doesn’t seem like current mental interventions are generally designed to be tailored to individuals based on their personality. The typological format of the model would make it easy to be used as a control and for interventions to be targeted at specific groups (based on their temperament).
The key feature that makes OPS unique seems to be that it tracks changing personality throughout the day—but what is it about that feature that makes you believe that it could be a better model with more predictive power? Just more granularity?
Mental limits: The concept of the animals, which illustrates the different mental states which one can move in and out of throughout the day, appears to dictate the mental capacity one has for doing different activities. In the same way temperament dictates how much mental capacity an individual has for social interactions (a well-known manifestation of extroversion), temperament, it seems, will also dictate how much mental capacity a person has for a range of activities. Trialling new things and taking in new information, for example. (one reason why some people read a lot more than others). This would be highly relevant to career planning.
Strategy: Another way in which the animal cycle appears to manifest, is in the way people approach a goal. It seems, individuals tend to take similar approaches to tackling different problems (even if the problems require different approaches). It is suggested that the approach individuals are inclined to take (E.g. does the individual tend to plan or do lots of research before getting started), is in large part a reflection of the order in which they cycle through the animals. These could be useful insights, particularly in the context of improving decision making.
Uniqueness: I think there are other important points that make the OPS distinct. For example, the interconnected nature of the model which shows how the different facets of personality interact with each other. I think this would probably give it an extra boost in predictive power, relative to models like the FFM which look at traits separately.
Practicality and adoptability – Predictive power isn’t the only thing that will dictate the real-world utility of a model. A type-based model is far more practical (and engaging) than a continuous model that looks at separate traits like the FFM. The modular framework (taxonomic structure) would also make it very practical, allowing the model to be used to divide a population at different levels (i.e. 2 types, 4 types, 8 types etc.).
This is probably the biggest bottleneck to convince an EA to get involved here. Have Dave & Shannon published peer-reviewed papers that have results that can be replicated?
Still building the samples: Dave & Shannon have not yet tested the model. They still need to build the population samples. The problem with testing a model of personality is that you have to build a population sample to conduct the test– which involves building an assessment. Therefore, you cannot evaluate the model without simultaneously evaluating the assessment. But there is no way of knowing that your assessment is working because you do not yet know whether the model is even valid. Dave & Shannon have therefore stuck with the double-blind observational test, with trained operators, to try to ensure they type people as best they can.
Focusing on evaluation: Regarding EAs, at this point I’m focusing on evaluating the model. I mentioned a few cause areas where this model might be of use. But I’m not expecting EAs working in these areas to apply the model until its validity has been tested. So, I’m asking whether there is any EAs who want to help accelerate the evaluation of the model. I realise this is more of a high-risk career move, where direct impact is less guaranteed.
Have they tried to come into contact with established institutions? What if the best next step is for Dave & Shannon to get into graduate school and go for a PhD doing this as their research?
Keeping their heads down—No, seem focused on building their own team first before making next move. They seem to want to wait until they have their own evidence base before sharing their findings.
PhD unlikely: I’m pretty sure neither of them have undergraduate degrees. So they might have a bit of a hard time getting onto a PhD programme. I think having their own independent research team give them more flexibility (and less time constraints).
If you’re still reading, I hope this answers your questions.
I’ve only skimmed the post, but I couldn’t really understand where it might be related to EA work. You have listed lots of possible uses, but would you mind expanding on one example which you find most promising?
Hi EdoArad, thanks for the question. Apologies for the lengthy response.
I guess there are 2 separate issues that I’m trying to address:
The significance of personality psychology and psychometrics in EA. In particular, the utility in developing and applying a model of temperament (let’s call it model X) which has substantial predictive power, whilst also being practical and adoptable. (Your question highlights that I haven’t properly addressed this – I will try and write something to rectify that and link it here)
Is the OPS model X? And if so, what bottlenecks exist in its development and application that might be worth EAs addressing?
For the OPS the big bottleneck is the fact that its predictive validity has not been formally evaluated. I am, at this stage, suggesting that it may be worth the time of a few EAs to get this bottleneck removed sooner than it otherwise would be. To be clear, I am not suggesting that were the OPS to be proven valid, it would only be useful to the EA community to the extent that EAs and EA organisations would use it. If the model were proven valid it would presumably receive a lot of investment from other individuals and organisations (businesses, research organisations etc.). The potential returns (e.g. research findings) from this investment could provide a big boost for certain EA cause areas.
Regarding a specific cause area. Let’s take mental health. I do not pretend to be an expert, of any sort, in this area. But here is my basic thinking.
For mental health, many of the serious mental disorders that have been identified, such as those in the DSM-5 and ICD-10, are labelled personality disorders. I believe this is because individual differences in predisposition to these conditions is related to individual differences in temperament. Many of those who end up with a personality disorder seem have lifestyles (or trauma) that accentuate a biological predisposition. A model like the OPS could help individuals identify which personality disorders they are prone to. Also, by applying the ideas of the model, more effective interventions could be designed to help treat these conditions and to avoid people getting them in the first place.
This idea extends to other mental disorders as well. It is suggested that the OPS can predict what sort of things an individual is likely to be afraid of and distressed by. In certain cases, it may be able to predict what is causing a person’s depression and/or anxiety. But it is not yet clear the extent to which it might be able to do this. More importantly, it may help identify the right treatment for an individual. Treatment for depression and anxiety is rarely tailored to the individual. But if there are significant differences in how our minds work, then it makes sense that treatments should be ‘psychometrically tailored’. Dave & Shannon suggest they themselves have been using their model to help clients overcome their mental health struggles.
I hope this makes things slightly clearer.
It’s possible that OPS could be useful to EA, but as stated in the post, the validity is not established. It’s hard for me to see how OPS has more predictive ability for mental illness (and subsequent treatment) than any other model of personality. The key feature that makes OPS unique seems to be that it tracks changing personality throughout the day—but what is it about that feature that makes you believe that it could be a better model with more predictive power? Just more granularity?
What are the key first steps that an EA could take? Are you looking for funding? Looking to connect with an established researcher in psychology, or an established institution?
“While they [Dave & Shannon] have taken steps to move to a more scientific approach, some have argued that they fall short of a truly scientific methodology. Nevertheless, that does not make their system invalid.”
This is probably the biggest bottleneck to convince an EA to get involved here. Have Dave & Shannon published peer-reviewed papers that have results that can be replicated? Have they tried to come into contact with established institutions? What if the best next step is for Dave & Shannon to get into graduate school and go for a PhD doing this as their research?
Hi jkmh, thanks for all your questions, it gives me the opportunity to layout my thinking without having to put it in the structure of an essay. I hope you’ll forgive me for answering in bullet points rather than prose.
Awareness: Firstly, I just want to make sure the EA community is aware that the OPS exists (in case I drop dead or something).
Feedback: Secondly, I’m hoping to get some feedback to see if others share my opinion that the OPS shows enough promise to suggest the model may have utility above and beyond existing systems - particularly the Five Factor Model. (based on their own subjective impressions – as that’s all we really have to go on at this point)
Build a team: If the feedback is positive, a next step could be to set up a research team to conduct the necessary research to test the validity of the model – if I can build that team with non-EA people to avoid wasting EA career hours that would probably be ideal.
Funding: Likewise, it would be a last resort to look to the EA community for any funding that would be needed.
Contacts: Contacts would be useful. I lack any connections with any researchers or institutions in the field. I can obviously try to contact these people independently, but if there is mutual contact within the EA community that would probably aid my cause. But again, at this stage, I am just looking for people’s thoughts on the model.
Learn the system: - Learn how the system works. One might find the system useful in their daily life. Of course, in using the model, one might be making some assumptions about the model’s validity. Fair warning: it can take some time before it clicks. There is a limited amount of material to learn from. Though, it is something than you can, to some extent, learn passively – once you understand the basic components of the system you can cross-reference them with observations of yourself and others as you go about daily life. If anyone is interested in learning the system, please reach out to me so I can try and speed up the process (by finding the best resources etc.).
Share their impressions: It would be better if we could have multiple opinions on how much promise the OPS shows from people who have learned the system – rather than just mine.
Get hired: Anyone based around Portland, Oregon can try and get hired. It seems Dave & Shannon are hiring part-time assistants as a trial run for operators. They are only just starting to build the team, so any new hires could have a significant influence on the direction of the organisation. They would probably benefit from having someone who has formal research training as this is a weak spot. Again, if anyone is interested, please reach out to me first so I can advise on what Dave & Shannon are looking for.
Accelerating testing (indirectly): EAs can help accelerate testing the predictive validity of the OPS, indirectly as well as directly, They could, for example, share the model with people who might be interested in exploring the model (e.g. psychology researchers they might know). Offering any relevant expertise would be useful, I don’t currently have any contacts in the field.
Accelerating testing (directly): If an EA wants to get a bit more involved, they can express an interest in joining an interest group. If there is sufficient interest, I could set up an EA personality psychology & psychometrics social media group where updates can be shared.
Imbalanced: The concept of the savior functions, particularly the dominant function, seems especially important when it comes to mental health. The model suggests that we are all significantly imbalanced, with these savior functions dictating a lot of our inclinations. It is this imbalance that seems to lead to a lot of mental disorders. While it is possible that models like the FFM could still capture this theorised imbalance in a different way, the self-report questionnaires might not pick it up. If this theorised imbalance is true, then we would want to ensure we identify it accurately in an individual before making a prediction about their likelihood of getting certain mental health conditions.
Usability: Predictive utility is more than just about predictive ability / power. It is also about how easy it for users to get at these predictions. The OPS is a model which individuals can learn about themselves and integrate into their everyday decision-making. The modular nature of the model means that individuals can gain a conceptualisation of how each component works and apply this understanding in their reasoning. This is not sort of model that can only be applied in a research setting.
An illustration of the mind: When one is trying to fix something, often the first thing to do is try to identify what specific part isn’t working properly. For therapists, doctors and others trying to tackle mental health issues, this isn’t so easy. There aren’t that many (as fair I’m aware) mental models which they can use to identify the source of a problem. The OPS could provide such a model, it illustrates parts of the mind which seem to be the source of a lot of mental disorders (particularly personality models). The model is also simple that mental health professionals can use it to communicate their ideas to their patients.
Tailoring new interventions: Ideally, new treatments and interventions will be trialled in advance. When conducting these trials, it seems like it would be useful to use personality as a control (to see if the intervention would have a different impact based on personality). This would involve putting each member of the population sample ‘in a box’ based on their personality. As fair as I’m aware (though I don’t claim to have done thorough research) this doesn’t normally happen. The continuous format of the FFM isn’t exactly conducive to this sort of thing. Moreover, new treatments could also be designed to target specific temperament types, which would presumably improve their effectiveness. I might be wrong but It doesn’t seem like current mental interventions are generally designed to be tailored to individuals based on their personality. The typological format of the model would make it easy to be used as a control and for interventions to be targeted at specific groups (based on their temperament).
Mental limits: The concept of the animals, which illustrates the different mental states which one can move in and out of throughout the day, appears to dictate the mental capacity one has for doing different activities. In the same way temperament dictates how much mental capacity an individual has for social interactions (a well-known manifestation of extroversion), temperament, it seems, will also dictate how much mental capacity a person has for a range of activities. Trialling new things and taking in new information, for example. (one reason why some people read a lot more than others). This would be highly relevant to career planning.
Strategy: Another way in which the animal cycle appears to manifest, is in the way people approach a goal. It seems, individuals tend to take similar approaches to tackling different problems (even if the problems require different approaches). It is suggested that the approach individuals are inclined to take (E.g. does the individual tend to plan or do lots of research before getting started), is in large part a reflection of the order in which they cycle through the animals. These could be useful insights, particularly in the context of improving decision making.
Uniqueness: I think there are other important points that make the OPS distinct. For example, the interconnected nature of the model which shows how the different facets of personality interact with each other. I think this would probably give it an extra boost in predictive power, relative to models like the FFM which look at traits separately.
Practicality and adoptability – Predictive power isn’t the only thing that will dictate the real-world utility of a model. A type-based model is far more practical (and engaging) than a continuous model that looks at separate traits like the FFM. The modular framework (taxonomic structure) would also make it very practical, allowing the model to be used to divide a population at different levels (i.e. 2 types, 4 types, 8 types etc.).
Still building the samples: Dave & Shannon have not yet tested the model. They still need to build the population samples. The problem with testing a model of personality is that you have to build a population sample to conduct the test– which involves building an assessment. Therefore, you cannot evaluate the model without simultaneously evaluating the assessment. But there is no way of knowing that your assessment is working because you do not yet know whether the model is even valid. Dave & Shannon have therefore stuck with the double-blind observational test, with trained operators, to try to ensure they type people as best they can.
Focusing on evaluation: Regarding EAs, at this point I’m focusing on evaluating the model. I mentioned a few cause areas where this model might be of use. But I’m not expecting EAs working in these areas to apply the model until its validity has been tested. So, I’m asking whether there is any EAs who want to help accelerate the evaluation of the model. I realise this is more of a high-risk career move, where direct impact is less guaranteed.
Keeping their heads down—No, seem focused on building their own team first before making next move. They seem to want to wait until they have their own evidence base before sharing their findings.
PhD unlikely: I’m pretty sure neither of them have undergraduate degrees. So they might have a bit of a hard time getting onto a PhD programme. I think having their own independent research team give them more flexibility (and less time constraints).
If you’re still reading, I hope this answers your questions.