Clearer Thinking's Study: How much is your anxiety level driven by your traits, and how much is it driven by your life circumstances?

How much is your anxiety level driven by your traits, and how much is it driven by your life circumstances?  

We ran a study to try to find out.

It seems obvious that dealing with a stressful life circumstance, like financial strain, or managing a chronic illness, might make a person more anxious than they would be otherwise. But it isn’t obvious how large of an impact this might have, what kinds of stressful experiences matter for increasing anxiety, or how long those effects might last. How does anxiety relate to stressful life circumstances compared to personality traits or other factors?

What we found was somewhat surprising. In our study, acutely stressful life events (like job loss) had less of a relationship with people’s anxiety levels than we expected. In contrast, we found that chronic stressors (like social isolation) accounted for a much greater share of the total anxiety people experience. 

We also found that the neuroticism Big Five personality dimension is a stronger predictor of current anxiety levels than either acute or chronic stressors, so although certain stressful circumstances do impact anxiety, people’s traits typically seem to more strongly predict their anxiety levels.

If you are struggling with anxiety, determining how much your circumstances are contributing to the problem can be helpful. This article offers some insights into figuring that out, as well as some guidance for using that information to manage your anxiety more effectively.

What is anxiety?

Before we get further into the results of the study, it is important to understand what anxiety is. An anxiety response is the brain preparing to react to something it has identified as a potential threat. It is similar to a fear response, although typically less intense. 

A fear response involves the brain reacting to a source of imminent danger (such as a predator) by ramping up adrenaline and triggering a fight-flight-freeze response. An anxiety response, by contrast, involves the brain responding to a threat that could happen in the future, rather than something immediate. Despite that key difference, an anxiety response is similar to a fear response because in both cases the brain triggers processes to prepare the body to respond to the threat. That is why the experience of being in an anxious state induces physiological changes that may include symptoms like a rapid heartbeat, sweating, and chest tightness.

We can think of anxiety as a system in the brain for detecting and preparing to respond to potential threats, which is a useful system to have! Anxiety is useful when it alerts us to real risks in ways that prompt us to take useful action. But it can become miscalibrated, and that’s when normal anxiety can change into an anxiety disorder. 

If the threat detection system is oversensitive to threats or responds disproportionately to them, the symptoms of the anxiety response itself can become very unpleasant and can get in the way of taking useful action to address the potential threats that the system is identifying. Think of it like a smoke detector - it’s really important to have a smoke detector to warn you in the presence of a real fire, but if your smoke detector goes off every time you cook, those false alarms make it unhelpful in your day-to-day life.

While the threat that triggers a fear response is usually pretty obvious, the trigger for an anxiety response - the possible future threat that has been identified by the brain - can often be unclear.

The brain identifies potential threats by using what a person has observed and experienced in their life as data. The threat detection process operates subconsciously and much more quickly than a reasoning process. 

Often a person rationally knows that a risk is minor, but their anxiety response may treat the situation as a potential major threat. For example, a person may feel anxious about asking a friend for a favor, despite rationally understanding that being rejected in this instance wouldn’t be a big deal and that making the ask is unlikely to harm their relationship with their friend. In this case, quick, subconscious processes have pattern matched this situation as similar to previously experienced or observed situations where something bad happened or where there was perceived danger, so the situation was flagged as potentially risky. 

It can be difficult to retrain the brain to realize that a situation is not threatening once it has been categorized as threatening because it takes many instances of something going smoothly to dull the salience of a similar past experience that was bad or dangerous. And it doesn't even require something really bad happening for your brain to associate a pattern with danger: the mere perception of possible danger can be enough. For instance, someone who becomes afraid of an overly friendly dog may have perceived a risk that the dog would attack them, even though it just wanted to play, and this experience may lead to anxiety in the future when dogs are nearby. 

How does anxiety interact with life circumstances?

That threat detection system is always operating, attempting to alert you to potential danger. How many alerts you receive from that system will depend on a number of factors. One important factor is how many potential threats there are in the environment you’re currently experiencing. A high-stress environment is rich in threats, which means that anxiety responses will be triggered more frequently than if the current environment is low-stress, regardless of the individual’s unique traits and past experiences. For example, two individuals living in a warzone are both likely to experience high levels of anxiety, regardless of their individual differences. This seems obvious; however, in situations that are less extreme, it’s often difficult to objectively assess the threat level of the environment. 

People who are experiencing a lot of anxiety due to internal factors tend to attribute their anxiety to things going on in their lives and tend to report their life circumstances as being stressful. For example, if two people with different jobs both report that their jobs are extremely stressful, it is difficult to objectively compare the level of external stress those two people are under just by looking at their self-reports about their level of job stress. This is challenging for researchers and clinicians, as well as for individuals trying to make accurate inferences about their own anxiety and circumstances.

Being able to accurately distinguish the role of the environment is important because it shapes which interventions are likely to be most effective. A person who is experiencing high levels of anxiety due to life stressors may want their anxiety levels to go down because their anxiety is getting in the way of them coping with their problems, but their anxiety itself isn’t their primary problem. The most effective strategies for reducing anxiety for this person may be addressing the factors that are causing their stress in the first place (so long as they have some control over those factors). For example, If a person is experiencing a lot of anxiety due to financial strain, their anxiety is likely to improve if they find a better-paying job. The most effective intervention to help this person might be assistance in finding a better job rather than teaching them anxiety management techniques. If, instead, the person is experiencing a lot of anxiety, which they attribute to financial strain, but that anxiety doesn’t diminish when their financial situation dramatically improves, it seems likely that there are internal factors in their threat detection system that are leading them to have levels of anxiety that are disproportionate to their situation. For this person, help with techniques for managing anxiety, like cognitive behavioral therapy (where they may learn to challenge unhelpful thoughts) or mindfulness techniques (where they learn to observe their thoughts less judgmentally), may be more useful interventions.

Our study

In order to understand the relationship between stressful life circumstances and anxiety, Spencer Greenberg, Gregory Lopez, and I developed a questionnaire and asked 419 people whether they experienced a list of chronic and acute stressors and when they experienced them (the exact wording for these chronic and acute stressor questions is in the appendix).

Chronic stressors are ongoing stressful circumstances that persist in time, like being a caregiver for a loved one with a severe illness, being unemployed for months, or experiencing lasting social isolation.

Acute stressors are discrete events that happen at one point in time, like an unexpected death in the family, a serious accident, or being fired from a job.

We developed our list of questions by looking at research that identified the most common high-stress events that people experience. The questions explored specific objective events rather than asking people to rate subjectively how stressful their lives were because that allowed us to disentangle a person’s internal states from their external experiences. 

For the chronic stressors, we asked people to respond with how recently they experienced the stressor, if ever. The chronic stressors we asked about were: 

• Abusive relationship

• Caregiving

• Chronic illness or disability

• Chronic severe illness of a loved one

• Discrimination

• Estrangement from family

• Financial strain

• Incarceration (note that no currently incarcerated people were included in the study, so the incarceration chronic stressor could not be a currently ongoing one in our sample)

• Parenting young children

• Social isolation

• Unemployment

• Unsafe environment

• Work strain

To assess acute stressors, we asked people how frequently they experienced certain events in the last 5 years. For those who had experienced the event at least once, we asked how recently they last experienced it. The events that we asked about were:

• Death of a loved one

• Job loss (not including planned departures)

• Severe accident or medical emergency (requiring going to the hospital)

• Criminal violence (e.g., being mugged, threatened with a weapon, or assaulted)

• Divorce or breakup from a cohabiting partner

• Housing loss due to eviction, foreclosure, fire, or other emergency

• Witnessing sudden death, violence, or life-threatening incident

• Disconnection from a close loved one (cutting off contact)

• Arrest

• Disaster evacuation

Participants also responded to short inventories assessing their current level of anxiety (GAD-7) and answered questions assessing their level of the personality trait neuroticism. The GAD-7 focuses on people’s experiences over the last two weeks, and is used as a measure of a person’s current level of anxiety. The trait neuroticism questions, by contrast, ask about more lasting personality traits that people ascribe to themselves related to anxiety, for example, asking whether people consider themselves to be someone who is easily frightened, or think of themselves as a worrier. This is a measurement of the neuroticism dimension of the Big Five personality model.

The GAD-7 anxiety measure asks people to think about their last two weeks and indicate how frequently they have experienced the following things:

• Feeling nervous, anxious, or on edge

• Not being able to stop or control worrying

• Worrying too much about different things

• Trouble relaxing

• Being so restless that it is hard to sit still

• Becoming easily annoyed or irritable

• Feeling afraid, as if something awful might happen

For each question the answer options are “not at all,” “several days,” “more than half the days,” “nearly every day.”

Our measure of the personality trait neuroticism comes from measures of the “Big 5” personality traits. Participants respond to statements about themselves, choosing from 7 answer options ranging from “strongly disagree” to “strongly agree” how well the statement describes them. The statements we used were:

• I am not a worrier. [reverse scored]

• I am easily frightened.

• I have fewer fears than most people. [reverse scored]

As you can see, these questions focus on a person’s general sense of how anxious or fearful of a person they are, while the GAD-7 questions focus on specific anxiety symptoms experienced in the last two weeks.

What we found

The main results from our study were that chronic stressors tend to matter much more in terms of increasing people’s total anxiety level than acute stress events. We suspect that this difference might be due to ongoing chronic stressors affecting people differently than discrete acute stressor events. While acutely stressful events show a much weaker lasting relationship with anxiety over months or years, our study wouldn’t capture the kind of immediate anxiety response that an acutely stressful event might trigger in the moment and immediate aftermath of the event. What we can say is that in the months or years after an acutely stressful event, people’s anxiety levels are typically not heavily influenced by that event. Being under a state of persistent stress from an ongoing chronic stressor is more strongly related to increases in anxiety over the time periods measured in our study, but that relationship also weakens as more time passes after the chronic stressor has ended.

In our study we found that the effects of stressors diminish over time. A few years after a chronic stressor is over, the anxiety levels reported by someone who experienced the stressor aren’t any different from someone who didn’t experience it. This speaks to the natural resilience that people have. Although difficult circumstances do have an impact on our anxiety, after they are over we typically recover and return to our baselines over time.

One hypothesis we had that we didn’t find support for was that people who are higher in the neuroticism trait would respond more strongly to chronic stressors, and show more of an increase in anxiety in response to them than people who were low in neuroticism. We tested this by looking at the interaction between neuroticism and chronic stressors in our regression model, but we did not find a significant result (p = 0.06). It’s possible that this interaction does exist and we didn’t find it with our sample, but if the effect does exist it is likely to be very small.

Although we do have some preliminary results showing that certain specific stressors influence anxiety more than others, the very small number of people who had experienced certain stressors recently in the sample means that we have to interpret those results cautiously. If we had a larger dataset we would be able to determine with more confidence whether specific stressors matter more than others. 

Because the incidence of any specific stressor was rare in our sample, we added up the number of different acute stressors and the number of different chronic stressors that each person reported. In early analyses, we tried several different ways of analyzing the stressors, and found that this simple approach worked best for capturing the impacts of stressors. We also examined a few different ways to divide recent from non-recent events, since there were fewer recent events and more older events.

The graph below shows what we found when we looked at a subset of our data including people who reported one or zero chronic stressors from any of the chronic stressor types. We tried looking at the recency of stressors in a number of different ways, and determined that the most informative one was to compare recent chronic stressors (meaning those currently ongoing or that had ended within the last 3 months) and older chronic stressors, meaning stressors that occurred more than 3 months ago but less than 5 years ago).  Looking at the graph below you can see that someone who experienced a chronic stressor more than 5 years ago has about the same GAD-7 score measuring their anxiety level over the last two weeks as someone who experienced no chronic stressors at all.  You can also see that a person with a recent chronic stressor reports an anxiety level about twice as high as people with no stressor or a stressor more than 5 years ago. The people who experienced a chronic stressor more than three months ago but within the last 5 years report a GAD-7 score about halfway between those other two groups.  

Based on this analysis we consider recent chronic stressors to be those that are currently ongoing or that occurred within the last three months, and older chronic stressors to be more than three months in the past but within the last five years. 

We did the same analysis looking at the acute stress events:

In this graph you can see that experiencing one acute stress event within the last five years but more than two years ago corresponds to a GAD-7 score that is about the same as having not experienced any acute stress events. Based on this analysis we decided to exclude acute stressor events that were more than two years in the past from our model. For those experiencing one acute stress event within the last month to within the last two years, their GAD-7 score is higher, but it’s not really possible to distinguish between the more recent and the older events within this time range from this analysis. We conducted subsequent analyses that led us to consider acute events within the last year to be recent, and those that were more than a year ago but within the last two years to be older.

Our overall model was a linear regression that looked at the relationship between neuroticism, recent chronic stressors, older chronic stressors, recent acute stressors and older acute stressors as independent variables and the GAD-7 anxiety measure as the dependent variable. Here are the results of that model:

Table 1: Acute and chronic stressors and neuroticism in a linear regression model predicting GAD-7 (R2 = 0.44) 

Variable	Coefficient (non-normalized)	p-value
pureNeuroticismSumScore	0.58	<0.00001
acuteRecentSum	0.21	0.3
acuteOlderSum	0.53	0.045
chronicRecentSum	0.75	0.0000232
chronicOlderSum	0.44	0.00651

Looking at these results you can see that acute recent stressors don’t have a significant relationship with GAD-7, and the relationship between acute older stressors and GAD-7 is on the border between statistical significance and nonsignificance. The relationships between GAD-7 and both chronic stressor variables as well as neuroticism are clearly significant.

Looking at the bar graph above it may seem surprising that acute stressors are not statistically significant in this regression model. The difference between the graph above and this model is that this model reports the amount of the variance in the dependent variable that can be specifically predicted from all independent variables simultaneously, while taking all of the other independent variables into account. That means that this model shows that knowing about a person’s recent or older acute stressor experiences isn’t adding extra ability to predict current anxiety levels beyond what is explained by neuroticism and recent or older chronic stressors.

That result led us to simplify our model to only include chronic stressors and neuroticism.

Table 2: Simplified Model - chronic stressors and neuroticism in a linear regression model predicting GAD-7 (R2 = 0.43) 

Variable	Coefficient (non-normalized)	Coefficient (normalized)	p-value
pureNeuroticismSumScore	0.58	0.52	<0.00001
chronicRecentSum	0.83	0.20	<0.00001
chronicOlderSum	0.59	0.18	<0.00001

Overall we learned a few important things from these analyses:

• Acute stressors in the past months or years are not useful for predicting current anxiety, but chronic stressors do predict increased anxiety. Even very serious acute events don’t seem to have much impact on a person’s anxiety levels in the months after the event has happened, when you take into account a person’s level of neuroticism and the chronic stressors they have experienced. Chronic stressors, on the other hand, do predict a person reporting a higher level of current anxiety. 

• Recent chronic stressors have a somewhat stronger relationship to anxiety than older chronic stressors, but both do matter for predicting current anxiety. Looking at the non-normalized results of the regression, you can see that one additional recent chronic stressor predicts an increase of .83 in GAD-7 score. One additional older chronic stressor predicts an increase of .59 in GAD7 score. The GAD-7 has a scoring range from 0 to 21. A score of 5-9 is interpreted as mild anxiety, 10-14 is moderate, and 15 or higher is severe. An increase of over half of a point in the context of these scoring ranges is a substantive increase.

• Neuroticism predicts a little bit more about someone’s current anxiety level than the chronic stressors they have experienced. By looking at the normalized coefficients, you can compare the impact of neuroticism on GAD-7 with the impact of chronic stressors on GAD-7. The neuroticism coefficient of .52 is larger than even the sum of the chronic stressor variable coefficients (.38). The R2 value (the total amount of the variance in anxiety explained by the model as a whole) is .43, which means that this model explains about 43% of the variation in GAD-7 scores in our sample of participants. If you remove neuroticism from the model the R2 value drops to .18,  meaning that the stressor variables alone explain about 18% of the variance in anxiety while neuroticism alone explains about 25% of the variance in anxiety. There is also 57% of the variation in GAD-7 score that is not explained by our model. That remaining variation is due to other factors impacting anxiety that we haven’t included, as well as noise. Explaining 43% of the variance in the dependent variable with only 3 independent variables means this model performs fairly well, given how simple it is.

What about specific stressors?

Among the specific chronic stressors that had a reasonably large number of occurrences in the study sample, those with the largest relationships with GAD-7 were: 

• work strain

• disability or chronic illness

• financial strain

• social isolation

We don’t have enough evidence to suggest that the other stressors aren’t important as well due to small counts in some of the categories, but we can suggest that these four are stressors to watch out for in terms of their relationship to anxiety.

Applying this research to your own life

If you are experiencing a high level of anxiety, and you’re trying to understand whether your situation is to blame or whether you could use help managing anxiety responses that may be a bit disproportionate to your circumstances, there are a few things to learn from this research.

Look specifically at the ongoing and recent stressful situations in your life. Especially if you are under severe financial or work strain, managing a chronic illness or disability, or experiencing social isolation, those factors could be having a substantial impact on your current anxiety level. Seeking help with managing those particular stressors (and working to improve them if they involve factors you can control) may be the most important step to managing the associated anxiety symptoms you may be experiencing. 

With currently ongoing chronic stressors, it’s useful to think about whether the situation you are dealing with is something that you can change or something that you’re going to need to continue managing for the foreseeable future. 

If the chronic stressor is a situation you can change, that may be the most effective approach to managing the anxiety that is coming from that stressor. For example, if your work environment has become especially toxic, looking for a new job or finding a way to change who you work with at your current job may be a good approach.

If you are managing a chronic stressor that you are unlikely to be able to change and that you are likely to be continuing to deal with for a long time, recognize that these ongoing stressors take a toll. It may be helpful to seek support for both dealing with the chronic stressor and managing anxiety symptoms that it generates. Even in objectively difficult situations, sometimes psychological treatments can help you cope better with the situation.

If the chronic stressors in your life are further in the past and you’re still experiencing a high level of anxiety, or if you find that you tend to feel anxious even when external events in your life are going well, consider tools and treatments that address the anxiety itself.

In any of these cases, if you’re experiencing a level of anxiety that is unpleasant or is getting in the way of things you need to or want to do, it’s useful to know that anxiety itself is not dangerous to you. Some of the symptoms of anxiety, like a racing heartbeat, can feel scary themselves, which can lead to feeling even more anxious. Remembering that anxiety is coming from your threat detection system trying to warn you about something, but that your threat detection system may be miscalibrated, can be helpful for putting anxiety symptoms in context. Seeking help from a therapist or self-help resources may be useful regardless of the underlying sources of your anxiety, but understanding whether the source is likely to be situational may guide what you choose to focus on when working with either a therapist or self-help resource, or even when you're working through your challenges without formal help.

Appendix: Chronic Stressors Survey Questions

Chronic Stressor Type	Our Question: Over the course of your life have you ever...
Abusive relationship	…lived with a partner that was abusive to you, or had a high-conflict relationship with a partner you lived with involving frequent shouting matches, throwing or breaking things, or physical fighting?
Caregiving	...acted as a primary caregiver for a loved one who was disabled or chronically ill and needed your help with basic self-care activities on a daily or near-daily basis?
Financial strain	...experienced financial strain severe enough that you were unable to pay your regular bills for the basic necessities of housing, food, medicine, and utilities for multiple months in a row, or had to file for bankruptcy?
Chronic illness or disability	...had a chronic illness or disability severe enough that you had to stop working for a month or more?
Unemployment	...been unemployed for 3 months or more when you were able to work and were seeking a job?
Unsafe environment	...lived in a neighborhood with a high rate of crime and violence that resulted in you taking precautions to protect yourself or your home that you wouldn't have otherwise taken?
Social isolation	...spent at least 3 months where you didn't have anyone in your life who you could ask to help you if you needed help with something like an errand or help with a task or other favor?
Estrangement from family	...been estranged from at least one member of your immediate family such as a parent, child, or sibling or have an extremely conflictual relationship with an immediate family member involving frequent heated arguments and/or periods of disconnection lasting a month or more?
Discrimination	...experienced regular discrimination based on people's perceptions of your race, gender, nationality, religion, sexual orientation, or other identity category to which you belong or are perceived to belong?
Parenting young children	...been responsible for at least half of the caregiving for at least one child under the age of 5?
Chronic severe illness of a loved one	...had a close loved one, such as a parent, child, sibling, partner or close friend, going through a chronic life-threatening illness? For example: having a sibling with cancer.
Incarceration	...spent at least one month in jail or prison?
Work strain	...worked for at least three months in an environment with hostile coworkers, personality conflicts with your supervisor, deadline pressures that meant often working outside of regular hours, or exacting standards on high stakes projects that you felt that you couldn't meet?

For each of these questions the answer options were: 

• Currently ongoing

• Not currently, but in the last 3 months

• More than 3 months ago, but in the last 12 months

• More than a year ago, but within the last 5 years

• More than 5 years ago

• Never

Acute Stressor Events Survey Questions

Acute Stressor Event Type	Consider the events of the last 5 years of your life.
Death of a loved one	Has a close loved one to you died? For example, a parent, sibling, spouse or partner, child or close friend.
Job loss (not including planned departures)	Have you experienced the sudden loss of a non-temporary job? For example: been fired, quit or walked off the job without a planned departure, or were let go or laid off on short notice.
Severe accident or medical emergency (requiring going to the hospital)	Have you had an accident or medical emergency serious enough that it required going to the hospital? For example: a car accident, heart attack, or stroke.
Criminal violence (e.g., being mugged, threatened with a weapon, or assaulted)	Have you experienced criminal violence? For example: being threatened with a weapon, mugged, physically attacked, or a break-in of your home while you were present.
Divorce or breakup from a cohabiting partner	Have you had a break up or divorce from a partner you were living with that resulted in at least one of you moving out?
Housing loss due to eviction, foreclosure, fire, or other emergency	Have you lost your housing due to eviction, foreclosure, fire, or other emergency?
Witnessing sudden death, violence, or life-threatening incident	Were you present for or have you witnessed the immediate aftermath of the sudden death or life-threatening incident of another person? For example: witness to a murder, suicide, violent beating, grevious accident, stroke, or enountering a victim at the scene shortly after an accident, suicide or violent attack.
Disconnection from a close loved one (cutting off contact)	Have you had conflict with a person close to you that was serious enough that it resulted in you no longer communicating with that person? For example: disconnection from a close friend, child, sibling, partner or parent.
Arrest	Have you been arrested?
Disaster evacuation	Have you had to evacuate from where you lived due to natural or human-caused disaster? For example: fire, flood, severe storm, chemical spill, or critical infrastructure failure.

For each acute stressor event everyone was given the question above with the following answer options asking about the frequency of this event in the last 5 years:

• Not in the last 5 years

• Once in the last 5 years

• Twice in the last 5 years

• Three or more times in the last 5 years

If participant indicated that the event had happened at least once in the last 5 years they were given a follow up question asking how long ago the most recent occurrence of this event had been, and they were given the following answer options:

• Within the last week

• More than a week ago, but within the last month

• More than a month ago, but within the last year

• More than a year ago, but within the last 2 years

• More than 2 years ago, but within the last 5 years

If the participant answered “not in the last 5 years” to the first question they didn’t receive the follow up question for that event type, and the answer to the follow up question for that event type was automatically set to "not within the last 5 years.”