One framework that explains variation in both ill-being (such as stress and depression) and wellbeing is the Basic Psychological Needs Theory of Self-determination Theory (SDT [10]). People have three basic psychological needs that must be satisfied to experience growth and wellbeing. Autonomy is the perception that any behavior in which an individual engages in is a result of their choice to behave that way, rather than a result of external factors and feelings of obligation. Competence is the perception of feeling effective in a certain context. Finally, relatedness is the perception of being listened to and cared for by others.

Satisfaction of these three basic psychological needs positively predicts hedonic wellbeing, the state of feeling good and being satisfied with life [11], and eudaimonic wellbeing, which involves personal growth, meaning, and purpose [12]. This relationship has been demonstrated with people of various ages, gender [12], and cultures [13]. It has also been demonstrated across numerous contexts, including the COVID-19 pandemic; individuals who reported that their autonomy, competence, and relatedness satisfaction increased during lockdowns experienced greater wellbeing compared to those who did not experience increases in their need satisfaction [14].

Individuals can also experience need frustration, which occurs when individuals feel prevented from feeling satisfied in their basic psychological needs. Researchers have found that need satisfaction in general positively predicts wellbeing (but need frustration was unrelated), and that need frustration positively predicts ill-being, such as depressive symptomology (but need satisfaction was unrelated) [13, 15]. One of the only papers that have investigated the unique influence of autonomy, competence, and relatedness satisfaction and frustration of well and ill-being within the same time point was conducted during COVID-19 pandemic (early April 2020) [16]. Although their results were mostly in line with prior research with need satisfaction uniquely predicting wellbeing and need frustration uniquely predicting ill-being, there were some exceptions. Both positive and negative affect predicted life satisfaction through competence frustration, but competence satisfaction was unrelated to life satisfaction. Similarly, relatedness frustration was not a consistent predictor of general distress. Importantly, this paper only examined one type of wellbeing and one type of ill-being, so questions remain as to the relationship with need satisfaction and frustration on well and ill-being and whether these relationships look different for multiple types of wellbeing and ill-being.

Investigating the three needs separately seems especially important during the COVID-19 pandemic. It could be that people’s autonomy feels especially threatened due to lockdown measures that prevent people from carrying out their normal routines, or that competence is especially threatened due to the amount of new information people are taking in regarding how to keep themselves safe, or that social distancing measures have people feeling particularly lonely and not related. Past research in self-determination theory suggests that people benefit from being fulfilled in all three needs regardless of the importance they place on that need, such as feeling a stronger desire to master skills (to feel competent) over relating to others [13]. Perhaps the pandemic has created a context in which satisfaction of one particular need (or frustration of a particular need) explains more or less variation in wellbeing (or ill-being). For example, Skewes et al. [17] studied faculty members undergoing the promotion and tenure process, which they described as an inherently low autonomy situation. However, many faculty members were still doing well if they were able to compensate for the low autonomy by focusing on ways to increase their competence and relatedness.

Although global pandemics are not new, multiple aspects of the COVID-19 pandemic make it unprecedented (e.g., transmissibility, economic impact) [18] and a worthy context in which to examine how established psychological theories perform in the prediction of mental health in extraordinary times. The aim of the current study was to examine the extent to which need satisfaction and frustration predicted mental health, consistent with SDT. We investigated several markers of both positive and negative functioning—hedonic wellbeing, or satisfaction with life, eudaimonic wellbeing, or growth and purpose, stress, and depressive thinking. Further, this study examined mental health over time, to support the existing literature that has focused on the beginning of the pandemic (e.g., 4).

All study procedures were approved by the University’s Institutional Review Board and the study was pre-registered prior to data analysis.

Specifically, the measures and hypotheses/research questions were pre-registered after data was collected at Time 2 but before data was collected at Time 3 (https://aspredicted.org/w9ps5.pdf). Pre-registration was not done prior to Time 1 data collection because these data were originally collected as part of a larger dataset.

In April 2020, a sample of Americans adults was recruited through Prolific (a participant recruitment company) to complete an online survey, entitled “Understanding Daily American Life during a Pandemic” (Time 1). Participants were contacted again in late July/early August 2020 via Prolific and asked to participate in an originally unplanned follow-up study (Time 2) and one more time in late January/early February 2021 (Time 3). Participants were compensated with US$4.50 at Time 1, US$3.50 at Time 2, and US$2.50 at Time 3 (corresponding to an approximate hourly compensation rate of US$10.00 for each time point).

According to the U.S. Coronavirus Tracker [1], during the 3 days of Time 1 data collection, the average number of deaths per day was 1,746 with an average number of 31,036 new cases. During the 9 days of Time 2 data collection, the average number of deaths was 1,101 per day with an average number of 58,020 new cases. During the 11 days of Time 3 data collection, the average number of deaths was 3,165 with an average number of 129,193 new cases per day. The number of days the study was available in Time 1 was based on the amount of time it took us to reach our target participant rate of 400. This amount of time was higher in Times 2 and 3 to allow for more participants to participate.

The final sample represented a total of 193 participants who completed the study at all three time points and passed three or more of the four attention checks, representing a retention rate of 47% from the original 407 participants from Time 1. The sample size was chosen to match the availability of funds, but an a priori power analysis suggests that we would need about 150 participants to detect a medium effect size, indicating that we have adequate power to test our primary hypotheses. Demographics of this final sample can be found in Table 1. Participants in our sample were relatively diverse in terms of age, race, employment status, and geographic region.

In order to determine if participants who responded to all three time points differed from those we lost due to attrition, we conducted a series of independent sample t-tests. Our results indicated that at Time 1, all participants were relatively equally satisfied in their need satisfaction and in their hedonic wellbeing (all t-values < 1.42). However, those who failed to return after Time 1 had lower multi-faceted (combined hedonic/eudaimonic) wellbeing, greater ill-being, and had greater levels of COVID-19 experience, compared to those who responded to all three time points. Furthermore, participants who failed to return after Time 1 tended to be younger (all t-values > 2.12). Those who failed to return after Time 2 were equivalent to those who responded to all time points in need satisfaction and wellbeing, but had higher rates of autonomy and competence frustration, as well as stress (t-values > 2.00).

Except for demographic variables (age, racial identity, gender, geographic region, employment status), all measures were assessed at all time points with higher scores indicating higher levels of the measured variable. Except for questions pertaining to experiences with COVID-19, all measures asked participants to reflect on the previous 7 days Reliability coefficients for all relevant measures can be found in Table 2.

All variables (except COVID-19 experience at Time 1 and Time 2) spanned mostly the full range of possible values and were relatively normally distributed, indicating that participants varied widely in how they were coping with the pandemic at all three time points, with most participants scoring around the mean of the distribution.

A series of one-sample t-tests examined whether participants scored significantly above or below the mid-point of each measure of well and ill-being (Table 2). Need satisfaction scores were higher than the midpoint at all time points and ill-being and need frustration scores were significantly lower than the midpoint at all time points. Surprisingly, wellbeing scores were not different from the midpoint at any time point.

The results of the current study raise several important questions that we encourage future research to consider. For example, it is unclear whether people were especially sensitive to changes in autonomy and competence during the COVID-19 pandemic, a time characterized by increased regulation and much uncertainty, or whether similar results would emerge outside of the pandemic. Research should continue to investigate whether there are various contexts that influence the extent to which individuals are especially frustrated in their autonomy, competence, or relatedness. Notably, factors directly concerning the pandemic influence differences in wellbeing, such as the degree of isolation that people experience during lockdowns and the number of days that individuals have been in lockdowns [14] and certainly factors not directly related to the pandemic influence wellbeing as well such as meaning in life [31] sleep quality [32], financial wellbeing [33], and even travel [34]. Because these were all impacted during the pandemic to varying degrees, future research should investigate the downstream effects of these factors on wellbeing as well as their unique effects on need satisfaction and frustration.

Although the current study adds to our understanding of both the COVID-19 pandemic and SDT, the findings should be considered in light of several limitations. Using an online research platform allowed us to gather a sample from around the country and such samples have generally been found to provide quality data (e.g., [35]); however, the types of people providing data online during a pandemic may differ in important ways from the general population (e.g., access to the internet, more time at home). Additionally, our attrition analyses indicated that the participants who failed to return after Time 1 were doing less well than those who responded to all three time points. This may partially account for the fact that participants in our sample generally reported higher than average levels of need satisfaction and lower than average need frustration. However, our data did span mostly the full range of values for every variable, which indicates that there was still high variance in our participants’ experiences.

Participants were also asked to reflect on the preceding week when completing measures, which may or may not have adequately captured the full variability in people’s lives. Research conducted during the pandemic has shown that people’s daily experiences do vary (even during lockdowns) and that these experiences have differential impacts on wellbeing [7]. Consistent with previous work that has examined daily variability in SDT needs and outcomes (e.g., [31, 36]), future research would benefit from the use of experience sampling techniques in order to capture events that predict both need satisfaction and frustration, as well and ill-being.

The COVID-19 pandemic has led to dramatic shifts in much of life, including how individuals work, maintain relationships, and accomplish daily tasks. Our participants reported a wide range of mental health experiences during the pandemic, with most doing relatively well. Satisfaction and frustration of autonomy, competence, and relatedness explained significant amounts of variance in multiple indices of wellbeing and ill-being. Of particular interest was the unique importance of autonomy satisfaction, which predicted more wellbeing and less ill-being. This may indicate that it is important for people to feel a sense of choicefulness, especially during periods of high restriction which serve to protect the health and safety of everyone. Public health experts may try to encourage people to maintain choicefulness, finding things that they intrinsically enjoy doing, especially during lockdowns when autonomy may be especially threatened. Of course, it is important to recognize that maintaining choicefulness will be easier for some, such as those who have access to appropriate resources, and that both psychological needs and physical needs are predictive of mental health [37]. In sum, our findings represent an important addition in our understanding of mental health during the COVID-19 pandemic, as well as an important contribution to the self-determination theory literature regarding the relationship between need frustration and well and ill-being.