Research and advocacy of progressive and pragmatic policy ideas.

Research Public Goods

A Year of Living Under COVID-19

Part 4: What Are the Most Significant Drivers of Mental Health During this Global Pandemic?

Age, gender, and marital status emerge as greatest predictors of negative mental health in our study.

By Aziff Azuddin, Edwin Goh & Ziad Razak30 May 2021

In February this year, we began a study to assess the effects of the then year-long pandemic on Malaysians’ mental health. At the time, a second nation-wide lockdown had just been declared. Data for this mental health study was collected via online survey which reached nearly 1,000 respondents from across the country*.

*Respondents’ mental health was measured by their responses to the DASS-21 questionnaire. For more details on how the DASS-21 questionnaire was applied, respondent demographics and the study’s overall methodology, please refer to Part 1. Please note that this is not a nationally representative stratified sample.

A descriptive analysis of the data was published in three parts, each tackling a different aspect of the general deterioration experienced over the past year. Part 1 showed that over half of surveyed respondents experienced worsened mental health over the past year. Women and youths self-reported lower mental health levels compared to other demographic groups. There was also a clear relationship between physical and mental well-being; those who reported worsened physical health over the past year also experienced worsened mental health.

Part 2 explored the impact of living arrangements and social connection on mental well-being. Social isolation appeared to be a factor; respondents who live alone self-reported worse mental well-being compared to those who live with others. Additionally, respondents who work from home every day also appeared to have worse mental health levels compared to those who worked from home less often.

Part 3 looked at the impact of changes to employment and income on mental well-being. The unemployed appeared most vulnerable mental health-wise, and similarly for part-time employees. Respondents earning less than RM 5,000 also appeared to experience worse mental well-being relative to other income groups.

In this fourth and final instalment of the research series, we sought to understand which of the variables included in our study were the most statistically significant in driving high scores in depression, anxiety and stress levels. We conducted a multiple variable analysis on the data set, using multiple linear regression.

Regression analysis is a statistical method used to determine the strength of the relationship between a dependent variable (in this study, self-reported mental health scores) and one or more independent or predictor variables (for example, age, gender etc.) in a regression model. Multiple regression analysis allows us to determine the relative contribution of each of the predictor variables to the model, in order to identify the most significant predictors.

Results of the Multiple Variable Analysis

The multiple variable analysis on our data set found age and gender as highly significant predictors of depression, anxiety and stress while marital status is a highly significant predictor of depression and anxiety (Table 1). Other significant predictors are change in income and level of household income. Comparatively, other variables such as living alone (i.e. household occupancy) and working from home frequency were not as significant as suggested by the descriptive analysis.

Table 1: Predictors of mental well-being

With respect to age, there is a very significant likelihood that people aged 35 years and younger are experiencing worse mental health in terms of depression, anxiety and stress levels compared to older age groups. Uncertainty over future prospects, financial distress, social isolation and lack of quality sleep have been attributed as the likely causes by a global cross-sectional COVID-19 and well-being study published this year.

As for gender, there is a very significant likelihood that women have worse depression, anxiety and stress levels compared to men. A study published last year exploring the impact of the pandemic on women shows that gender-related challenges such as parenting challenges, uneven distribution of domestic care and the increased risk of domestic violence are some of the key contributors to women’s mental health difficulties.

Marital status is a very significant predictor of negative mental health as well; there is a very significant likelihood that people who are single during the pandemic have worse depression and anxiety levels compared to those who are married. A relationship quality and COVID-19 study published last year showed that single individuals had poorer mental health compared to those who were in relationships of good quality. This is supported by a social relationships and health study that showed how social support systems and a sense of meaning found in relationships led to better mental health and health behaviours overall.

Change in income is a significant predictor for depression, anxiety and stress. Based on our data set, there is a significant likelihood that a fall in income over the past year is accompanied by higher levels of depression, anxiety and stress compared to an increase or no change in income levels.

The absolute level of household income is a significant predictor for anxiety; there is a significant likelihood that individuals with household incomes less than RM5,000 experience higher anxiety levels compared to those with higher household incomes.

In Part 2, we discussed the impact of social isolation on self-reported mental health scores, particularly in those living alone and those who work from home (WFH) daily. Although the descriptive analysis showed a somewhat clear pattern in negative mental health scores, the variables of household occupancy and WFH frequency were only weakly significant as predictors of negative mental health.

Employment status (e.g. whether full-time, part-time or unemployed), number of financial dependents and availability of personal space were not included in the regression model as these variables were highly correlated with some of the above-mentioned variables and were very weakly significant in explaining the patterns observed in negative mental health scores.


At the time of this writing, a full national lockdown had just been declared as it became clear that the third Movement Control Order then underway was not enough to bring down the rate of virus transmission. With case numbers breaching 9,000 a day, it is hard to visualise better days. But if we look to the history of past global pandemics and if the necessary precautions are taken, we can reasonably hope that the current worrying situation will gradually abate over the next one to two years.

However, we remain concerned about the long-term implications of the pandemic, beyond the immediate crisis. COVID-19 and successive lockdowns have amplified and exacerbated problems latent in Malaysian society, including widespread mental health challenges

Our study found that young people, women, unmarried Malaysians and those with reduced incomes are the ones most likely to have negative mental health after the year-long pandemic (and counting). These findings are generally consistent with other mental well-being studies, both local and worldwide.

These are groups that require particular mental health outreach and support, now and after the pandemic, alongside the various measures of economic support that have been announced. It is crucial that these groups are intentionally targeted for mental health support services by the government and civil society.

In addition, it is also necessary to have institutional reform. We have to rethink how existing ministries and agencies are involved in mental health support. Although the Ministry of Health leads public mental health, mental health support and services are fragmented across different ministries, such as the Ministry of Women, Family and Community Development and the Ministry of Education. There is a need to rebalance assistance budgets in relevant ministries to ensure greater recognition for and a higher level of support for mental health modules within various existing socioeconomic assistance and relief programmes.

There is also an argument for decentralising service provision and resources, and empowering regional and local authorities to address healthcare issues based on detailed local knowledge and contexts, particularly amongst vulnerable populations.

The fallout of the COVID-19 pandemic will be felt for many years, or even decades, to come. We need to address these very real social challenges now, if we are serious about helping Malaysians navigate their way safely into the post-pandemic future.

Academics and researchers interested in the data and further details of this research project may contact the principal researcher at

The Centre is a centrist think tank driven by research and advocacy of progressive and pragmatic policy ideas. We are a not-for-profit and a mostly remote working organisation.