What if the strongest clues to student anxiety and depression are not the pandemic itself, but personality patterns and childhood emotional abuse that remain important after the variables are separated? University students are at high risk for anxiety and depression. The main objective was to tease apart variance in symptom severity uniquely attributable to four associated variables that are frequently confounded: exposure to childhood emotional abuse, alexithymia, sensory processing sensitivity, and anxiety sensitivity.

Alexithymia, sensory processing sensitivity, and anxiety sensitivity can co-occur, and all three are related to adverse mental health outcomes. Because these variables are rarely studied together, it is difficult to assess the role each might play in mental health outcomes or how each relates to exposure to emotional abuse and other childhood adverse experiences.

The primary aim was to tease apart the variance in university students’ mental health outcomes uniquely attributable to each variable. The first key objective was to present data collected during the COVID-19 pandemic on alexithymia, sensory processing sensitivity, anxiety sensitivity, childhood emotional abuse, pandemic-related impacts, anxiety, and depression in a large, mixed-sex sample of university students.

The study was mainly descriptive, to characterize the sample, and expected rates of anxiety and depression to be high, with many study variables positively related to one another. The second objective was to test for possible sex differences in the study variables, given mixed findings in the literature.

The expectation was that females would score higher than males on sensory processing sensitivity, anxiety sensitivity, pandemic-related impacts, anxiety, and depression. It was unclear whether males and females would differ on alexithymia or experiences of childhood emotional abuse.

The third objective was to examine how the pandemic may have differentially impacted male and female students’ exercise routines and whether this was related to current mental health. The expectation was that decreasing activity after the pandemic began would be a risk factor for poor mental health, whereas maintaining or increasing physical activity would be protective.

The fourth objective was to assess the unique contributions of personality variables and childhood emotional abuse to mental health while controlling for current physical activity, COVID-19 impacts, and sex. The variables were frequently confounded, so teasing apart each unique contribution was intended to clarify risk and protective factors affecting university students during this stressful time.

The findings may have important implications for theory and clinical practice. The final sample included 410 individuals with a mean age of 19.7 years, a standard deviation of 3.4, and an age range from 17 to 44. The majority identified as female, at 75.4 percent, and as Canadian citizens, at 83.7 percent.

A large proportion identified as White or European, at 44.7 percent. Because pandemic restrictions prevented in-person testing, participants completed an online survey measuring demographics, alexithymia, sensory processing sensitivity, anxiety sensitivity, childhood emotional abuse, physical activity, pandemic impacts, depression, and anxiety.

The survey also included a measure of conscientious responding to check for poor effort. Participants completed the survey remotely through Qualtrics at a time and location of their choosing. The procedures were approved by the Psychology and Sociology Research Ethics Board at the University of Manitoba, and participants gave informed consent.

The DASS-21 measures symptoms of depression, anxiety, and stress, but this investigation administered only its anxiety and depression subscales. Participants rated 14 statements about the past week on a four-point Likert scale ranging from zero, meaning the statement did not apply at all, to three, meaning it applied very much or most of the time.

Each subscale produced a total score out of 42, with higher scores indicating greater symptom severity. Individuals could be assigned to normal, mild, moderate, severe, or extremely severe symptom categories using manual cut points. The study characterized the full sample of 410 individuals with descriptive statistics on the study variables and used those data to illustrate correlations between the variables.

Scores on all variables covered virtually the whole range, and most variables were positively correlated with one another, with moderate to large effect sizes. Almost one-third of the sample, 30.5 percent, scored at or above the traditional cutoff for alexithymia.

Over half of the sample reported moderate to extremely severe symptoms of depression and anxiety, with 52.9 percent reporting each. Figure one is a scatterplot of standardized Highly Sensitive Person Scale scores against Anxiety Sensitivity Index-three scores, with points colored by lexithymic, borderline, or alexithymic classification on the Toronto Alexithymia Scale.

The points span a broad range, showing how sensitivity and anxiety-sensitivity scores are distributed across these three alexithymia groups. This matters because it characterizes the full sample and makes the relationships among the study variables visible.

Welch’s t-tests were used to look for possible sex differences, with a Bonferroni-corrected significance threshold of p less than or equal to zero point zero zero seven. There was no sex difference in mean alexithymia scores, and the proportion above the traditional alexithymia cutoff was similar in males and females.

Females had higher mean sensory processing sensitivity scores than males, and more than twice as many females as males were classified as highly sensitive: 35.0 percent versus 15.5 percent. Females also scored significantly higher than males on the Anxiety Sensitivity Index-3.

Table three compares ninety-seven males with three hundred nine females across psychological measures, coronavirus impact, mood, and weekly activity. The authors report significant t-test values for HSPS, ASI-3, and mCIS at the table’s listed p-values, while DASS D and Total METs show p-values of zero point five seven nine and zero point two two one, respectively.

These comparisons matter because the authors used them to assess possible sex differences, while accounting for multiple tests with a stricter significance threshold. Female sex continued to show a trend toward predicting higher Coronavirus Impact Scale scores in both models.

After controlling for all other variables, sensory processing sensitivity, a history of childhood emotional abuse, and higher physical activity emerged as the strongest predictors. Table 4 presents hierarchical regression models predicting negative COVID-19 impacts, measured by the modified Coronavirus Impact Scale.

In Step 1, HSPS, emotional abuse history, and physical activity level have statistically significant associations, with p values of 0.010, less than 0.001, and less than 0.001, respectively. Step 2 adds depression and anxiety scores, but neither is statistically significant: DASS D has p equal to 0.649, and DASS A has p equal to 0.206.

The depression model was significant and accounted for 44.9 percent of the variance in depression scores. All three personality variables, childhood emotional abuse, and male sex emerged as significant predictors of depression. The anxiety model was also significant and accounted for 51.1 percent of the variance in anxiety scores.

Again, all three personality variables and childhood emotional abuse emerged as significant predictors, while COVID impacts and physical activity accounted for no unique variance in either model after the other variables were controlled. Table five reports separate multiple regressions predicting depression and anxiety from alexithymia, sensory processing sensitivity, anxiety sensitivity, childhood emotional abuse, coronavirus impact, physical activity, and sex.

Anxiety sensitivity was a significant predictor for both outcomes, with beta values of zero point two zero three for depression and zero point four six one for anxiety, while physical activity was not significant for either outcome. The table matters because it identifies which variables accounted for unique variance after the predictors were considered together.

Sensory processing sensitivity was the most important of the three personality variables in explaining individual differences in perceived COVID-19 impacts. Perceived COVID-19 impacts were also stronger among those reporting childhood emotional abuse, those engaging in higher physical activity, and, to a lesser extent, females.

Once personality and childhood emotional abuse were accounted for, pandemic-related impacts did not improve prediction of anxiety or depression in the university student sample. A limitation is that the study took place during the pandemic, when quarantine, social distancing, and institutional restrictions on in-person testing were in place.

The study therefore relied on online data collection techniques, which may have reduced the accuracy with which certain constructs were measured. Even with the Conscientious Responders Scale, participants may have been less attentive and careful online than they might have been during in-person testing.

The study demonstrates that alexithymia, sensory processing sensitivity, anxiety sensitivity, and emotional abuse suffered in childhood each make a unique contribution to predicting mental health outcomes. Assessing variance uniquely attributable to each variable was important because past research has often ignored their correlations, potentially confounding previously described links with other measures.

The central finding is that alexithymia, sensory processing sensitivity, anxiety sensitivity, and childhood emotional abuse each contributed unique variance to mental health outcomes, while COVID-19 impacts and physical activity did not uniquely predict anxiety or depression.