Research | Open Access | Volume 9 (3): Article 114 | Published: 11 Jul 2026
Menu, Tables and Figures
Kingsley Eyram King-Kuadzi1,&
1Department of Psychology, Heritage Christian University, Amasaman-Accra, Ghana
&Corresponding author: Kingsley Eyram King-Kuadzi, Department of Psychology, Heritage Christian University, Amasaman-Accra, Ghana, Email: kek024a@hcu.edu.gh, ORCID: https://orcid.org/0009-0002-1107-3814
Received: 05 Apr 2025, Accepted: 10 Jul 2026, Published: 11 Jul 2026
Domain: Mental Health
Keywords: Gambling-related harm, psychological distress, youth mental health, Ghana, public health
©Kingsley Eyram King-Kuadzi, Journal of Interventional Epidemiology and Public Health (ISSN: 2664-2824). This is an Open Access article distributed under the terms of the Creative Commons Attribution International 4.0 License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Cite this article: Kingsley Eyram King-Kuadzi, Sports betting frequency and psychological distress among youth in Greater Accra, Ghana: An exploratory cross-sectional study. Journal of Interventional Epidemiology and Public Health. 2026; 9(3):114. https://doi.org/10.37432/jieph-d-26-00120
Introduction: Sports betting is increasingly prevalent among young people in Ghana, yet there is limited local evidence on how betting frequency and borrowing to bet relate to psychological distress.
Methods: A cross-sectional exploratory survey was conducted among 100 youths aged 18-30 years in selected university campuses, betting centres and youth-dominated spaces in Greater Accra. Participants were recruited using purposive and convenience sampling after eligibility screening. Psychological distress was measured with a nine-item adapted Depression Anxiety Stress Scale used as a general distress index, not as a diagnostic instrument. Descriptive statistics, Spearman correlation, independent-samples t-test, Mann-Whitney U sensitivity analysis and multivariable linear regression with HC3 heteroscedasticity-robust standard errors were used. Statistical significance was set at p < 0.05.
Results: Psychological distress scores ranged from 0 to 27 (mean 7.50, standard deviation 5.66). Betting frequency was positively associated with psychological distress in bivariate analysis (Spearman r_s = 0.335, 95%CI: 0.155 – 0.503, p = 0.001). Respondents who borrowed money to bet reported a slightly higher mean distress score than those who did not, but the difference was not statistically significant (mean difference: 0.15; 95%CI: -2.62 to 2.92; p = 0.914). The Mann-Whitney U sensitivity analysis also showed no statistically significant difference by borrowing status (U = 790.5, p = 0.744). In the adjusted model, betting frequency remained independently associated with higher distress (adjusted beta 2.78, 95%CI: 1.71 – 3.86, p < 0.001), while borrowing remained non-significant.
Conclusion: More frequent sports betting was associated with higher psychological distress among youths in this non-probability exploratory sample. The findings do not establish causality. Betting frequency may help identify youth who warrant further psychological assessment, but it should not be treated as a validated screening tool until it has been tested in larger, longitudinal, and more representative studies.
Sports betting has become one of the fastest-growing forms of gambling among young people globally, partly because mobile technology, internet access, online betting platforms and digital payment systems have made gambling more accessible [1]. Among young people, sports betting is often framed as recreation, a socially acceptable pastime and, in some settings, a perceived opportunity for quick financial gain [2]. Despite this normalisation, public health concerns have increased because repeated betting may be linked to emotional, behavioural and financial harms.
Gambling-related behaviours have been associated with psychological distress, depressive symptoms, anxiety, stress, financial strain, interpersonal conflict and reduced well-being [1,3,4]. Young adults may be particularly vulnerable because this developmental stage is often characterised by peer influence, experimentation, financial instability and ongoing development of self-control and decision-making capacities [2,5]. Under these conditions, sports betting may shift from a leisure activity to a repeated behavioural pattern with mental health implications. Borrowing money to bet is a specific behaviour that requires closer attention. It may indicate financial strain, loss-chasing, reduced behavioural control or attempts to recover previous betting losses [3,6]. Borrowing may also generate distress through debt pressure, shame, interpersonal conflict and difficulty repaying borrowed funds [6,7]. However, borrowing is not always captured with sufficient detail in youth gambling studies, and a simple yes-or-no measure may underestimate the psychological importance of debt burden, repayment difficulty and subjective debt stress.
In Ghana, sports betting has become increasingly visible in urban areas where football culture, mobile money, digital platforms and youth unemployment intersect. Greater Accra is an important setting because it is highly urbanised, densely populated, and characterised by widespread access to betting centres, smartphones, and online financial transactions [8,9]. Recent Ghanaian evidence has linked gambling severity with psychological distress among young people [4]. However, there remains a need for additional local epidemiologic evidence on specific betting behaviours, particularly betting frequency and borrowing to bet.
This study examined the relationship between the frequency of sports betting and psychological distress among youth in Greater Accra, Ghana. It also assessed whether psychological distress differed between youth who had borrowed money for sports betting and those who had not. The study was framed as exploratory and hypothesis-generating because the sampling approach was non-probability-based, the borrowing measure was narrow, and the adapted distress scale was not intended to diagnose depression, anxiety, or stress disorders.
Study design and setting
This study used a cross-sectional quantitative design to estimate associations among sports betting frequency, borrowing behaviour, and psychological distress at a single point in time. The design was appropriate for examining statistical associations but did not permit causal inference. The study was conducted in selected areas of Greater Accra, Ghana, including university campuses, betting centres, and youth-dominated social spaces where sports betting was common. Reporting was checked against the relevant STROBE guidance for observational studies, particularly the descriptions of setting, participants, variables, bias, statistical methods, and limitations.
Study population, eligibility and recruitment
The target population was youths aged 18-30 years who had previously engaged in sports betting. Eligible participants were required to be within the age range, report lifetime experience of sports betting, be able to complete the questionnaire and provide informed consent. Individuals were excluded if they were below 18 years, above 30 years, had never engaged in sports betting, declined consent or submitted substantially incomplete questionnaires.
Participants were recruited using purposive and convenience sampling. Purposive sampling was used to identify youth spaces where sports betting was likely to occur, while convenience sampling was used to recruit eligible participants who were available at the selected sites. Before completing the questionnaire, potential respondents were screened verbally for age, previous sports betting experience and willingness to participate. A total of 100 usable questionnaires were included in the analysis. No questionnaire was excluded for missing outcome data because all included records had complete psychological distress items. Respondents were included if they were aged 18 to 30 years, had engaged in sports betting, were willing to participate voluntarily, and provided informed consent. The study was conducted from 4 December 2025 to 15 February 2026. 123 participants were approached; 8 declined, and of the 115 who agreed, 100 were found to be eligible.
Sample size justification
The sample size was determined pragmatically because the study was an exploratory cross-sectional survey using a hard-to-enumerate population of youths involved in sports betting. A sample of 100 provides approximately 80% power at a two-sided alpha of 0.05 to detect a correlation of about 0.28 or larger between betting frequency and psychological distress. For the borrowing comparison, the available group sizes provided adequate power only for moderate-to-large differences; therefore, the borrowing analysis was interpreted cautiously, especially because only 21 respondents reported borrowing money to bet. The adjusted regression model was also interpreted cautiously because the sample size limited the number of parameters that could be estimated reliably.
Data collection and measures
Data were collected using a structured self-administered questionnaire. The questionnaire captured socio-demographic characteristics, sports betting behaviour, borrowing behaviour, psychological distress and perceived fear of legal or debt-related consequences. Questionnaires were administered online using a Google Form version, where needed to improve access and participation. Sports betting frequency was measured as an ordinal variable and coded so that higher values represented more frequent betting. The exact categories used in the questionnaire were coded as follows: 1 = never; 2 = rarely; 3 = sometimes; 4 = often; and 5 = very often. Although the response options included “never”, eligibility required lifetime sports betting experience. Two respondents selected “never” for current betting frequency; these responses were retained and interpreted as current non-engagement among respondents with prior betting experience rather than as lifetime non-betting.
Borrowing behaviour was measured with a dichotomous item asking whether the respondent had ever borrowed money to place bets. The response was coded as yes or no. Additional borrowing-related items assessed the source of borrowed money, the number of times borrowed in the past month, and failure to repay, but the primary borrowing analysis used a yes-or-no variable due to small cell sizes in the more detailed borrowing categories. This measure did not capture the amount borrowed, repayment difficulty, current debt burden or subjective debt stress.
Psychological distress was measured with a nine-item adapted short version of the Depression Anxiety Stress Scale. Items assessed depressive, anxiety-related and stress-related symptoms over the past week, including low mood, difficulty winding down, low self-worth, meaninglessness, trembling, difficulty relaxing, fear without reason, low initiative and dryness of mouth. Each item was scored from 0 (“did not apply to me at all”) to 3 (“applied to me very much”). Total scores ranged from 0 to 27, with higher scores indicating greater psychological distress. The scale was used as a general distress index rather than as three separate subscales or as a diagnostic screening tool because the study did not administer the full DASS-21 structure and the adapted index has not been validated specifically among Ghanaian youth involved in sports betting. The shortened distress index was used to reduce respondent burden in field-based and betting-centre recruitment contexts. The questionnaire was reviewed by experts in counselling psychology, mental health and research methods for content relevance. The psychological distress items showed good internal consistency in this sample (Cronbach’s alpha = 0.86), but internal consistency alone was not interpreted as evidence of construct, criterion or diagnostic validity. The nine distress items are provided in Supplementary Appendix A.
Data management and statistical analysis
Data were coded, cleaned and analysed using statistical software. Descriptive statistics were used to summarise socio-demographic characteristics, betting behaviour and psychological distress. Missing data were checked before analysis. All 100 records had complete psychological distress data. Spearman’s rank-order correlation was used to examine the bivariate relationship between ordinal betting frequency and psychological distress. A bootstrap 95% confidence interval was estimated for the Spearman correlation. An independent-samples t-test compared mean distress scores between respondents who had borrowed money to bet and those who had not. Normality of distress scores within borrowing groups was assessed using the Shapiro-Wilk test and inspection of distributional summaries. Homogeneity of variance was assessed using Levene’s test. The borrowing comparison was interpreted with caution because the borrowing group was small.
A multivariable linear regression model was then fitted to estimate the adjusted association between betting frequency and psychological distress, controlling for borrowing status, age group, gender, educational level, employment status, monthly income, and prior mental health diagnosis or treatment. Heteroscedasticity-robust standard errors were used because the outcome distribution was not perfectly normal. The betting amount was not included in the primary model because 41 records contained a non-substantive code in the amount variable, making direct interpretation unsafe. A sensitivity model restricted to respondents with substantive betting amounts was estimated and is described in the text. Results are reported with 95% confidence intervals and p-values. Statistical significance was set at p < 0.05.
Ethical considerations
Approval for the study was obtained from Heritage Christian University under reference number HCU/SHaSS/DPsy/GC.26. Written informed consent was obtained from all participants before they completed the questionnaire. Participants were informed about the purpose of the study, confidentiality protections, anonymity, voluntary participation, and their right to decline any question or withdraw before completing the questionnaire without penalty.
The study involved no invasive procedure, clinical intervention, collection of direct personal identifiers, or access to sensitive medical records. No names, phone numbers, residential addresses, or other direct identifiers were collected. Completed questionnaires and electronic data were securely stored and accessed only by the researcher. Data were reported in aggregate form to protect participants’ identities. Given the sensitivity of discussing sports betting, borrowing behaviour, and psychological distress, participant privacy and well-being were actively safeguarded. Recruitment was conducted with attention to privacy, particularly at betting centres and youth-dominated spaces. Participants were approached individually where possible and were not required to disclose their betting behaviour publicly. Betting operators were not asked to identify eligible participants, and participation was not linked to access to betting services.
Participants were allowed to complete the questionnaire privately and away from peers, betting operators, recruiters, or any other persons who could influence their responses. Where participants experienced discomfort or reported distress, they were advised to contact a qualified counsellor or mental health professional, including counselling or mental health services available through their campus, community, or nearest health facility.
Socio-demographic characteristics
A total of 100 youths who had engaged in sports betting were included. Nearly half were aged 20-25 years (49.0%), 32.0% were aged 26-30 years, and 19.0% were below 20 years, as shown in Table 1. Most respondents were male (64.0%), undergraduates (58.0%) and students (62.0%). Monthly income was most commonly between GH₵500 and GH₵999 (38.0%).
Psychological distress and betting behaviour
Psychological distress scores ranged from 0 to 20, with a mean of 7.50 and a standard deviation of 5.66. This indicates considerable variation in distress within the sample. As presented in Table 2, betting frequency was distributed as follows: never/current non-engagement, 2.0%; rarely, 41.0%; sometimes, 34.0%; often, 6.0%; and very often, 17.0%. Although mean distress did not increase monotonically across all betting-frequency categories, the overall ordinal association between betting frequency and distress was positive and statistically significant. The lower mean distress score in the “often” category should be interpreted with caution because this subgroup contained only 6 respondents.
Betting frequency was positively associated with psychological distress (Spearman r_s = 0.335, 95% CI: 0.155 – 0.503, p = 0.001), as summarised in Table 3. The result suggests that respondents who reported more frequent betting also tended to report higher psychological distress.
Borrowing behaviour and psychological distress
Twenty-one respondents (21.0%) reported borrowing money to bet, while 79 (79.0%) did not. Respondents who borrowed money to bet had a mean distress score of 7.62 (SD 6.86), compared with 7.47 (SD 5.34) among those who did not borrow money to bet, as shown in Table 4. The mean difference was small and not statistically significant (0.15; 95% CI: -2.62 – 2.92; t(98) = 0.108, p = 0.914; Cohen’s d = 0.03). The bivariate and sensitivity analyses in Table 3 further showed that borrowing money to bet was not significantly associated with psychological distress. Levene’s test did not indicate unequal variances (p = 0.369). Shapiro-Wilk tests indicated non-normality in both groups. The Mann-Whitney U sensitivity analysis was also non-significant (U = 790.5, p = 0.744), with median distress scores of 4.0 among those who borrowed and 7.0 among those who did not borrow. The borrowing finding was therefore interpreted as inconclusive rather than evidence that borrowing is unrelated to distress.
Multivariable analysis
In the adjusted linear regression model using HC3 robust standard errors, betting frequency remained independently associated with higher psychological distress after controlling for borrowing status, age group, gender, education, employment status, income and history of mental health diagnosis or treatment. As shown in Table 5, the adjusted beta for betting frequency was 2.78, with a 95% CI: 1.71 – 3.86 and p < 0.001. Gender was entered into the model as a binary variable, with female respondents used as the reference category. After adjustment for other covariates, male gender was significantly associated with lower psychological distress scores compared with female gender (β = -4.61, 95% CI: -7.05 – -2.16, p < 0.001). Borrowing money to bet was not independently associated with distress in the adjusted model shown in Table 5 (adjusted beta 1.83, 95% CI: -0.88 – 4.55, p = 0.186). The adjusted model explained 33.8% of the variance in distress scores.
Regression diagnostics supported cautious interpretation. The model included eight predictors plus an intercept among 100 participants, leaving 91 residual degrees of freedom. Residuals showed mild non-normality (Shapiro-Wilk p = 0.021), while the Breusch-Pagan test did not show statistically significant heteroscedasticity (p = 0.179); HC3 robust standard errors were nevertheless retained. Influence diagnostics showed five observations above the Cook’s distance screening threshold of 4/n and nine observations with leverage above 2p/n; these cases were retained because they were plausible survey values and no data-entry errors were identified. Variance inflation factors indicated notable collinearity among the age, education, and employment/income covariates (maximum VIF = 21.99), so the demographic coefficients were not overinterpreted; the main interpretation focuses on betting frequency, while demographic covariates were included primarily for adjustment. In a sensitivity model restricted to respondents with substantive betting-amount data (n = 59), betting amount was not significantly associated with distress (p = 0.863); however, this sensitivity analysis was underpowered and should be interpreted cautiously.
This study found that more frequent sports betting was associated with higher psychological distress among youth in Greater Accra. The bivariate association was small to moderate and statistically significant; it remained significant in the primary adjusted model. This finding is consistent with evidence that greater gambling involvement and gambling severity are linked to poorer mental health outcomes among young people [1,3,4].
The finding should not be interpreted causally. Because the study was cross-sectional, the temporal direction cannot be established. Frequent betting may increase distress through financial losses, frustration, regret, uncertainty or loss-chasing. Conversely, distressed youth may engage in betting more often as a coping strategy, distraction or perceived financial solution. A bidirectional relationship is also plausible. The appropriate interpretation is therefore that betting frequency was associated with psychological distress in this sample, not that betting caused distress.
Borrowing money to bet was not significantly associated with psychological distress in either the bivariate comparison, the Mann-Whitney sensitivity analysis or the adjusted model. This result should be interpreted cautiously. Only 21 respondents reported borrowing money to bet, which limited statistical power. In addition, borrowing was measured as a simple yes-or-no item. This may not capture the dimensions of borrowing most relevant to distress, such as amount borrowed, borrowing frequency, repayment difficulty, debt burden, shame or conflict with lenders. The finding, therefore, does not show that borrowing is harmless; rather, it suggests that the crude borrowing measure used in this study failed to distinguish distress levels within this sample.
The Ghanaian context is important. Youth betting occurs within a social environment shaped by football culture, mobile money, unemployment, peer influence and the availability of physical and online betting platforms. In such a setting, asking about betting frequency in youth counselling, campus mental health services and community health promotion may help identify young people who warrant further assessment, psychoeducation, financial counselling or psychological support. This should be understood as a preliminary public health implication, not as evidence that betting frequency is already a validated screening test.
The study contributes local evidence to the growing public health discussion on sports betting, digital gambling and youth mental health in Africa. Its most defensible implication is preventive and hypothesis-generating rather than causal: youth who bet frequently may benefit from early mental health assessment, education on gambling-related harm, and interventions that challenge beliefs about betting as a reliable financial strategy. Future studies should test predictive validity, thresholds, sensitivity, specificity and follow-up outcomes before betting frequency is used as a formal screening marker.
Limitations
The study used a small non-probability sample of 100 respondents, so the findings cannot be generalised to all youth in Greater Accra or Ghana. The cross-sectional design prevents causal inference and does not establish whether betting preceded distress or distress preceded betting. Psychological distress was measured using an adapted nine-item distress scale, not the full DASS-21 diagnostic screening structure. Although internal consistency was good, Cronbach’s alpha alone does not establish validity; further psychometric validation in Ghanaian youth betting populations is needed. Borrowing behaviour was measured narrowly as a dichotomous variable, and the borrowing group was small. The measure did not capture the amount borrowed, borrowing frequency, repayment difficulty, debt burden or subjective debt stress. Self-report data may have been affected by recall bias and social desirability bias, especially because gambling and debt can be sensitive behaviours. Some variables, including betting amount, had coding limitations that constrained full adjustment in the primary multivariable model. The multivariable model was exploratory. Several covariates were coded as ordinal scores for parsimony, and multicollinearity was present among age, education, employment and income variables. Adjusted demographic coefficients should therefore not be interpreted as stable category-specific effects.
Sports betting frequency was associated with psychological distress among youth in Greater Accra, Ghana. Borrowing money for betting was not significantly associated with distress in this sample, but this finding is inconclusive because the borrowing group was small and borrowing was measured narrowly. The findings suggest that frequent betting may be a possible marker of emotional vulnerability that can prompt further assessment in counselling, campus mental health and community health settings. This implication is preliminary and requires validation in larger, more representative and longitudinal studies before betting frequency can be used as a formal screening tool. Future studies should clarify temporality, causal pathways, predictive validity and the role of financial strain.
What is already known about the topic
What this study adds
The author acknowledges all respondents who participated in the study and the individuals and institutions that facilitated data collection and ethical compliance.
Data availability
The dataset used for this study is not publicly deposited because it contains sensitive self-reported information on gambling behaviour, borrowing and psychological distress among young people. De-identified data may be made available by the corresponding author upon reasonable request, subject to ethical restrictions.
| Variable | Frequency (n=100) | Percentage |
|---|---|---|
| Age | ||
| Less than 20 | 19 | 19.0 |
| 20-25 | 49 | 49.0 |
| 26-30 | 32 | 32.0 |
| Gender | ||
| Male | 64 | 64.0 |
| Female | 36 | 36.0 |
| Educational level | ||
| Undergraduate | 58 | 58.0 |
| Graduate | 33 | 33.0 |
| Postgraduate | 9 | 9.0 |
| Employment status | ||
| Student | 62 | 62.0 |
| Employed | 25 | 25.0 |
| Unemployed | 13 | 13.0 |
| Monthly income | ||
| Less than GH₵500 | 25 | 25.0 |
| GH₵500-GH₵999 | 38 | 38.0 |
| GH₵1000-GH₵1999 | 25 | 25.0 |
| GH₵2000 and above | 12 | 12.0 |
| Code | Current betting frequency category | Frequency | Percentage | Mean distress | SD |
|---|---|---|---|---|---|
| 1 | Never/current non-engagement | 2 | 2.0 | 7.50 | 0.71 |
| 2 | Rarely | 41 | 41.0 | 5.22 | 5.42 |
| 3 | Sometimes | 34 | 34.0 | 9.50 | 5.41 |
| 4 | Often | 6 | 6.0 | 1.83 | 0.75 |
| 5 | Very often | 17 | 17.0 | 11.00 | 4.14 |
Table 3. Psychological distress and bivariate analyses
| Analysis | Statistic | Estimate | 95% CI | p-value |
|---|---|---|---|---|
| Psychological distress score | Mean (SD); range | 7.50 (5.66); 0-20 | ||
| Betting frequency and distress | Spearman rs | 0.335 | 0.155 to 0.503 | 0.001 |
| Borrowing and distress | Mean difference; Cohen’s d | 0.15; d = 0.03 | -2.62 to 2.92 | 0.914 |
| Borrowing and distress sensitivity analysis | Mann-Whitney U | 790.5 | Not applicable | 0.744 |
| Borrowed money to bet | N | Mean distress | SD | SE | Median (IQR) | Mean rank |
|---|---|---|---|---|---|---|
| Yes | 21 | 7.62 | 6.86 | 1.50 | 4.0 (2.0-14.0) | 48.64 |
| No | 79 | 7.47 | 5.34 | 0.60 | 7.0 (3.0-12.0) | 50.99 |
Table 5. Multivariable linear regression model for psychological distress with coding clarified
| Predictor | Adjusted beta (95% CI) | P value | Robust SE | Coding/reference group |
|---|---|---|---|---|
| Betting frequency | 2.78 (1.71-3.86) | <0.001 | 0.55 | Ordinal score: 1 = never/current non-engagement to 5 = very often |
| Borrowed money to bet | 1.83 (-0.88 to 4.55) | 0.186 | 1.39 | Binary indicator: yes = 1, no = 0; reference = no |
| Age group | -2.07 (-4.48 to 0.35) | 0.094 | 1.23 | Ordinal score: 1 = <20, 2 = 20-25, 3 = 26-30 |
| Male gender | -4.61 (-7.05 to -2.16) | <0.001 | 1.25 | Binary indicator: male = 1; reference = female |
| Educational level | -1.61 (-3.74 to 0.53) | 0.140 | 1.09 | Ordinal score: 1 = undergraduate, 2 = graduate, 3 = postgraduate |
| Employment status | 1.61 (0.10-3.12) | 0.036 | 0.77 | Ordinal score: 1 = student, 2 = employed, 3 = unemployed; interpreted only as adjustment |
| Monthly income | 0.28 (-0.94 to 1.50) | 0.657 | 0.62 | Ordinal score: 1 = <GH₵500 to 4 = GH₵2000 and above |
| Prior mental health history | -0.83 (-7.18 to 5.52) | 0.797 | 3.24 | Binary indicator: yes = 1, no = 0; reference = no |
Note. HC3 heteroscedasticity-robust standard errors were used. Categorical socio-demographic variables were entered as ordinal scores for parsimony in this exploratory model; therefore, demographic coefficients should not be read as category-specific causal estimates.