A deep dive into our new study in Aging & Mental Health, and why the determinants of perceived ageism are not what most policy frameworks assume.
By Prof. Dr. Mehmet Nurullah Kurutkan & Serhat Fırat Estimated reading time: 10 minutes
The Paradox That Started It All
Turkish culture is famous for its reverence toward elders. We give up our seats on public transport. We address older people as amca, teyze, dede, or nine. We expect elderly parents to live with their children rather than in institutional care. By most cultural measures, Türkiye should be one of the least ageist societies on earth.
And yet, in our newly published study in Aging & Mental Health — drawing on the nationally representative Türkiye Elderly Profile Survey (TYPA) 2023 with 13,758 valid records from adults aged 50 and over — we found that 82.6 percent of participants reported experiencing ageism. Across every age group we examined, the prevalence exceeded 80 percent.
That number is not a typo. It is a paradox, and it tells us something important: explicit cultural respect can coexist with — and even mask — implicit, structural, and everyday discrimination.
This blog post walks you through what we did, what we found, and why we think the determinants of perceived ageism demand a fundamentally different policy response than the one most governments currently deploy.
Why We Needed a New Approach
The ageism literature is rich, but it has three persistent blind spots that we set out to address:
First, most studies analyze a handful of predictors in isolation. A paper on depression and ageism here, a paper on digital literacy and ageism there — but no study, to our knowledge, has evaluated social participation, digital competence, health status, economic resilience, healthcare access, geriatric depression, loneliness, happiness, education, gender, marital status, and lifestyle behaviors together within a single, integrated, nationally representative analytical framework.
Second, the 50–64 age cohort — what we call the “near-elderly” or “future-elderly” population — is almost always missing from ageism research. This is a serious omission. Ageism begins being perceived well before official retirement age, and if we want early-warning systems and preventive policy, we need to understand the cohort that is about to become the target of structural discrimination.
Third, most existing work relies either on theory-driven regression (which assumes linear relationships) or on data-driven machine learning (which is agnostic to theory but often uninterpretable). We wanted to do both — and have them talk to each other.
Our Study in One Paragraph
We analyzed data from TYPA-2023 (collected by TURKSTAT from 22,640 households nationwide between October and December 2023). After cleaning, we had 13,758 valid records of adults aged 50+. We measured perceived ageism using an 11-item scale, which exploratory factor analysis split into two nearly orthogonal sub-dimensions: structural/institutional ageism (6 items, e.g., respect for elderly rights, valuing experience, state protection) and interpersonal/everyday ageism (5 items, e.g., verbal abuse, social exclusion, negative media portrayal). We then ran OLS regression models stratified across five age groups (50+, 50–64, 65–74, 75+, 65+) — ten models in total — and paired this with eight machine learning classifiers (XGBoost, GBM, Random Forest, Extra Trees, Decision Tree, SVM, KNN, and Logistic Regression), interpreted using SHAP analysis. Every predictor variable was anchored in at least one of three established aging theories: Stereotype Embodiment Theory (Levy, 2009), Terror Management Theory (Greenberg et al., 1986), and the WHO Comprehensive Ageism Framework (2021).
Finding #1: Geriatric Depression Is the Dominant Risk Factor — Everywhere
If there is one headline from this study, it is this: geriatric depression emerged as the strongest predictor of perceived ageism across every configuration we tested.
Let us be precise about what “every configuration” means here. Depression dominated across:
- 5 age groups (50+, 50–64, 65–74, 75+, 65+)
- 3 dependent variable specifications (binary aggregate ageism, continuous structural sub-score, continuous interpersonal sub-score)
- 2 analytical methods (parametric OLS regression and non-parametric XGBoost with SHAP)
That is 13 distinct analytical configurations, and in every single one, geriatric depression came out on top. More strikingly, its effect on structural ageism intensifies with age: the standardized beta climbs from 0.137 in the 50–64 group, to 0.152 in the 65–74 group, to 0.159 in the 75+ group. This age gradient is exactly what Terror Management Theory predicts — as death becomes more salient and psychological buffers weaken, perceived discrimination deepens.
This consistency is not a statistical curiosity. It is a policy imperative. Routine GDS-15 depression screening in primary care should be as standard as blood pressure measurement for everyone aged 50 and over. Mental health services for older adults are not a luxury; they are the single highest-return intervention target our data identified.
Finding #2: The Two Faces of Ageism Behave Very Differently
Here is where the study gets genuinely surprising. When we separated ageism into its structural and interpersonal dimensions, the predictor profiles diverged dramatically.
Consider social participation. In our original binary model (which simply asked “did the person experience ageism, yes or no”), social participation appeared to be broadly protective. But when we split the outcome:
- For structural ageism, the effect of social participation was essentially negligible (β = 0.033, and actually slightly positive in the aggregate).
- For interpersonal ageism, social participation was the strongest single protective factor across the entire study, reaching β = −0.232 in the 75+ group — the largest beta coefficient in any of our ten models.
The same binary-model masking showed up for education. In the aggregate binary model, education looked irrelevant (p = 0.922). But in the dimensional models, education displayed a classic suppression effect: positive for structural ageism (β = 0.050, meaning more educated older adults perceive more institutional discrimination) and slightly negative for interpersonal ageism. These opposite effects canceled each other out in the binary model, hiding a story that turns out to matter enormously for policy. Educated elderly people are not insulated from ageism; they are more aware of their rights and therefore more likely to recognize structural violations that less-educated peers might normalize.
Economic hardship followed a similar pattern — invisible in the binary model (p = 0.727), but a statistically significant predictor of interpersonal ageism once the outcome was properly decomposed.
The methodological lesson is blunt: treating ageism as a single construct obscures the mechanisms that make it modifiable. The policy lesson is equally sharp: a single-lever intervention will not work. Structural ageism is driven by institutional contact points (healthcare access, policy environments, state protection), while interpersonal ageism is driven by the thickness of one’s social fabric.
Finding #3: Digital Competence Protects the Young-Old but Loses Its Power in the Old-Old
Digital literacy is widely promoted as a universal remedy for ageism. Our data tell a more nuanced story.
For adults aged 65–74, digital competence was a strong dual protector — significant for both structural (β = −0.063) and interpersonal (β = −0.121) ageism. Being able to navigate e-government, mobile health apps, and social media meaningfully reduced perceived discrimination on both fronts.
For adults aged 75 and over, however, digital literacy’s interpersonal protective effect disappeared entirely (β = −0.028, p = 0.366). It retained a modest structural benefit, but it no longer reduced everyday, face-to-face discrimination.
This matters because billions of dollars are being invested globally in “digital inclusion” programs for older adults — often with the tacit assumption that one program fits all. Our findings suggest otherwise. For the 75+ group, direct digital training will not solve interpersonal ageism. Resources should shift toward face-to-face, neighborhood-based social participation programs — which, as we just saw, yield the largest protective effect observed in the entire study.
In practical terms: digital literacy programs for ages 65–74, community-based social programs for ages 75+, and hybrid models during the transition years. Age-undifferentiated policy is inefficient policy.
Finding #4: Methods That Agree and Methods That Disagree
One of the most valuable things a combined TD/DD design gives you is the ability to see where methods agree (high confidence) and where they diverge (areas for deeper investigation).
Where OLS and XGBoost converged: Geriatric depression, digital literacy, healthcare access, and alcohol use. These four variables surfaced as important under both analytical lenses. We treat these as the most reliable policy targets in the study.
Where they diverged: Social participation reached conventional statistical significance for structural ageism in OLS (β = 0.033, p = 0.002), but XGBoost classified it as practically irrelevant for that dimension — suggesting that our very large sample size inflated statistical significance beyond practical importance. Conversely, general health and economic status sometimes fell below OLS significance thresholds but were assigned meaningful importance by tree-based algorithms in specific subgroups.
We also learned something important about predictive ceilings. Our ML models achieved ROC-AUC values between roughly 0.61 and 0.66 — moderate, not spectacular. Systematic hyperparameter optimization experiments failed to push performance meaningfully higher (maximum ΔAUC = 0.006). This is not a modeling failure; it is a substantive finding. Structural ageism’s R² of 0.038 tells us that individual-level survey data simply cannot predict institutional discrimination well, because institutional discrimination is driven by macro-level mechanisms that live above the individual. Policy environments, administrative practices, media framing, and political rhetoric do not show up in a household survey. They should not be expected to.
What This Means for Clinicians, Policymakers, and Health Managers
Let me translate these findings into concrete recommendations, age-stratified, because that is where the study earns its keep:
For adults aged 50–64 (the near-elderly cohort): Intervention windows are widest here, and relationships are largely linear. Prioritize tobacco cessation, early digital literacy programs, and social network maintenance as people transition out of intensive workplace interaction.
For adults aged 65–74 (the young-old): This is where digital inclusion delivers its highest return. Active programs in e-government navigation, mobile health, and social media engagement protect against both structural and interpersonal ageism. Routine GDS-15 screening should be standardized in primary care — depression in this group operates through a threshold effect on structural ageism, meaning early intervention, before the threshold is crossed, is disproportionately effective.
For adults aged 75 and over (the old-old): Pivot decisively toward face-to-face, community-based social participation. Intergenerational mentoring, neighborhood activity groups, life-story work, and meaning-making programs yield the largest protective effects we measured anywhere in the study. Depression in this group accumulates gradually rather than through thresholds, so sustained psychosocial support — not just crisis intervention — is the right frame. And pay particular attention to elderly men: gender becomes significant only in the 75+ group for structural ageism (β = 0.078, p = 0.007), with men perceiving higher institutional discrimination — likely reflecting autonomy and status loss in institutional interactions.
Across all age groups: Expand home-care services and mobile health units to address the consistent association between healthcare access difficulty and perceived ageism. Train healthcare staff in geriatric communication. Address alcohol use, which showed negative effects across every age group.
The Turkish Sociocultural Context — And Why It Matters Internationally
I want to return to the paradox I opened with, because the Turkish context genuinely shapes how these findings should be read, and how they should not be generalized.
Türkiye is in the middle of a rapid demographic and structural transition. The proportion of older adults rose from 8 percent in 2014 to 10.6 percent in 2024. Urbanization is separating adult children from elderly parents. The traditional extended-family safety net is weakening. Pension systems are strained. Public services are digitizing faster than the oldest cohorts can follow.
In this environment, the fact that social participation emerges as the strongest protective factor — and especially the strikingly large β = −0.232 in the 75+ group — suggests that community engagement is functioning as a compensatory mechanism as family-based support erodes. Meanwhile, the education suppression effect (educated elderly perceive more structural ageism) likely reflects heightened awareness of rights violations in a society that officially reveres but structurally marginalizes its elderly.
For international readers: direct comparison of our 82.6 percent ageism prevalence with figures from Western contexts may be misleading. The item set, the cultural framing of respect, and the survey modality all interact with results. But the structural insights — the two-dimensional nature of ageism, the dominance of depression, the age-varying effect of digital literacy, the divergence between structural and interpersonal predictors — travel well. We expect these patterns to replicate, with local calibration, in other aging societies.
Limitations (Because Every Serious Paper Has Them)
Our design is cross-sectional, so we cannot speak to causality with the confidence a panel design would permit. Some composite indices (Social Participation, α = 0.635; Payment Hardship, α = 0.585) showed lower-than-ideal internal consistency, though tetrachoric-based EFA confirmed their factor structures and Pearson-based Cronbach’s alpha systematically underestimates reliability for dichotomous items. Perceived ageism as a construct is sensitive to social-desirability and framing effects. And the revised two-dimensional operationalization — though internally validated — should be replicated in independent Turkish samples and other national contexts before being treated as settled.
Future work should pursue panel data, natural experiments, or pre-post intervention designs that approximate causality, and cross-national replication is essential.
A Closing Thought: What Numbers Can and Cannot Do
The reason we built this study the way we built it — a theory-driven backbone, a data-driven exploration, a two-dimensional outcome, and five age strata — is that ageism is not the kind of phenomenon that yields to a single model. It is woven through psychology, sociology, economics, institutional design, and cultural narrative. Any methodology that pretends otherwise is flattering itself.
But numbers, used carefully, can cut through paradox. Our data show that a society can revere its elders in word and still fail them in practice. They show that the 82.6 percent of our respondents who report experiencing ageism are not complaining about loss of respect — they are telling us about depression that goes unscreened, social networks that have thinned, digital systems that have left them behind, and healthcare encounters that treat age as a diagnosis.
None of that is inevitable. All of it is addressable. That is why we did this work.
Read the full article: Firat, S., & Kurutkan, M. N. (2026). Determinants of the Perceived Ageism Index using theory-driven and data-driven approaches: evidence from the Türkiye Elderly Profile Survey 2023. Aging & Mental Health. https://doi.org/10.1080/13607863.2026.2656224
About the authors: Serhat Fırat is affiliated with the Department of Health Management, Faculty of Health Sciences, Hakkari University. Prof. Dr. Mehmet Nurullah Kurutkan is affiliated with the Department of Health Management, Düzce University, where his research spans patient safety, quality in health, health policy, and bibliometric analysis in health management.
If you found this post useful, consider sharing it with colleagues in gerontology, public health, health management, or social policy. Comments and reactions welcome.
When Respect Hides Discrimination: What 13,758 Older Turkish Adults Taught Us About Ageism
A deep dive into our new study in Aging & Mental Health, and why the determinants of perceived ageism are not what most policy frameworks assume.
By Prof. Dr. Mehmet Nurullah Kurutkan & Serhat Fırat Estimated reading time: 10 minutes
The Paradox That Started It All
Turkish culture is famous for its reverence toward elders. We give up our seats on public transport. We address older people as amca, teyze, dede, or nine. We expect elderly parents to live with their children rather than in institutional care. By most cultural measures, Türkiye should be one of the least ageist societies on earth.
And yet, in our newly published study in Aging & Mental Health — drawing on the nationally representative Türkiye Elderly Profile Survey (TYPA) 2023 with 13,758 valid records from adults aged 50 and over — we found that 82.6 percent of participants reported experiencing ageism. Across every age group we examined, the prevalence exceeded 80 percent.
That number is not a typo. It is a paradox, and it tells us something important: explicit cultural respect can coexist with — and even mask — implicit, structural, and everyday discrimination.
This blog post walks you through what we did, what we found, and why we think the determinants of perceived ageism demand a fundamentally different policy response than the one most governments currently deploy.
Why We Needed a New Approach
The ageism literature is rich, but it has three persistent blind spots that we set out to address:
First, most studies analyze a handful of predictors in isolation. A paper on depression and ageism here, a paper on digital literacy and ageism there — but no study, to our knowledge, has evaluated social participation, digital competence, health status, economic resilience, healthcare access, geriatric depression, loneliness, happiness, education, gender, marital status, and lifestyle behaviors together within a single, integrated, nationally representative analytical framework.
Second, the 50–64 age cohort — what we call the “near-elderly” or “future-elderly” population — is almost always missing from ageism research. This is a serious omission. Ageism begins being perceived well before official retirement age, and if we want early-warning systems and preventive policy, we need to understand the cohort that is about to become the target of structural discrimination.
Third, most existing work relies either on theory-driven regression (which assumes linear relationships) or on data-driven machine learning (which is agnostic to theory but often uninterpretable). We wanted to do both — and have them talk to each other.
Our Study in One Paragraph
We analyzed data from TYPA-2023 (collected by TURKSTAT from 22,640 households nationwide between October and December 2023). After cleaning, we had 13,758 valid records of adults aged 50+. We measured perceived ageism using an 11-item scale, which exploratory factor analysis split into two nearly orthogonal sub-dimensions: structural/institutional ageism (6 items, e.g., respect for elderly rights, valuing experience, state protection) and interpersonal/everyday ageism (5 items, e.g., verbal abuse, social exclusion, negative media portrayal). We then ran OLS regression models stratified across five age groups (50+, 50–64, 65–74, 75+, 65+) — ten models in total — and paired this with eight machine learning classifiers (XGBoost, GBM, Random Forest, Extra Trees, Decision Tree, SVM, KNN, and Logistic Regression), interpreted using SHAP analysis. Every predictor variable was anchored in at least one of three established aging theories: Stereotype Embodiment Theory (Levy, 2009), Terror Management Theory (Greenberg et al., 1986), and the WHO Comprehensive Ageism Framework (2021).
Finding #1: Geriatric Depression Is the Dominant Risk Factor — Everywhere
If there is one headline from this study, it is this: geriatric depression emerged as the strongest predictor of perceived ageism across every configuration we tested.
Let us be precise about what “every configuration” means here. Depression dominated across:
- 5 age groups (50+, 50–64, 65–74, 75+, 65+)
- 3 dependent variable specifications (binary aggregate ageism, continuous structural sub-score, continuous interpersonal sub-score)
- 2 analytical methods (parametric OLS regression and non-parametric XGBoost with SHAP)
That is 13 distinct analytical configurations, and in every single one, geriatric depression came out on top. More strikingly, its effect on structural ageism intensifies with age: the standardized beta climbs from 0.137 in the 50–64 group, to 0.152 in the 65–74 group, to 0.159 in the 75+ group. This age gradient is exactly what Terror Management Theory predicts — as death becomes more salient and psychological buffers weaken, perceived discrimination deepens.
This consistency is not a statistical curiosity. It is a policy imperative. Routine GDS-15 depression screening in primary care should be as standard as blood pressure measurement for everyone aged 50 and over. Mental health services for older adults are not a luxury; they are the single highest-return intervention target our data identified.
Finding #2: The Two Faces of Ageism Behave Very Differently
Here is where the study gets genuinely surprising. When we separated ageism into its structural and interpersonal dimensions, the predictor profiles diverged dramatically.
Consider social participation. In our original binary model (which simply asked “did the person experience ageism, yes or no”), social participation appeared to be broadly protective. But when we split the outcome:
- For structural ageism, the effect of social participation was essentially negligible (β = 0.033, and actually slightly positive in the aggregate).
- For interpersonal ageism, social participation was the strongest single protective factor across the entire study, reaching β = −0.232 in the 75+ group — the largest beta coefficient in any of our ten models.
The same binary-model masking showed up for education. In the aggregate binary model, education looked irrelevant (p = 0.922). But in the dimensional models, education displayed a classic suppression effect: positive for structural ageism (β = 0.050, meaning more educated older adults perceive more institutional discrimination) and slightly negative for interpersonal ageism. These opposite effects canceled each other out in the binary model, hiding a story that turns out to matter enormously for policy. Educated elderly people are not insulated from ageism; they are more aware of their rights and therefore more likely to recognize structural violations that less-educated peers might normalize.
Economic hardship followed a similar pattern — invisible in the binary model (p = 0.727), but a statistically significant predictor of interpersonal ageism once the outcome was properly decomposed.
The methodological lesson is blunt: treating ageism as a single construct obscures the mechanisms that make it modifiable. The policy lesson is equally sharp: a single-lever intervention will not work. Structural ageism is driven by institutional contact points (healthcare access, policy environments, state protection), while interpersonal ageism is driven by the thickness of one’s social fabric.
Finding #3: Digital Competence Protects the Young-Old but Loses Its Power in the Old-Old
Digital literacy is widely promoted as a universal remedy for ageism. Our data tell a more nuanced story.
For adults aged 65–74, digital competence was a strong dual protector — significant for both structural (β = −0.063) and interpersonal (β = −0.121) ageism. Being able to navigate e-government, mobile health apps, and social media meaningfully reduced perceived discrimination on both fronts.
For adults aged 75 and over, however, digital literacy’s interpersonal protective effect disappeared entirely (β = −0.028, p = 0.366). It retained a modest structural benefit, but it no longer reduced everyday, face-to-face discrimination.
This matters because billions of dollars are being invested globally in “digital inclusion” programs for older adults — often with the tacit assumption that one program fits all. Our findings suggest otherwise. For the 75+ group, direct digital training will not solve interpersonal ageism. Resources should shift toward face-to-face, neighborhood-based social participation programs — which, as we just saw, yield the largest protective effect observed in the entire study.
In practical terms: digital literacy programs for ages 65–74, community-based social programs for ages 75+, and hybrid models during the transition years. Age-undifferentiated policy is inefficient policy.
Finding #4: Methods That Agree and Methods That Disagree
One of the most valuable things a combined TD/DD design gives you is the ability to see where methods agree (high confidence) and where they diverge (areas for deeper investigation).
Where OLS and XGBoost converged: Geriatric depression, digital literacy, healthcare access, and alcohol use. These four variables surfaced as important under both analytical lenses. We treat these as the most reliable policy targets in the study.
Where they diverged: Social participation reached conventional statistical significance for structural ageism in OLS (β = 0.033, p = 0.002), but XGBoost classified it as practically irrelevant for that dimension — suggesting that our very large sample size inflated statistical significance beyond practical importance. Conversely, general health and economic status sometimes fell below OLS significance thresholds but were assigned meaningful importance by tree-based algorithms in specific subgroups.
We also learned something important about predictive ceilings. Our ML models achieved ROC-AUC values between roughly 0.61 and 0.66 — moderate, not spectacular. Systematic hyperparameter optimization experiments failed to push performance meaningfully higher (maximum ΔAUC = 0.006). This is not a modeling failure; it is a substantive finding. Structural ageism’s R² of 0.038 tells us that individual-level survey data simply cannot predict institutional discrimination well, because institutional discrimination is driven by macro-level mechanisms that live above the individual. Policy environments, administrative practices, media framing, and political rhetoric do not show up in a household survey. They should not be expected to.
What This Means for Clinicians, Policymakers, and Health Managers
Let me translate these findings into concrete recommendations, age-stratified, because that is where the study earns its keep:
For adults aged 50–64 (the near-elderly cohort): Intervention windows are widest here, and relationships are largely linear. Prioritize tobacco cessation, early digital literacy programs, and social network maintenance as people transition out of intensive workplace interaction.
For adults aged 65–74 (the young-old): This is where digital inclusion delivers its highest return. Active programs in e-government navigation, mobile health, and social media engagement protect against both structural and interpersonal ageism. Routine GDS-15 screening should be standardized in primary care — depression in this group operates through a threshold effect on structural ageism, meaning early intervention, before the threshold is crossed, is disproportionately effective.
For adults aged 75 and over (the old-old): Pivot decisively toward face-to-face, community-based social participation. Intergenerational mentoring, neighborhood activity groups, life-story work, and meaning-making programs yield the largest protective effects we measured anywhere in the study. Depression in this group accumulates gradually rather than through thresholds, so sustained psychosocial support — not just crisis intervention — is the right frame. And pay particular attention to elderly men: gender becomes significant only in the 75+ group for structural ageism (β = 0.078, p = 0.007), with men perceiving higher institutional discrimination — likely reflecting autonomy and status loss in institutional interactions.
Across all age groups: Expand home-care services and mobile health units to address the consistent association between healthcare access difficulty and perceived ageism. Train healthcare staff in geriatric communication. Address alcohol use, which showed negative effects across every age group.
The Turkish Sociocultural Context — And Why It Matters Internationally
I want to return to the paradox I opened with, because the Turkish context genuinely shapes how these findings should be read, and how they should not be generalized.
Türkiye is in the middle of a rapid demographic and structural transition. The proportion of older adults rose from 8 percent in 2014 to 10.6 percent in 2024. Urbanization is separating adult children from elderly parents. The traditional extended-family safety net is weakening. Pension systems are strained. Public services are digitizing faster than the oldest cohorts can follow.
In this environment, the fact that social participation emerges as the strongest protective factor — and especially the strikingly large β = −0.232 in the 75+ group — suggests that community engagement is functioning as a compensatory mechanism as family-based support erodes. Meanwhile, the education suppression effect (educated elderly perceive more structural ageism) likely reflects heightened awareness of rights violations in a society that officially reveres but structurally marginalizes its elderly.
For international readers: direct comparison of our 82.6 percent ageism prevalence with figures from Western contexts may be misleading. The item set, the cultural framing of respect, and the survey modality all interact with results. But the structural insights — the two-dimensional nature of ageism, the dominance of depression, the age-varying effect of digital literacy, the divergence between structural and interpersonal predictors — travel well. We expect these patterns to replicate, with local calibration, in other aging societies.
Limitations (Because Every Serious Paper Has Them)
Our design is cross-sectional, so we cannot speak to causality with the confidence a panel design would permit. Some composite indices (Social Participation, α = 0.635; Payment Hardship, α = 0.585) showed lower-than-ideal internal consistency, though tetrachoric-based EFA confirmed their factor structures and Pearson-based Cronbach’s alpha systematically underestimates reliability for dichotomous items. Perceived ageism as a construct is sensitive to social-desirability and framing effects. And the revised two-dimensional operationalization — though internally validated — should be replicated in independent Turkish samples and other national contexts before being treated as settled.
Future work should pursue panel data, natural experiments, or pre-post intervention designs that approximate causality, and cross-national replication is essential.
A Closing Thought: What Numbers Can and Cannot Do
The reason we built this study the way we built it — a theory-driven backbone, a data-driven exploration, a two-dimensional outcome, and five age strata — is that ageism is not the kind of phenomenon that yields to a single model. It is woven through psychology, sociology, economics, institutional design, and cultural narrative. Any methodology that pretends otherwise is flattering itself.
But numbers, used carefully, can cut through paradox. Our data show that a society can revere its elders in word and still fail them in practice. They show that the 82.6 percent of our respondents who report experiencing ageism are not complaining about loss of respect — they are telling us about depression that goes unscreened, social networks that have thinned, digital systems that have left them behind, and healthcare encounters that treat age as a diagnosis.
None of that is inevitable. All of it is addressable. That is why we did this work.
Read the full article: Firat, S., & Kurutkan, M. N. (2026). Determinants of the Perceived Ageism Index using theory-driven and data-driven approaches: evidence from the Türkiye Elderly Profile Survey 2023. Aging & Mental Health. https://doi.org/10.1080/13607863.2026.2656224
About the authors: Serhat Fırat is affiliated with the Department of Health Management, Faculty of Health Sciences, Hakkari University. Prof. Dr. Mehmet Nurullah Kurutkan is affiliated with the Department of Health Management, Düzce University, where his research spans patient safety, quality in health, health policy, and bibliometric analysis in health management.
If you found this post useful, consider sharing it with colleagues in gerontology, public health, health management, or social policy. Comments and reactions welcome.