class: center, middle, inverse, title-slide .title[ # The Causal Effects of School Entry Age ] .subtitle[ ## A Methodological and Empirical Analysis ] .author[ ### Mehmet Yerlikaya ] .date[ ### April 2025 ] --- class: center, middle # The Causal Effects of School Entry Age ## A Methodological and Empirical Analysis ### Mehmet Yerlikaya #### Master's Thesis in Economics #### Advisor: [Advisor Name] #### April 2025 --- # Research Question & Narrative Framework .pull-left[ ## Central Research Question: "How does school entry age causally affect educational and labor market outcomes across different contexts and time horizons?" ## The Narrative Journey: - The age paradox: Initial advantages may become long-term disadvantages - Methodological challenges in establishing causality - From quasi-experimental methods to policy implications ] .pull-right[ <div class="grViz html-widget html-fill-item" id="htmlwidget-68aad693baadaf2da649" style="width:432px;height:432px;"></div> <script type="application/json" data-for="htmlwidget-68aad693baadaf2da649">{"x":{"diagram":"\n digraph research_framework {\n graph [rankdir=TB, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=lightblue, fontname=Helvetica, fontsize=12, margin=0.2]\n \n q [label=\"Research Question\", fillcolor=\"#1c5253\", fontcolor=\"white\"]\n \n node [fillcolor=lightyellow]\n short [label=\"Short-Term\nOutcomes\"]\n long [label=\"Long-Term\nOutcomes\"]\n labor [label=\"Labor Market\nOutcomes\"]\n \n node [shape=ellipse, fillcolor=lightgreen]\n policy [label=\"Policy\nImplications\"]\n \n q -> short\n q -> long\n q -> labor\n short -> policy\n long -> policy\n labor -> policy\n \n {rank=same; short; long; labor}\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ## Significance: > "The age at which children begin formal education represents one of the earliest and most consequential educational policy decisions." ] --- # Theoretical Framework & Mechanisms .pull-left[ ## Multiple Causal Pathways: ### Absolute Age Effects: - Developmental readiness for cognitive tasks - Maturity advantages in non-cognitive skills ### Relative Age Effects: - Peer comparisons influence self-concept - Teacher expectations and evaluations - Streaming and ability grouping ### Institutional Effects: - Track placement in early-tracking systems - Grade retention probabilities - Special education placement ] .pull-right[ <div class="grViz html-widget html-fill-item" id="htmlwidget-7a59ff0f61194f0be04d" style="width:432px;height:432px;"></div> <script type="application/json" data-for="htmlwidget-7a59ff0f61194f0be04d">{"x":{"diagram":"\n digraph causal_pathways {\n graph [rankdir=TB, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=lightblue, fontname=Helvetica, fontsize=12, margin=0.1]\n \n age [label=\"School Entry Age\", fillcolor=\"#1c5253\", fontcolor=\"white\"]\n \n node [shape=ellipse, fillcolor=lightyellow]\n abs [label=\"Absolute Age\nEffects\"]\n rel [label=\"Relative Age\nEffects\"]\n inst [label=\"Institutional\nEffects\"]\n \n node [shape=box, fillcolor=lightgreen]\n cog [label=\"Cognitive\nOutcomes\"]\n noncog [label=\"Non-Cognitive\nOutcomes\"]\n attain [label=\"Educational\nAttainment\"]\n labor [label=\"Labor Market\nOutcomes\"]\n \n age -> abs\n age -> rel\n age -> inst\n \n abs -> cog\n abs -> noncog\n rel -> cog\n rel -> noncog\n inst -> attain\n \n cog -> attain\n noncog -> attain\n attain -> labor\n \n {rank=same; abs; rel; inst}\n {rank=same; cog; noncog}\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ] --- # Methodological Challenges .pull-left[ ## The Fundamental Challenge: > "Simple comparisons of outcomes by school entry age will confound the causal effect with unobserved factors." ## Three Key Identification Problems: 1. **Selection Bias** - Parents' strategic decisions - Correlation with family background 2. **Age-at-Test Confound** - Perfect collinearity with school starting age - Maturity effects vs. schooling effects ] .pull-right[ 3. **Institutional Sorting** - Different educational tracks - Grade retention policies - Special education placement <div class="grViz html-widget html-fill-item" id="htmlwidget-6e6bcee87c146867b6f5" style="width:432px;height:360px;"></div> <script type="application/json" data-for="htmlwidget-6e6bcee87c146867b6f5">{"x":{"diagram":"\n digraph methodological_challenges {\n graph [rankdir=LR, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=\"#E8E8E8\", fontname=Helvetica, fontsize=12]\n \n bias [label=\"Selection\nBias\", fillcolor=\"#FFD6D6\"]\n confound [label=\"Age-at-Test\nConfound\", fillcolor=\"#D6EAFF\"]\n sorting [label=\"Institutional\nSorting\", fillcolor=\"#D6FFD6\"]\n \n node [shape=ellipse, fillcolor=\"#FFFFD6\"]\n quasi [label=\"Quasi-Experimental\nDesigns\"]\n \n bias -> quasi [label=\" requires \"]\n confound -> quasi [label=\" requires \"]\n sorting -> quasi [label=\" requires \"]\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ## Methodological Imperative: > "Without a credible research design for uncovering causal effects, empirical work serves no useful purpose." ] --- # Quasi-Experimental Approaches .pull-left[ ## Instrumental Variables (IV): - Uses expected school entry age based on birth date - Key assumption: Birth date is exogenous - Exploits cutoff rules as natural experiment ## Regression Discontinuity Design (RDD): - Compares children born just before/after cutoffs - Key assumption: Continuity in potential outcomes - Controls for seasonal birth patterns ] .pull-right[ ## Difference-in-Differences (DiD): - Exploits policy changes in entry age rules - Controls for time-invariant confounders - Accounts for cohort effects <div class="grViz html-widget html-fill-item" id="htmlwidget-e35b847c20695175b553" style="width:432px;height:360px;"></div> <script type="application/json" data-for="htmlwidget-e35b847c20695175b553">{"x":{"diagram":"\n digraph methods {\n graph [rankdir=TB, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=\"#1c5253\", fontcolor=\"white\", fontname=Helvetica, fontsize=12]\n \n selection [label=\"Identification\nChallenges\"]\n \n node [shape=ellipse, style=filled, fillcolor=\"#DBF3F4\", fontcolor=\"black\"]\n iv [label=\"Instrumental\nVariables\"]\n rdd [label=\"Regression\nDiscontinuity\"]\n did [label=\"Difference in\nDifferences\"]\n \n node [shape=box, style=rounded, fillcolor=\"#F3FAFC\", fontcolor=\"black\"]\n iv_a [label=\"Birth date as\ninstrument\"]\n rdd_a [label=\"Cutoff\ncomparisons\"]\n did_a [label=\"Policy\nchanges\"]\n \n selection -> iv\n selection -> rdd\n selection -> did\n \n iv -> iv_a\n rdd -> rdd_a\n did -> did_a\n \n {rank=same; iv; rdd; did}\n {rank=same; iv_a; rdd_a; did_a}\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ] --- # Educational Outcomes - Short-Term Effects .pull-left[ ## Test Score Advantages: - Consistent finding across countries - Stronger effects for mathematics - Declines but persists through primary school ## Non-Cognitive Outcomes: - Improved self-regulation - Reduced inattention/hyperactivity - Enhanced social-emotional development ## Heterogeneous Effects: - Larger for boys than girls - Varies by socioeconomic status - Differences by pre-school experience ] .pull-right[ <img src="school_entry_age_presentation_files/figure-html/effect-sizes-plot-1.png" style="display: block; margin: auto;" /> ] --- # Educational Outcomes - Long-Term Effects .pull-left[ ## Diminishing Academic Advantages: - Test score advantage declines with age - Largely disappears by secondary school - Convergence through educational career ## Educational Attainment Effects: - Mixed evidence on high school completion - Variable effects on college attendance - Persistence depends on educational system ] .pull-right[ <img src="school_entry_age_presentation_files/figure-html/long-term-trends-1.png" style="display: block; margin: auto;" /> ## Track Placement: - Strong effects in early-tracking systems - Potential long-term consequences for educational trajectories ] --- # Labor Market Outcomes .pull-left[ ## Earnings Effects: - Small negative effect on earnings until ~age 30 - Trade-off: maturity advantage vs. lost experience - Returns to education vs. returns to experience ## Cross-Country Variation: - Smaller negative effects in credential-based systems - Larger in experience-valued labor markets - Institutional context matters significantly ] .pull-right[ <img src="school_entry_age_presentation_files/figure-html/labor-market-1.png" style="display: block; margin: auto;" /> ## Life-Course Effects: - Later labor market entry - Delayed family formation - Timing of major life transitions ] --- # Socioeconomic Inequality Implications .pull-left[ ## SES Differences in Effects: - Mixed evidence on differential impacts - Home environment quality is key mediator - Enrichment opportunities during delay period ## Parental Decision-Making: - Higher-SES parents more likely to "redshirt" - Strategic advantages in flexible systems - Information asymmetries matter ] .pull-right[ <img src="school_entry_age_presentation_files/figure-html/ses-differences-1.png" style="display: block; margin: auto;" /> ## Policy Implications for Equity: - Rigid vs. flexible policies have different distributional consequences - Need targeted support for disadvantaged students ] --- # Policy Implications .pull-left[ ## School Entry Cutoffs: - Any cutoff creates a distribution of relative ages - Earlier cutoffs lead to older entry age cohorts - Trade-offs between short and long-term effects ## Flexibility and Support: - More flexibility in initial grade placement - Targeted instructional support for youngest students - Developmental screening before school entry ] .pull-right[ <div class="grViz html-widget html-fill-item" id="htmlwidget-291cd6c2adbd656ad21f" style="width:504px;height:360px;"></div> <script type="application/json" data-for="htmlwidget-291cd6c2adbd656ad21f">{"x":{"diagram":"\n digraph policy_options {\n graph [rankdir=TB, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=\"#1c5253\", fontcolor=\"white\", fontname=Helvetica, fontsize=12]\n \n issue [label=\"Relative Age\nEffects\"]\n \n node [shape=ellipse, style=filled, fillcolor=\"#DBF3F4\", fontcolor=\"black\"]\n cutoffs [label=\"Cutoff Date\nPolicies\"]\n flex [label=\"Flexibility\nPolicies\"]\n support [label=\"Support\nPrograms\"]\n training [label=\"Teacher\nTraining\"]\n \n node [shape=box, style=rounded, fillcolor=\"#F3FAFC\", fontcolor=\"black\"]\n cut_opt [label=\"• Earlier/Later cutoffs\n• Regional variation\n• Quarterly entry\"]\n flex_opt [label=\"• Assessment-based entry\n• Flexible progression\n• Grade retention policies\"]\n sup_opt [label=\"• Targeted interventions\n• Extra resources\n• Differentiated instruction\"]\n train_opt [label=\"• Age-awareness training\n• Bias recognition\n• Modified assessment\"]\n \n issue -> cutoffs\n issue -> flex\n issue -> support\n issue -> training\n \n cutoffs -> cut_opt\n flex -> flex_opt\n support -> sup_opt\n training -> train_opt\n \n {rank=same; cutoffs; flex; support; training}\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ## Teacher Training: - Awareness of relative age bias in assessments - Modified evaluation practices - Differentiated instruction ] --- # Methodological Limitations & Future Research .pull-left[ ## Persistent Challenges: - Selection bias with strategic birth timing - Age-at-test vs. school entry age effects - Long-term follow-up data gaps ## Future Research Directions: - Mechanisms of relative age effects - Interaction with educational policies - Heterogeneous effects across populations - Alternative interventions to reduce disadvantages - Life-course outcomes beyond education and earnings ] .pull-right[ <img src="school_entry_age_presentation_files/figure-html/research-gaps-1.png" style="display: block; margin: auto;" /> ## Data Needs: - Actual entry age vs. expected entry age - Pre-school experience information - Long-term longitudinal tracking ] --- # Conclusion - The Age Paradox Revisited .pull-left[ ## The Temporal Paradox: - Short-term academic advantages - Diminishing effects over educational career - Potential labor market disadvantages ## The Contextual Contingency: - Effects vary by educational system - Labor market structure matters - Individual characteristics moderate effects ] .pull-right[ <div class="grViz html-widget html-fill-item" id="htmlwidget-2d9a0c40505191f5ca40" style="width:504px;height:360px;"></div> <script type="application/json" data-for="htmlwidget-2d9a0c40505191f5ca40">{"x":{"diagram":"\n digraph conclusion {\n graph [rankdir=TB, bgcolor=transparent]\n node [shape=box, style=filled, fillcolor=\"#1c5253\", fontcolor=\"white\", fontname=Helvetica, fontsize=12]\n \n entry [label=\"School Entry Age\nPolicy\"]\n \n node [shape=ellipse, style=filled, fillcolor=\"#F8F8F8\", fontcolor=\"black\"]\n balance [label=\"Balance of\nTrade-offs\"]\n context [label=\"Contextual\nConsiderations\"]\n flex [label=\"Flexible\nApproaches\"]\n \n node [shape=box, style=rounded, fillcolor=\"#DBF3F4\", fontcolor=\"black\"]\n short [label=\"Short-term\nbenefits\"]\n long [label=\"Long-term\ncosts\"]\n system [label=\"Educational\nsystem\"]\n market [label=\"Labor\nmarket\"]\n indiv [label=\"Individual\nfactors\"]\n screen [label=\"Developmental\nscreening\"]\n support [label=\"Targeted\nsupport\"]\n \n entry -> balance\n entry -> context\n entry -> flex\n \n balance -> short\n balance -> long\n context -> system\n context -> market\n context -> indiv\n flex -> screen\n flex -> support\n \n {rank=same; balance; context; flex}\n }\n","config":{"engine":"dot","options":null}},"evals":[],"jsHooks":[]}</script> ## The Policy Challenge: - No universal optimal school entry age - Need for complementary policies - Flexible approaches accommodating individual differences ] --- class: center, middle # Thank You & Questions ## Contact Information: ### Mehmet Yerlikaya ### [mehmet.yerlikaya@fau.de] ### [Department of Economics] .footnote[Mehmet Yerlikaya | April 2025]