Structured Reading Strategies and Comprehension Gains Among Ghanaian Basic School Pupils

Technical Summary: Structured Reading Strategies and Comprehension Gains Among Ghanaian Basic School Pupils

Problem Statement
Reading comprehension remains a critical challenge for lower primary learners in Ghana, particularly within public basic schools. Despite English being the official language and medium of instruction, a significant structural disadvantage exists for pupils unable to read and comprehend English, affecting their performance across the entire curriculum. The study identifies that reading difficulties in this context are not merely personal deficits but systemic issues stemming from unstructured instructional approaches, a lack of diagnostic remediation, and the absence of systematic lesson plans. Preliminary observations at Model 2A Basic School in Wenchi revealed that 80% of Basic Two (B2) pupils could not accurately read two- and three-letter words, identify alphabet letters, or answer basic comprehension questions. Existing literature suggests that teachers often rely on repetitive oral drilling rather than structured, strategy-based interventions, leading to ingrained reading difficulties that persist into junior high school.

Methodology
The study employed an action research design utilizing a single-group, pretest–posttest approach over a four-week intervention period.

• Participants: The sample consisted of 50 purposively selected Basic Two pupils (30 boys, 20 girls; average age 6) from a public basic school in Wenchi, Ghana. The selection was based on diagnostic observations indicating significant reading and comprehension deficits.
• Intervention: Three structured reading instructional strategies were implemented sequentially:
  1. Collaborative Strategic Reading (CSR): Small group reading with peer monitoring, vocabulary clarification, and summarization.
  2. Modelled Reading: Explicit teacher demonstration of fluent decoding, pacing, and "think-aloud" comprehension monitoring, followed by independent student practice.
  3. Question-Answer Relationship (QAR): Instruction on locating answers within text, specifically focusing on "Right There" (explicit) questions.
• Data Collection: Data were gathered via teacher-created comprehension tests administered at four distinct occasions: a pretest and a posttest following each of the three intervention phases. Tests included 5 multiple-choice and 5 open-ended questions (30 items total).
• Data Analysis: Performance was categorized into four bands: Below Average (0–14), Average (15–20), Above Average (21–25), and Excellent (26–30). Statistical analysis included descriptive statistics, a Chi-square test of homogeneity, a Cochran-Armitage trend test, and pairwise two-proportion z-tests. To address the issue of multiple comparisons, a Holm-Bonferroni step-down correction was applied. Additionally, Cramér's V was calculated for association strength, and Wilson score 95% confidence intervals were computed for key proportions.

Key Results
The intervention yielded significant improvements in reading comprehension performance, though the magnitude of gain varied across the sequential phases.

• Baseline vs. Final Outcome: The proportion of pupils scoring below average decreased from 60% at the pretest to 14% after the full intervention sequence. Conversely, the proportion of pupils scoring at "Above Average" or "Excellent" levels increased from 20% to 76%.
• Statistical Significance:
  • The overall distribution of performance bands across the four testing occasions varied significantly ($\chi^2(9, N=200) = 38.57, p < .001$).
  • A significant linear improvement trend was observed ($z = 5.16, p < .001$) for the transition from below average to at least average performance.
  • The cumulative improvement from pretest to the final posttest (Posttest 3) was highly significant ($z = 4.76, p < .001$) with a large effect size (Cohen's $h \approx 1.01$).
• Phase-Specific Findings:
  • CSR (Posttest 1): Showed a descriptive improvement (below average dropped to 48%) but was not statistically significant compared to the pretest ($p = .229$).
  • Modelled Reading (Posttest 2): Produced a statistically significant jump at the uncorrected .05 level ($p = .039$), with the below-average group dropping to 28%. However, this specific step did not survive the Holm-Bonferroni correction for multiple comparisons.
  • QAR (Posttest 3): Further reduced the below-average group to 14%. The gain from Posttest 2 to Posttest 3 approached significance ($p = .086$) but was not statistically significant on its own, likely due to a "ceiling effect" where most pupils had already improved.
• Residual Analysis: Adjusted standardized residuals indicated that the pretest had significantly more pupils in the "Below Average" band than expected, while the final posttest had significantly fewer. The intermediate steps (Posttest 1 and 2) did not show statistically significant deviations from homogeneity when analyzed individually.

Key Contributions and Claims
The paper claims that sequential, strategy-based reading instruction can make a significant contribution to early reading comprehension in Ghanaian basic schools.

• Evidence of Efficacy: The study provides empirical evidence that a low-cost, teacher-implementable sequence of CSR, Modelled Reading, and QAR strategies can shift the performance distribution of struggling readers.
• Cumulative Effect: The authors argue that the most robust finding is the cumulative effect of the three interventions taken together. While the study cannot definitively isolate the specific contribution of any single intervention due to the lack of a control group and the application of multiple comparison corrections, the data supports the philosophy that structured, explicit instruction yields better outcomes than unstructured repetition.
• Policy Relevance: The findings offer a model for school leaders and curriculum planners (e.g., Ghana Education Service) to move away from generic remediation toward diagnostic, strategy-based cycles.

Limitations and Modest Claims
The authors explicitly maintain a modest stance regarding the causal attribution of gains to specific strategies:

• Lack of Control Group: Without a control group, the study cannot rule out maturation, practice effects, or concurrent classroom instruction as alternative explanations for the observed gains.
• Statistical Independence: The statistical tests used (Chi-square, z-tests) assume independent samples, whereas the data involved repeated measures on the same 50 pupils. The authors interpret the results as a conservative, aggregate-level test of association rather than a precise individual-level causal estimate.
• Generalizability: The findings are restricted to a single class in one school and a short four-week window, limiting the ability to generalize to other regions or long-term retention without further study.
• Specific Strategy Attribution: The paper concludes that while the sequence of interventions is effective, the data does not conclusively establish the independent statistical significance of any single step (CSR, Modelled, or QAR) in isolation. The significant improvement is attributed to the entire intervention sequence.

Conclusion
The study concludes that structured literacy interventions are a viable and practical approach for addressing reading comprehension difficulties in Ghanaian basic schools. It recommends that teacher professional development focus on these specific strategies and that future research incorporate control groups and longitudinally linked individual data to refine causal claims.