Optimal correction for continuity in the chi-squared test in 2×2 tables

The Family Context in Cybervictimization: A Systematic Review and Meta-Analysis

“…Due to their diversity, the statistics and measures had to be converted to Fisher’s Z -values (Martin-Andrés & Luna del Castillo, 2004), to ensure that they could all be compared.…”

“…In this sense, it is necessary to highlight an aspect that could explain the existence of extreme data, in addition to the diversity of the data. Conversion to Fisher’s Z -values, despite being accepted in this type of methodology (Martin-Andrés & Luna del Castillo, 2004), poses a risk to the values of x > 0.5. This is due to the use of Student’s T curve, which implies that these measures can be distorted, moving away from the mean values, compared to the normal curve.…”

“…The statistical transformation to Fisher’s Z -values (Martin-Andrés & Luna del Castillo, 2004) was carried out using the statistical software CMA (Biostat, United States of America). This program was used to perform publication bias tests (Egger) to guarantee the quality and reliability of the data.…”

The Yates, Conover, and Mantel statistics in 2 × 2 tables revisited (and extended)

“…Subject to Model III, the traditional cc is that of the chi-square statistic of Yates (1934) 𝜒 2 Y (i) = n(| x 1 y 2 − x 2 y 1 | −0.5) 2 ∕{a 1 a 2 n 1 n 2 }. Conover (1974) proposed another cc which leads to the statistic 𝜒 2 C(i) = 𝜒 2 Y (i) + n 3 ∕{4a 1 a 2 n 1 n 2 }; this statistic is less conservative than the statistic 𝜒 2 Y (i) and performs better when the minimum quantity expected is higher than 2 (Martín Andrés, Herranz Tejedor, & Luna del Castillo, 1992).…”

“…Subject to Model III, the traditional cc is that of the chi‐square statistic of Yates (1934) $$$ {\chi}_{Y(i)}^2=n{\left(\mid {x}_1{y}_2-{x}_2{y}_1\mid -0.5\right)}^2/\left\{{a}_1{a}_2{n}_1{n}_2\right\} $$$. Conover (1974) proposed another cc which leads to the statistic $$$ {\chi}_{C(i)}^2={\chi}_{Y(i)}^2+{n}^3/\left\{4{a}_1{a}_2{n}_1{n}_2\right\} $$$; this statistic is less conservative than the statistic $$$ {\chi}_{Y(i)}^2 $$$ and performs better when the minimum quantity expected is higher than 2 (Martín Andrés, Herranz Tejedor, & Luna del Castillo, 1992).…”