المقارنة بين طريقتي التعويض بآخر قيمة ملاحظة وهوت دك لمعالجة البيانات المفقودة في ضوء معيار قوة الاختبار الإحصائي: دراسة محاكاة

الملخص EN

The study aimed to identify the most efficient ways to compensate for missing data.

this is by comparing the two methods of compensation with the Last Observation Carried Forward and hot deck Imputation to estimate the missing data.

To achieve the study goal.

the simulation was used to generate random numbers of different sample sizes (30+ 60+ and 90).

A simulation model of One-Way Analysis of Variance method was built (if sample sizes are equal between groups and sample sizes are not equal between groups) and assumptions are available for One Way Analysis of Variance method.

The statistical strength of the full model of these cases was calculated.

This data was then lost by 15%* 25% 40% respectively according to the MCAR mechanism and under the pattern of loss (general pattern).

The lost values were then estimated using the two methods of compensation.

Then conduct the One-Way Analysis of Variance test for estimated models and calculate the statistical strength of those models for the cases mentioned.

The two methods of compensation are then compared with the Last Observation Carried Forward and hot deck to estimate the missing data by comparing the statistical strength of those models estimated by the strength of the full statistical model.

The results of the study showed that the compensation method for (LOCF) is more efficient than the hot deck.

The LOCF method gives estimated values that lead to results closer to the full model according to the model's statistical test strength standard.