ความชุกความไม่มั่นคงทางอาหารของประชากร โดยใช้เกณฑ์การวัดประสบการณ์ความไม่มั่นคงทางอาหาร

43 7. Number and % of cases with extreme raw scores (0 or highest number of scale items) ___score 0 – 10759 cases (84.55%) and score 8 – 68 cases (0.53%)__ 8. Sample size used to perform statistical validation to check the quality of the data After any items may be dropped due to a high percentage of missing responses (point 6 above), this is the original sample minus incomplete and extreme cases. It is important to ensure that a large enough number of cases remain after those with incomplete responses or extreme raw scores are omitted in order for statistical validation to be reliabl e. 2 _______________1898____________________ Statistical validation Application of the Rasch model to assess the performance of the FIES in the present context and check the quality of data collecte d 3 : Summary of output: item parameters and fit statistics Item Item parameter Standard error Infit Outfit Proportion of affirmative responses Number of affirmative responses WORRIED -2.834 0.030 1.259 3.405 80.45 1527 HEALTHY -1.153 0.030 0.956 1.111 41.15 781 FEWFOOD -1.387 0.029 0.804 0.788 46.63 885 SKIPPED 1.072 0.057 0.954 0.811 12.49 237 ATELESS -0.258 0.036 0.935 0.813 27.19 516 RANOUT 0.276 0.042 0.896 0.759 22.44 426 HUNGRY 1.216 0.060 0.933 1.038 11.8 224 WHLDAY 3.355 0.138 1.018 0.584 1.9 36 2 See Lesson 3, slide 43 of the FIES e-learning course for sample size requirements for statistical validation. 3 Th e RM.weights software f or analyzing FIES data is available free from FAO.