Item Reliability of the Movement Assessment of Infants

To examine the item reliability of the Movement Assessment of Infants, 53 high- and low-risk 4-month-old infants were tested. Two examiners rated performance independently to provide an estimate of inter-observer reliability. A sub-sample of 29 infants was again assessed within one week of the initial session. The scores of the same observer over two repeated testings provided and estimate of intra-rater reliability. Percent of agreement and the Kappa statistic were used to estimate inter-observer and intra-observer reliability. Using the Kappa statistic guidelines, 2% of the items demonstrated excellent inter-observer reliability and 58% had fair to good inter-observer reliability; 10% of the items demonstrated excellent intra-observer reliability and 42% had fair to good intra-observer reliability. Agreement and stability of the items were also reported by the index of percent of agreement to help evaluate items with good agreement, yet low variability of response categories. Items that had acceptable reliability and items that were considered problematic in administration and scoring were identified. Factors were discussed that may affect the reliability of individual items that were scored by a single examiner on only one test administration.