Estimating Missing Data in Balanced Incomplete Blocks from Connected Designs
Henry Mendoza R. & Luís Alberto López
Abstract
Missing data in experiments cause loss in the original structure and therefore in the orthogonality of the experimental design . This paper estimates missing data in an incomplete connected balanced block design type ICBB, applying the Bartlett’s covariance method; a correction of the analysis of variance is done, and finally the results are applied in experimental data realized in incomplete blocks design with repetitions. The obtained results show that the corrected analysis of variance depends on the observed data and estimation of the missing data.
Key words: Covarianca Analysis, Incomplete Blocks, Missing Data, Within Block Analysis, Connected Designs.
PDF