The application of one-way analysis of variance by matrix, tacking into account replicate analysis, followed by post hoc tests when differences were relevant, showed for all matrices that method 3 is almost always in the group giving the higher carotenoid
content; exceptions were lutein in NIST SRM 2383, kale, leaf beat, and tomato, and -carotene and zeaxanthin in the in-house baby food control sample.