A field experiment on cotton yield resulted in a non-statistically significant interaction. An approach for follow-up examination between treatments based on least significant difference values was suggested to identify the effect regardless of insignificance. It was found that the classical formula used in calculating the significance of interactions suffers a possible shortage that can be eliminated by applying a suggested revision.
Sawan, Zakaria M.
"An Approach for Dealing with Statuses of Non-Statistically Significant Interactions between Treatments,"
Journal of Modern Applied Statistical Methods: Vol. 12
, Article 24.
Available at: http://digitalcommons.wayne.edu/jmasm/vol12/iss1/24