Application of Meta-Analysis to Yield Performance of Rubber (Hevea brasiliensis) Clones
- 1 Downloads
In Hevea, so far no attempt has been made to conduct a meta-analysis on yield performance of clones from many studies. With the objective to demonstrate the application of meta-analysis in Hevea research, yield performance of twelve clones repeated across many trials were subjected to meta-analysis. Two effect-size measures namely Standardised mean difference as Hedge’s g and log response ratio were used for combining results. Standardised mean difference in different studies ranged from −3.76 to 6.83. Random effect mean standardised mean difference in GT1 was 0.38. All the standardised mean difference values in RR1M 600 were negative. In clones RRII203, PB 235, PB 260, PB 217 and PB 311, standardised mean difference values were all positive while in other clones they were mostly negative. Random effect model mean standardised mean difference’s were significantly higher in clones RRII 203 and PB 235. In effect sizes, the percentage change in yield of GT 1 over RRII 105 ranged from negative 13 percent to positive 39 percent.
Among eight studies, four studies gave positive while the other four gave negative change in yield over the control RRII 105. Random effect overall yield was about 10 percent. In RRII 600, all the six studies produced negative change in yield. RRII 203 and PB 235 produced significantly higher yield over RRII 105 with percentage increases of about 20 and 36 percent respectively. Among the other clones, PB 260 produced a significantly higher yield of about 22 percent. Clear general patterns were identified in yield performance of treatment clones over the control. Clones RRII 203, PB 235 and PB 260 showed clear superiority in yield performance. Results of clones PB 217 and PB 311 indicated that further testing needs to be undertaken for taking a clear stand on the usefulness of these clones for the region. Comparison of standardised mean difference and the log response ratio effect size indices indicated that the log response ratio maybe more informative for quantifying the treatment effect in meta-analytic studies with Hevea brasiliensis.
KeywordsNatural rubber Hevea braseliensis meta-analysis yield performance effect size standardised mean difference log ratio of means random effect mean heterogeneity continuous data
Unable to display preview. Download preview PDF.
- 1.SIMMONDS, N.W. (1979) Principles of Crop Improvement. England: Longman Scientific and Technical, 408.Google Scholar
- 2.SIMMONDS N.W. (1989) Rubber Breeding. In: Rubber. (C.C. Webster and W.J. Baulkwill eds.). England: Longman Scientific and Technical, 85–124.Google Scholar
- 3.VARGHESE, Y.A. AND MYDIN, K.K. (2000) Genetic Improvement. In: Natural Rubber, Agromanagement and Crop Processing (P.J. George, and C.K. Jacob, eds.) Kottayam: RRII, 36–46.Google Scholar
- 4.LIPSEY, M.W. AND WILSON, D (2000) Practical meta-analysis. Thousand Oaks: SAGE Publications, 264.Google Scholar
- 6.COOPER H.M. (2009). Research synthesis and meta-analysis: A step-by-step approach 4th Edn. Thousand Oaks: SAGE Publications, 280.Google Scholar
- 14.HEDGES, L.V. AND OLKIN, I. (1985) Statistical Methods for Meta-Analysis. Orlando: Academic Press, 378.Google Scholar
- 20.COOPER, H., HEDGES, L.V. AND VALENTINE, J.C. (2009) The handbook of research synthesis and meta-analysis, 2nd Edn. New York: The Russell Sage Foundation, 610.Google Scholar
- 22.COHEN, J. (1988) Statistical Power Analysis for the Behavioral Sciences. New York: Routledge Academic, 579.Google Scholar