Includes bibliographical references

Simple and piecewise linear regression methods were used on cross-country as well as pooled-cross-country data to estimate the relationship of enrollment ratios to per capita GNP growth. Dummy intercept shifters were used for countries of the OPEC (Organization of Petroleum Exporting Countries) and East Asian Countries (EAC). Dummy variables were also used for the threshold values of the enrollment ratios and for the periods in the pooled regressions. Results showed that: (1) The positive relationship of secondary enrollment ratios to per capita GNP growth were stronger and their estimated regressions better fitted than those of primary ones. (2) The intercept of the estimated regressions - throughout the analysis - shifted downward once the OPEC dummy variable was introduced into the model and shifted upward when the EAC dummy was included. (3) The effect of the enrollment ratios up to the threshold values was much more powerful on economic growth than was that beyond the threshold values. (4) Pooled estimated enrollment/ economic growth regressions showed some negative structural change in the periods 1980 and 1992 as compared to the reference period 1950.