The results of

random-effect model (pD=20.9 and DIC=23.3) were similar to the fixed-effect model (pD=15.9 and DIC=67.7).

If significant heterogeneity was present (P < 0.1, [I.sup.2] > 50%), a

random-effect model was selected.

Next, Columns (7) through (9) report results from the application of fixed-effect model, Columns (10) through (12) provides a similar report from the estimation of a

random-effect model, and, finally, the last three columns do the same job for a random-coefficient model, where coefficient randomness is assumed to apply only to the coefficient of our concern, the level of corruption (cp i).

A

random-effect model was used to summarise the results when the heterogeneity was high or medial, otherwise a fixed-effect model was used.

Heterogeneity was measured using [I.sup.2].18 When [I.sup.2] is less than 50% and p>0.10, the results were considered homogeneous and the fixed-effect model was used; when [I.sup.2] is greater than 50% but less than 75%, results were considered heterogeneous and the

random-effect model was used.

A fixed-effect model (Mantel-Haenszel) was applied in case of absence of heterogeneity between studies and otherwise a

random-effect model was performed.

Homogenous data was calculated using the fixed-effect model, and

random-effect model was employed when there was statistically significant heterogeneity.

Two approaches are developed to capture the unobserved heterogeneity: fixed-effect model and

random-effect model. The fixed-effect model examines cross-sectional and/or time-serial difference in the intercept term.

Lagrangian multiplier test statistic of 21.78 and p-value = 0 supports the preference for the

random-effect model rather than for the pooled model.

In a

random-effect model, there is an extra factor added in the equation to control the country-specific random element.

According to the

random-effect model, each respondent bases both the initial and the follow-up questions on a WTP amount that has mean value equal to the respondent's true WTP; but the actual WTP amount used when responding to a particular payment question is subject to random error.