Purpose The study aims to assess the moderating role of celebrity characteristics in the relationship between celebrity endorsement and telecommunication companies’ reputation. Design/methodology/approach Analysis of results was based on 700 customers in the telecommunication sector. Confirmatory factor analysis was conducted to check for validity and reliability of the observed items. A hierarchical regression model was estimated to test the various hypotheses set for the study. Findings The study finds that celebrity endorsement in itself had no significant effect on the reputation of telecommunication companies. Celebrity attractiveness, likeability and trustworthiness had a direct effect (positive) on the reputation of telecommunication companies and positively moderated the effect of celebrity endorsement on telecommunication company reputation. Celebrity expertise had no direct effect on telecommunication company reputation but positively moderated the effect of celebrity endorsement and telecommunication company reputation. Research limitations/implications This study was purely quantitative. Future study could consider a mixed approach and include senior management members of the telecom firms for an in-depth interview. Practical implications In signing on celebrities as brand ambassadors, management must pay particular attention to celebrity attractiveness, likeability and trustworthiness. This would be more rewarding to the firms. Originality/value The study adds to the little empirical knowledge available on celebrity endorsement in sub-Saharan Africa and telecommunication sector in particular.

Controls are coordinated horizontal or vertical positional data for land and geographic information systems forming a framework to which surveys are started, referenced and adjusted. Controls on KNUST campus were established over three decades ago with new ones being added as and when t h e y a r e required. Conventional survey methods like trilateration, triangulation and traversing were used and are still being used for densification of controls. These approaches involve the use of low precision instruments such as analogue/optical theodolites and leveling equipment. Since the establishment of these controls, t h e y have not been investigated b u t t h e y are being used for p r o j e c t development. A modern method that uses satellite positioning techniques such as GPS is currently in operation and has numerous advantages in the establishment of control networks. GPS control surveys were carried-out on seven controls and ellipsoidal coordinates were obtained in the World Geodetic System 84 reference frame. The Cartesian coordinates were projected onto the Universal Transverse Mercator frame. A two dimensional conformal transformation was done using existing KNUST boundary coordinates to Ghana National coordinates based on the War Office ellipsoid. Precise Level routines were carried out on the seven controls for the computations of orthometric heights. The method of least squares adjustment, root mean square errors (RMSE), standard errors (SE) and residuals derived were used to analyse the differences in horizontal positions and heights of the controls. The linear displacement between the computed and the existing coordinates were within the range of 0.015m to 0.014m. The RMSE were 0.048 and 0.106, whiles the SE were also 0.057 and 0.125 for the northings and eastings respectively. The computed and existing heights differed between-0.075m and-0.004m with a mean downward movement of 0.011m.