Reactor power affects the temperature of fuel and coolant of a nuclear power plant. The change in temperature of fuel and coolant modify the reactivity of a nuclear reactor. The change in reactivity due to the change of reactor power is known as power defect. The power defect depends on the various parameters such as reactivity coefficients due to thermal hydraulics feedback, the response of fuel and coolant temperature due to variation of reactor power, etc. The reactivity coefficients are significantly varied due to different operating conditions, changes in fuel characteristics and fuel burn-up during fuel residence time inside the reactor. This wide variation of reactivity coefficient are in general technically specified in the form of lower and upper bound for safety analysis purpose. A thought experiment has been carried out considering those reactivity coefficients contain stochastic variability and ignorance (i.e., lack of knowledge). The uncertainty involved in reactivity coefficients are captured by defining them with probability box (p-box). After propagating the p-box of reactivity coefficients through the theoretical model of power defect, the p-box of power defect has been generated. In the pinching method, one of two reactivity coefficients will be fixed at their average value and observation on the change of area of p-box of power defect has been made for sensitivity analysis. Based on the reduction of area of p-box of power defect, the sensitivity of these two reactivity coefficients has been analyzed. The parametric studies of variation of sensitivity for five different power drops (i.e., $$10\%$$10%, $$25\%$$25%, $$50\%$$50%, $$75\%$$75% and $$100\%$$100%) have been studied and quantified in this paper. It is found that the reactivity coefficient due to coolant temperature is more sensitive than reactivity coefficient due to the fuel temperature on power defect. It is also found that the sensitivity does not depend on amount of power drop.