A critical assessment of potential measurement biases in the technology acceptance model: three experiments