We use data from a randomized controlled trial in Ethiopia and examine the causal effects of HIV/AIDS education, home‐based voluntary HIV counseling and testing (VCT), and conditional cash transfers (CCT) for facility‐based VCT on HIV/AIDS knowledge and demand for HIV testing. HIV/AIDS education significantly increases HIV/AIDS knowledge but has a limited effect on testing take‐up. However, when HIV/AIDS education is combined with either home‐based VCT or CCT for facility‐based VCT, take‐up increases substantially by about 63 and 57 percentage points, respectively. We also demonstrate evidence of persistence in test‐taking behavior, where past HIV testing does not dampen demand for testing. Lastly, we find suggestive evidence that home‐based VCT could be more effective at detecting HIV‐positive cases relative to CCT for facility‐based VCT. Our findings highlight the importance of geographic accessibility in the testing decision and persistence in demand for HIV testing. Copyright © 2016 John Wiley & Sons, Ltd.