Modeling human response time data with an exponential model: model comparison issues

Is this still an issue?