This paper studies the impact of investor attention to oil prices on returns, volatility, and covariances of three exchange traded funds representing oil, gold, and the stock market. For this purpose, we suggest a new multi-variate volatility model based on open, high, low, and closing prices that incorporates the impact of investor attention on returns, volatility, and covariances. We find that this model, which incorporates Google searches for "oil prices" as an exogeneous variable, outperforms other considered multivariate volatility models, and dem-onstrates that Google searches for "oil prices" can explain and forecast covariances between returns of oil, gold, and the stock market.