Just google "t table" and lots show up for the p value for different t values given the df (and the t can be calculated from mean, sd, and n as shown in intro books), or use the functions available in most statistics software. For example, in the freely available R environment, it is the function pt( ).