You can append these answers directly onto the "accumulated data" file if you use the proper nomenclature, then import using the "Paper-and-Pencil Data" option. So if you have a grid question called "direct" with 18 rows, one per MaxDiff item, that were given ratings on a 1-5 scale, your accumulated data file would have MXD_1_b, MXD_1_w,..., MXD_18_b, MXD_18_w, direct_r1, direct_r2,..., direct_r17, direct_r18. If in doubt as to what the "direct" variables are named in the software, you can use the "print study" function in the file menu to print a graphical PDF that will display the variable names. Or, alternately, if you're familiar with HTML and don't mind doing a little digging, you can preview the "direct" question, then right-click and view source. Be aware that this is case sensitive, and won't work if your variable names don't match ("direct_r3_c1" instead of just "direct_r3", for instance, or capital D "Direct" instead of lowercase).