Recommendations in line with full disclosure (see, for example, http://www.ehps.net/ehp/index.php/contents/article/view/777/ehp.v14.i4.p77): Raw data should be stored. For each publication, also the final data (on which the reported analyses are based) should be stored. Syntaxes should be made available as part of a replication package that allows others to replicate the steps you made in the analyses.
How syntax files can be made available differs per software package. For example, when using R, the syntax files (i.e., having the file name extension .r or .rd) provide an overview of all the steps taken during analyses. It is highly recommended to document these steps (e.g., by using comments) so both yourself and others understand decisions/steps that were made during analyses (an example: http://sciencerep.userfriendlyscience.com/files/11/rcode.r). When using SPSS, the syntax files can also be saved separately (i.e., having the file name extension .sps). Also here you can add comments to document the decisions/steps taken during analyses. Moreover, syntax commands can also be automatically added to the output files (.spv-files) in SPSS (tick the box ‘Display commands in the log’ under Options > Viewer).
Besides data files in their original format, it is warranted to also supply them in non-proprietary format, such as plain text, Open Document Format (ODF), Portable Document Format (PDF), or Hyper Text Markup Language files (HTML), or, for images, Portable Network Graphics (PNG) or Scalable Vector Graphics (SVG). This is important because opening proprietary formats require the purchase of specific software, which other researchers may not have. Syntaxes (e.g. for SPSS, SAS, or R) are provided in their original format, because these generally already are in plain text format.