Lewis dot structures or VSEPR structures, and those that require hefty amounts of computer time to set up and solve (ab initio MO theory). The text we are using includes many full color diagrams of bonds, but student still struggle with how these two dimensional representations "work" in three dimensions.
Ad hoc models — mock ups of molecules built by hand from mundane materials such as cardboard and wire — have a venerable history in chemistry. Watson and Crick used cutouts of the bases to figure out how they paired along the helix; Smalley built a paper model of pentagons and hexagons to see how C60 could be constructed.
So last week, I brought paper, tape and some simple molecular models (tubes and small metal centers which I buy by the bag to hand out to students) to class and asked students to build a valence bond model for acetaldehyde (implicated in hangovers - acetaldehyde, not valence bond theory, though the latter certainly can make students queasy). Students cut, paste, built and discussed, producing what you see in the slideshow.
With thanks to Danqui Luo, Tess McCabe, Kai Wang, Ben Kaufmann and all the students of Chem 103 who built and photographed the models.
Sixty-four years later: How Watson and Crick did it
13 hours ago in The Curious Wavefunction