The key to making more accurate measurements is to slow down the rate of changing temperature of the probes. The slower the probe's temperature increases, the more accurate is the basic assumption that the probe is in a state of thermo equilibrium..
Sand Cryostat
Heat loss will be primary though the top surface of the sand. A way to help reduce this is to use some fiberglass insulation. A little insulation between the top of the sand, and the lid of the cryostat will make quite a difference. One needs to take extra precautions if using the fiberglass insulation with liquid nitrogen. It is very easy for beads of liquid nitrogen to become trapped within the fiberglass. If the fiberglass is handled with bare hands, it could appear to be at room temperature, but if squeezed, would force the liquid nitrogen out, like water out of a sponge
Four Point Probes
The four point probes can also be improved by very carefully bending the cable so that the open face of the probe is parallel to the top surface of the sand. The reasoning behind this is that the superconductor pellet thickness is approximately 5mm. Hence any gradient in the sand's temperature that corresponds to a change in depth, will be minimalized.
Ice Water Bath
By using two separate containers, one placed inside another, a more constant temperature of 0ºC can be achieved. The inner container would have crushed ice, water and the reference junction. The outter container, would consist of larger cubes of ice, and just enough water to float the ice. It is important that the inner container is as completely surrounded by the ice as possible. Ideally this means both on the sides, and the bottom. If possible, it is best to use a piece of Styrofoam to cover the top, as to insulate your bath from the outside room temperature.
Two point probes (thermocouple probes)
These are the most difficult probes to try to insulate. Unlike the Four Point Probes that can be buried out of sight in the sand, these require the user to visually watch and observe the temperature at which the Meissner effect disappears. The best method for performing this experiment, is to use a small thick plastic pyrex petri dish. Place the probe in the center of the dish, then using large grained washed sand, completely surround the probe, such that the top of the probe is level with the top surface of the sand. Slowly add the liquid nitrogen to the sand, until the sand and probe are complete immersed. Carefully place the small cylindrical magnet over the probe, then place a small clear plastic cup over the probe, magnet, and some of the sand The edges of the cup can be buried in the sand somewhat. Just as long as there is still space between the lip of the cup and the bottom of the dish. The reason for the space would be to allow the nitrogen gas to slowly seep out, instead of building up.
Information Courtesy of CSI Superconductors