The competitive market model as commonly described in textbooks includes a number of assumptions that are thought to be necessary to reach the efficient allocation of resources and stable price predicted by the model. Experiments have demonstrated that most of these assumptions are much stronger than conditions that lead reliably to effiecient outcomes.
The "homogenous good" aspect of perfect competition is the norm in the laboratory. In fact, it generally takes a good deal of design to implement differentiated products in the lab.
Perfectly Divisible Good
Most laboratory experiments involve goods which can be traded only in integer amounts. This variance with the ideal situation of perfect competition is rarely very important, except that it makes the math more challenging in approaching laboratory environments. In many cases, the discontinuity can mean there is not a single equilibrium price, but a range of prices.
Whether or not perfect information appears in the laboratory depends on the question, "information about what"? It is common in many laboratory experiments to make the prices for trades public information. In this sense, laboratory information is more complete than information in many real-life markets. On the other hand, the norm in laboratory experiments is that information about private costs and values is not made public. Anecdotally, making more of this sort of information available can make markets less efficient.
No Transactions Costs
Some small amount of transactions costs are inevitable in any transaction, but the norm in laboratory experiments is to have very small transactions costs. Occasionally, when the "no transactions costs" seems particularly important, experimenters have subsidized trade by paying subjects a small flat fee for taking part in a trade, in addition to their normal earnings.
This is the norm in market experiments, unless the experiment is specifically examining externalities
Free Entry And Exit
This is clearly not the norm in laboratory experiments. Free exit is normally possible, simply in terms of experimental subjects deciding not to take part. Free entry is normally not possible, since a typical laboratory experiment has a specifically limited number of buyers and sellers.
This difference between the theory of free entry and the laboratory can be seen in that sellers in experimental markets typically do make some profits. This result can be emphasized in class discussions (along with arguments about whether it was "normal profit").
Large Number Of Buyers And Sellers
It is commonly difficult to make an experiment with a large number of buyers and sellers, each "small" relative to the entire demand or supply in the market. How much this matters in practice depends on the particular environment and market institution. There is one famous result that "four is a large number" in a double auction market.
Analytically, price taking is sometimes used as an assumption, while sometimes it is considered a result of other assumptions (such as free entry and exit). In the laboratory, it is worthwhile to examine with a class when price-taking is a good assumption. After experiments, ask class members if they were price takers. Alternatively, did they withhold demand or supply?