You can  use both but it is very important what you want to score. Another thing is whether you score large or small bowel.

In general, I use the  section for scoring where the villae and crypts are seen in a longitudinal manner. Thus one can evaluate the markers in all cell there including Panet cells in the small intestine and crypt base in the large bowel. Not to mention that in a longitudinal structure section, one can also evaluate the submucose glands in duodenum as well as other structures like the plexuses for ex.

Depends on what you want to score and which markers you use. Most IHC studies use semi-quantitative scoring systems which evaluate the "force" of staining (none, weak, moderate, strong) and the prevalence (amount of positively stained cells and\or areas). There are a lot of different approaches. Look for IRS score, for example. 

What are you trying to analyse? Without that information, it is like asking "how long is a piece of string"? It depends...............

Thank you for all the answers. I am doing IHC on colon samples. In my lab, IHC scoring is done manually. To determine whether there is an overexpression or underexpression of a protein, the staining intensity of an antibody is compared between tumor tissue and the normal epithelial tissue. The scoring of intensity of tumor tissue is straight forward but when it comes to normal epithelial part, there is a problem. Ivan's answer was helpful but what if only cross section existed in the sample, how should I score?

What IHC markers do you use? Do you perform all stages of the experiment or only IHC? Actually what you are discussing is a question of proper tissue orientation in a paraffin block during embedding stage. For further information see: Manual of surgical pathology / [edited by] Susan C. Lester ; illustrators, Christopher A. French, Shogun G. Curtis. -- 3rd ed. It describes all major steps of proper tissue handling. 

Thank you for your answer Nickolay. IHC markers I used includes beta catenin and APC antibodies. I am studying on wnt signaling. IHC is just part of the study. My problem is scoring the staining after IHC. I am only interested in how to score if only cross section existed in the sample. My samples are old FFPE block which I have no control of the tissue orientation. Is there a paper or anyone who is experience that can help?

Make a simple scoring system. When there is no guideline for scoring I do the following:

I plot area and intensity:

Area in percentage 10-30%< - 2, >30-50%< - 3, >50-80%< - 4, >80%-5

Intensity: 0(no immune reactivity), 1 (detectable but faint or patchy in the cell), 2 (weak homogeneous or moderate irregular), 3 (moderate homogeneous or strong patchy), 4 (strong homogeneous)

Then plot area by intensity and sum the different results

1(area) x 4(intensity) + 2(area)x3(intensity) +4(area)x0(intensity)=10

It works for me :)

Good luck