Eric Marder wrote this book much before consumer behavior books atarted coming in like anything during this decade.

Highlights from the book:

The main concern of Choice Research is the congruence between the choice situation in the questionnaire and the choice situation in the world.

... it doesn't matter whether the choice consists of voting in a booth or voting in a survey. It doesn't matter whether it consists of selecting a brand from a shelf or electing a brand while playing a game in a questionnaire. Provided we can keep the competitive frame, the information and the accessibility the same in the questionnaire as in life, the choices in the questionnaire will be the same as the choices in life. By saying that the information in the questionnaire is the same as the information in life, I mean that the totality of what the customer knows about the brands at the time of making choices in the questionnaire is the same as the totality of what she knows about the brands at the time of making the choices in the supermarket. The central idea is simple enough, possibly self-evident.

The three 'laws' of choice:
1. The laws of congruence: Congruent choice situations have congruent choice vectors.
2. The law of primacy: An individual for whom, at the moment of choice, n brands are tied for first place in brand strength, chooses each of these brand with probability 1/n.
3. The law of persistence: The effect produced by a message is made up of two components: a transient effect and an intrinsic effect. The transient effect decays rapidly. The intrinsic effect lasts indefinitely.

Based upon these three laws, there are three specific market research techniques which enable the researcher to determine choice accurately.
STEP (Strategy Evaluation Program) - This research methodology follows directly from Marder's first law, and is used to determine how the likely market share of a product may differ under alternative branding strategies. In this two-step market research methodology, respondents first are given a booklet with information about all possible brand choices, and asked a few questions about the extent to which they like each. They then are given a certain number of stickers, which they allocate to the brands, giving s many or as few stickers as they want to each one. The share of all stickers that each brand receives when the results for all respondents are totaled up correlates to its actual market share. By varying the information that respondents get about the particular brand under consideration (and keeping the information provided about all other brands the same) the most effective branding strategy can be determined.
VEST (Volume Estimation Test) - This is a method for estimating the likely market penetration of new or repositioned products based upon the ownership of items in other product categories that are not substitutable for one another (e.g. cell phones, toasters, VCRs). Respondents are presented with a list of these non-substitutable items (including the new product to be tested) and asked to indicate whether they would like to own the product, and whether or not, in fact, they do own the product. The results are then tabulated, and an adjustment made for differences between the 'want to own' and 'actually own' scores for all products except the subject product. The 'want to own' score for the subject item is then adjusted by this factor, to arrive at an estimated market penetration rate for the product.
SUMM (Single Unit Marketing Model) - This is a procedure to estimate the attractiveness of various product attributes, and their contribution to overall brand choice on the part of an individual. A set of mutually exclusive product attributes (that are meaningful when considering the type of product being assessed) are listed, with various categories within each identified. (For example: 'colour' might be identified as one product attribute, and various sub categories within it specified: red, green, blue. 'Shape' might be another product attribute, and given the options square, round and oval. And so on, where as many as ten or twelve such product attributes are specified.) Respondents are then asked to select their preferences for each product attribute, allocating 100 points among the choices for each attribute (e.g. they may give 60 points to 'blue' for colour, and 20 points each to'green' and 'red', indicating a strong preference for blue'). Next they indicate whether each brand in the set of products being considered reflects each possible sub-category of the product attribute, scoring a simple 'yes' or 'no' on each (e.g. they may indicate a belief that believe that 'product X' is available in blue only). Summing and weighting the results of this analysis for an individual allows a prediction of the brand choice of that individual in the marketplace, and the cumulative results of this analysis for many individuals enables overall market share estimates to be made. However, the real use of this technique is in determining how manipulating the attributes of the product or service in question will impact on market share - a very powerful technique indeed!

The various 'principles' of Choice Research (28 in all)
The Selling Strategy Principle: Explicitly or implicitly, every brand has a selling strategy that consists of the same universal syllogism: Because my brand has attributes A,B,C,...... it will satisfy your desires X,Y,Z,...better than any other brand.
The Price Effect Principle: On the average, a price increase of 10 percent will produce a share decrease of around nine percent, but there is a great deal of variability in this result. One time in five the loss will be much larger, and one time in five there will be no loss at all.
The Name Principle: A name is worth money. For durables, a good name may permit charging as much as twenty percent more for the brand, on the average; in some cases, as much as fifty percent more. It can also be of comparable value for consumables, but only rarely.
The Image Principle: A brand's future is built on its past. It is easier to give a brand the right image in the first place than to change a wrong image once it has taken hold.
The Laws of Choice is a difficult and highly technical book at times, but contains some extremely useful insights and research techniques. It should be required reading for all market researchers.

[From the Great Books Series. Also see The Success Manual  - Encyclopedia of Advice, which contains summaries of 100+ Most useful books.]