French term (edited): manipulé

rather than "handled"

experiment is fine for "manipulation"

one of the definitions of manipulation is "engineering" in the sense of manoeuvering or manipulating things

"used" or handled as you suggested for "manipulés"

"operation" or experiment as you suggested for "manipuation.

Another idea for you !

Just a thought - could manipulation actually mean manipulation ? There is some work on "manipulation marketing"... see here
http://www.marketing-etudiant.fr/forum/le-marketing-cree-t-i...
However, experiment feels right to me.

French term (edited): manipulation / manipulé

manipulation --> experiment (= I agree with your own propsal)

Les trois facteurs manipulés
-->
The three factors in play

webref::

Glossary A 3 level design that starts with a 2 level fractional factorial and some ... one or more factors in play that are preventing an organisation reaching its ...
www.seglainternational.com.au/glossary/c/ - 26k - Cached - Similar pages

--------------------------------------------------
Note added at 12 hrs (2008-06-22 12:44:28 GMT)
--------------------------------------------------

The above webref - which is a glossary of statistics and other terms - also defines 'factor':

An input to a process that can be changed via experimentation.

This confirms that a 'factor' is a 'variable' (as used in other areas of mathematics) - but that the correct term in this kind of experimental study is 'factor'.

Several other entries in the glossary use 'in play' and experiment

--------------------------------------------------
Note added at 14 hrs (2008-06-22 14:41:49 GMT)
--------------------------------------------------

Final comment:

'Factor' in this context refers to the pre-set conditions for each individual experiment; the factors define the context in which observations are made. The factors do not change while the experiment is being executed, so they are not 'variables' (dependent or independent).

However, the factors may be represented as independent variables in the subsequent statistical analysis of a full set of experiments (8 sets in this case), with a view to showing how consumer response (or whatever...) is dependent on those factors.

So, it is correct to refer to 'factors' when designing the experiment - and 'independent variables' when doing the statistical analysis.

If the source text makes this important distinction (which it clearly does) then so too must the translation.

French term (edited): facteurs manipulés

The three independent variables were manipulated on two levels...

I agree that "experiment" works fine in the second instance.

--------------------------------------------------
Note added at 2 days21 hrs (2008-06-24 21:30:02 GMT)
--------------------------------------------------

Independent variables are the elements in the design of an experiment (not merely in the subsequent statistical analysis) which are systematically varied by the researchers in order to observe possible effects on dependent variables. Each of the items in question here (familiarity, quality, and "mentions RVP", whatever those may be) is being systematically varied (manipulated) between two levels; any scholar conducting quantitative research will affirm that these are independent variables.

[As a side note and contrary to Mediamatrix's assertion at 14 hrs, the researchers have not laid out eight individual experiments here but rather one experiment with eight conditions, one for each possible combination of the independent variables.]

In factorial design, it's not inappropriate to refer to the independent variables as "factors", but its also perfectly correct to refer to them as "independent variables" (for evidence of which see below). In terms of translating the phrase "facteurs manipulés" in this context, "independent variables" is precise, unambiguous, and stylistically appropriate, and it allows the translator to subsequently use the word "manipulated" for a natural-sounding translation of "le sont".

------
Some instances of the term "independent variables" in published marketing papers:

"To test these two hypotheses, a 2 x 2 factorial design was used. That is, each independent variable (price and store of purchase) had two levels, high and low."

Stafford, James E. and Ben M. Enis (1969). "The Price-Quality Relationship: An Extension". Journal of Marketing Research (6), 456-458.
---

"The communications were featured in a 2 x 2 factorial design, with two levels of source credibility (high and moderate) and two levels of the consumer situation (buy and lease). The success of these manipulations was checked by administering measures of perceived credibility and financial risk associated with buying or leasing. The critical dependent variables were subjects' attitudes toward the product and their behavioral intensions....

"Independent Variables

"Source credibility was varied through the description of the communicator presented at the outset of the message. ...

"The manipulation of the dimensions of the buy-or-lease situation was presented in the role-playing instructions. ..."

Harmon, Robert R. and Kenneth A. Coney (1982). The Persuasive Effects of Source Credibility in Buy and Lease Situations. Journal of Marketing Research (19), 255-260.
---

"The experimental variables in our study were the availability of an MDSS (yes/no) and the amount of time pressure (low / high). In the 2*2 factorial design, we created four experimental groups, consisting of 20 subjects each. The measure for cognitive style, i.e., field dependence, was treated as a covariate. ...

"For each subject, we measured the net marketing contribution or profit (PROFIT) as the performance variable. Since four periods were played, we have four (repeated) measurements for profit. We measured three independent variables:

"1. Use of a marketing decision support system (MDSS). ...

"2. Cognitive style (FIDE). ...

"3.Time pressure (TIPR). ..."

Van Bruggen, Gerrit H., Ale Smids, and Berend Wierenga (1998). Improving Decision Making by Means of a Marketing Decision Support System. Management Science (44:5), 645-658.
---

"We tested the hypotheses by using a 5 x 3 x 3 between-subjects factorial design (Figure 2) with five price levels (low, medium, high, too high, and absent), three brand levels (low, high, and absent), and three store levels (low, high, and absent). ...

"At the end of the experiment, subjects evaluated the price (very high to very low), brand name (very high quality to very poor quality), and store name (very high quality to very poor quality) on 7-point rating scales. Analysis of the manipulation check mean scores suggested that the manipulations of the three independent variables were perceived as intended."

Dodds, William B., Kent B. Monroe, and Dhruv Grewal (1991). Effects of Price, Brand, and Store Information on Buyers' Product Evaluations. Journal of Marketing Research (28), 307-319.
---

"Independent Variables

"In the experimental conditions, we manipulated four factors based on our hypotheses development (H3-H6): ( 1) the extent of mass customization (number of modules and number of levels per module), ( 2) the level heterogeneity, ( 3) the individual pricing of modules, and ( 4) the type and availability of a default version (for an overview of these factors and their levels, see Table 1). We also included a predefined part in the experiment that served as a baseline evaluation in the model. We measured consumer expertise using five aspects of consumer expertise about PCs (i.e., knowledgeable, competent, expert, trained, and experienced) on a seven-point scale (e.g., the measure for knowledge ranged from "not at all knowledgeable" to "very knowledgeable"). We adapted these measures of consumer expertise from Netemeyer and Bearden's (1992) work, and the coefficient alpha showed high reliability (.97).

"Design

"A fractional factorial design prescribed the variations over experimental conditions."

Dellaert, Benedict G. C. and Stefan Stremersch (2005). Marketing Mass-Customized Products: Striking a Balance Between Utility and Complexity. Journal of Marketing Research (42:2), 219-227.
---

"Method

"A PC-based homebanking service was simulated in a laboratory, whereby expectations and performance of the service response time were manipulated in a 2 × 2 factorial design. ...

"Manipulations

"The attribute chosen to manipulate both performance and expectations was response time. ...

"Analysis and Results

"Manipulation checks revealed that the subjects perceived the two independent variables as intended."
Wirtz, Jochen and John E. G. Bateson (1999). Consumer Satisfaction with Services Integrating the Environment Perspective in Services Marketing into the Traditional Disconfirmation Paradigm. Journal of Business Research (44:1), 55-66.
---

"Design and Subjects

"A 2 (high/low endorser credibility) × 2 (high/low corporate credibility) between-subjects factorial design was used to test the hypotheses. ...

"Independent Variables

The second page of the booklet contained the description of the corporation, either negative or positive (Goldberg and Hartwick, 1990). The students were instructed to read the excerpt that was said to be taken from a Wall Street Journal article. After the description, they were instructed to proceed without turning back.

"The next page showed the ad for Pride athletic shoes that contained a picture of either Florence Griffith-Joyner or Roseann Barr. The layout for each ad as well as the headline and copy were identical in each of the four conditions to eliminate the possibility of any potential confounds. The copy consisted of a short quote that was attributed to the endorser stressing the natural fit of the shoe. No other product attributes or benefits were provided."

Lafferty, Barbara A. and Ronald E. Goldsmith (1998). Corporate Credibility’s Role in Consumers’ Attitudes and Purchase Intentions When a High versus a Low Credibility Endorser Is Used in the Ad. Journal of Business Research (44:2), 109-116.