Editor's note: Doug Spaulding is director of quantitative research, Market Development, Inc., San Diego. David Taber is director of business development, MDI Interviewing Services, Chula Vista, Calif.

"I need a bid for some telephone fieldwork for a real straightforward study. The study would consist of 200 15-minute Spanish-language interviews among Hispanics from the Los Angeles ADI who are the primary grocery shopper for their household and who purchase orange juice. The incidence for the study is 60 percent."

If your field supplier gives you a price from the above request, then it may be time to get a new supplier!

Assuming both parties agree that incidence means effective study incidence (ESI) as defined by the Marketing Research Association (MRA) and the Council of American Survey Research Organization (CASRO), there are still a number of additional questions a good field supplier should confirm with you before quoting a price for any study.

For Hispanic studies, however, the following question must be asked before an estimate can be provided: "What type of sample source are you assuming with the quoted incidence?"

Furthermore, for Spanish-speaking Hispanic studies another question is just as critical: "Are you treating English-only speaking potential respondents as non-screened or as screened respondents who do not qualify for study participation?"

As a brief background, the ESI is calculated by dividing the total number of individuals screened and eligible to participate in the study into the total number of individuals screened, that is

ESI = Total number screened and eligible divided by Total number screened

A respondent would be considered to have been screened if she or he was asked all of the screening questions or terminated during the screening process because she or he did not qualify for the study.

A respondent would be considered to have been eligible if she or he "passed" all the screening questions and qualified for study participation. It does not matter whether the respondent actually completed the interview or not.

Sample source

Unlike the general market, there are a variety of accepted sample sources for conducting "random" custom research among Hispanics. The three most distinct sources are: a sample that covers all telephone households within a given market (full coverage), a sample that is limited to telephone households which are located in areas which are heavily populated by Hispanics (high-density), and a sample that is limited to listed telephone households who are headed by individuals with Hispanic surnames (listed surname).

For all Hispanic studies, the sample source has a substantial impact on the ESI. In fact, its impact is often greater than all of the other screening requirements combined. For example, results from a recent study in Los Angeles showed that the Hispanic incidence ranged from 24 percent using a full coverage sample, to 45 percent using a high-density sample, and up to 74 percent for a listed surname sample. This means that if a client claims the ESI is 60 percent and has not considered the sample source that the actual ESI will be much lower. It could range from 44 percent (60 percent x 74 percent) down to 14 percent (60 percent x 24 percent) for the most representative sample.

Treatment of English-only speaking individuals

No matter what sample source is used, the specific area that causes the most incidence discrepancies in Spanish-language studies is how potential respondents who speak only English are classified. The two options are:

1) Treat the respondent as a "language barrier" and consequently an individual who was not screened; or

2) Treat the respondent as a person who has been screened, and terminated, because she or he did not meet the study's requirement of being Spanish speaking.

Since the definition of ESI depends on whether a respondent has been screened or not, the final ESI will be different depending on which of the above options are used. For example, assume dialings from a Spanish-language research study revealed the following:

Not eligible

Speak English only:                50
Not Hispanic:                         5
Don't purchase orange juice:  15

Eligible
Qualified refusals:                  5
Completes:                           40

Using option 1 identified above, the ESI would be calculated as

ESI = 5+40/5+15+5+40 = 69%

Using option 2, however, the ESI would be calculated as

ESI 5+40/50+5+15+5+40 = 39%

Both of the above options can be, and are, used when determining costs for field studies. In the former case, the field supplier simply estimates a lower contact rate since the English-only speaking individuals are treated as non-contacts.

Additionally, neither of the options should be seen as incorrect. Option 1 does a better job of literally following the incidence definitions outlined by CASRO and MRA, while option 2 provides a more accurate estimate of the intended purpose of the ESI, that is, it better illustrates to clients how easy or difficult it is to reach a qualified respondent.

The distinction between most general market studies and Spanish-only studies is that with the former, a potential respondent who does not speak English is usually not included in the study because the interviewer can't determine whether he or she would qualify for study participation, whereas in the Spanish-only studies, a potential respondent who does not speak Spanish is not included in the study because he or she does not qualify for study participation.

When conducting Hispanic studies, particularly those that are conducted only in Spanish, the client and supplier would be well-advised to be fully aware of both the sample source being used and how non- Spanish-speaking individuals will be classified before the price for the fieldwork is finalized. This will reduce any possible disagreements on whether the price should be modified should the final ESI be higher or lower than anticipated.