Where do you get respondents?
This depends on the chosen method.
For home surveys, the interviewer visits the apartments of potential respondents and, should the person fit the selection criteria, invites him to participate in the study.
For telephone interviews, interviewers call random phone numbers selected by computer and ask people to answer the questionnaire.
For hall tests, respondent recruitment and preliminary selection criteria checks generally occur right on the street, by the study location.
For qualitative research (focus groups, in-depth interviews), respondents are recruited by recruiters – field department employees that use their social networks to search for people that meet selection criteria.
How do you get people to answer questions?
For short surveys (low amount of time lost by respondent), additional respondent motivation in the form of a gift is generally unnecessary. Respondents participate in the research out of interest for the survey and due to interviewers’ persuasiveness.
For long home surveys or hall tests, respondents are generally compensated for time spent with a small gift, whose value depends on the material status of the study’s target group.
For qualitative research, where respondents are invited to the Researcher’s office (focus groups and sometimes in-depth interviews), the incentive is never under 500 RUB. When surveying high-income groups, financial motivation (generally in the form of gifts) can reach several hundred dollars.
Despite interviewer persuasiveness and compensated participation, the share of those refusing to participate increases yearly, and this is a significant issue for the marketing research industry. A high share of refusals lowers study representativeness, because the opinions of those agreeing to be interviewed may differ from those of the people that refused.
The primary causes of increasing refusals are the growing deficit of free time and loss of interest for market research (studies are conducted very widely in big cities, so participation in them, to the majority, is no longer an interesting experience or something out of the ordinary).
One of the means to combat refusals is to decrease interview length: the chances that a person will agree to answer a 5-minute questionnaire are far greater than those for a one-hour questionnaire.
How do interviewers get inside residential buildings? After all, today, most apartment buildings have intercoms and security guards.
For the experienced interviewer, code locks and intercoms are not an issue. It is always possible to wait until one of the residents enters or exits the building to gain entry. Some interviewers even manage to guess lock codes, based on wear and tear on buttons.
As to security guards, these are a true ‘threat’ to the interviewer, since it is their job to keep strangers out of the building. Here, the interviewer has only one option – to use personal charm and demonstrate feats of charisma.
They are all lying!
In fact, when market research is conducted properly, the issue of respondent dishonesty is not at all significant. The percentage of compulsive liars in the population is so low that it cannot affect study results.
Data distortion may occur if so-called ‘sensitive’ questions (for example, about: incomes, sexual or deviant behaviour, etc.) or questions that the respondent cannot answer (for example, asking to evaluate the traits of a brand that the respondent has never consumed) are asked during the survey.
- For ‘sensitive’ topics, if possible, it is best to decide against collecting such information, or to use indirect questions
- Questions that the respondents cannot answer signify an error in study planning: the information gathered does not match the survey’s respondent selection
You pay them well for 2 hours in a focus group. Probably, some students participate daily?
Unfortunately, in recent years, so-called ‘professional respondents’ (people that participate in focus groups and in-depth interviews for profit) have appeared in big cities. At the same time, no matter how skillfully they imitate real consumers, their opinions can in no case be used to make decisions on a brand’s further actions on the market. Therefore, filtering out ‘professional respondents’ is a highly critical factor in conducting qualitative research.
To combat the ‘professionals’, O+K keeps its own database of qualitative survey participants and also uses OIROM’s database. Information about all participants of O+K focus groups and in-depth interviews is entered into the database.
Will your interviewers fly out to Khabarovsk if I need to conduct research in that city?
No, they will not.
For quantitative research, O+K has staff interviewers in Moscow and St. Petersburg. In other Russian and CIS cities, our studies are conducted by our long-term regional partners, professional organizations that specialize in field work for national research companies.
The quality of field work beyond Moscow and St. Petersburg is guaranteed by a long history of partnership and our constant supervision of their work. The majority of Russian marketing firms prefer to use partners for field work in other regions.
O+K’s membership in IRIS, an international association of research companies, lets us conduct research beyond the CIS.
At the same time, field work for qualitative research (focus groups, in-depth interviews) is done by O+K’s staff moderators.
How could 200, 500, or even 1,000 people speak for all of my product’s consumers?
You must admit that surveying ‘all the consumers’ is, in the majority of cases, an expensive and long task. Fortunately, mathematicians have cooked up sampling methods, which allow us to define deviations in study data to a certain degree of confidence (for example, stating with 95% confidence that the share of this brand’s consumers is 26% ± 5%). A bigger sample size makes for a smaller confidence interval (a smaller plus-minus figure).
A properly formed sample (objectively representing the target audience) of several hundred people can provide sufficiently precise answers about the preferences and opinions of a target audience of any size whatsoever.
‘Proper’ sampling (or rather, its representativeness) is provided by:
- Sufficient sample size
- Following random selection procedures when forming the sample
In any case, figures from a survey of 10 people cannot be used to evaluate thousands or millions of other consumers – too great is the chance that at least one atypical respondent will significantly distort the data.
A random selection procedure, providing all representatives of the target audience with equal chances of being in the sample, allows us to achieve representativeness. Equal chances of participating in the study guarantee a proper cross-section of the target audience and the ability to use study results to describe the entire target audience.
Where do you fabricate your data?
No respectable research firm would fabricate surveys at firm level. Those who had been guilty of such things have long since been exposed by Clients and colleagues, back in the dawn of Russian market research.
At individual interviewer level, however, drylabbing questionnaires unfortunately still exists, albeit rarely. O+K’s field department (like the field departments of other companies) is ever-diligent in its fight for information gathering quality.
To be sure of data quality, O+K uses the following interview testing standards:
- 100% visual supervision of field work when conducting personal interviews
- At least 25% control calls when conducting personal interviews
- Telephone interviewing listened in on by at least two CATI department supervisors at any given time
- Two levels of field work quality control in regional projects (on location and from the head office)
- Using a qualitative research ‘professional respondent’ database (focus groups and in-depth interviews)
I would like to purchase ready-made research. Have you done something in my market?
Unfortunately, we do not offer the results of completed research. O+K only does market research to order.
All information collected and analysis belongs exclusively to the Client that ordered this specific study and cannot be provided to third parties.