Grupo Radar

Quantitative research is usually done via surveys which are systematic information collections from a sample and through a pre-elaborated questionnaire.   This is used in order to achieve results that can be projected to a certain target, for example individuals who live in Montevideo and metropolitan area and have at least one credit card.

With these techniques, questions such as “how many”, “who”, “with what frequency”, “where” or “when” can be answered with precision and intend to obtain numeric and objective measures of facts, habits, behaviors or opinions.  On the other hand, they do not allow us to go deeper with the question “why” in which case a Qualitative method would be more effective.

The questionnaires should be carefully worded by very trained professionals so that they will not contain ambiguous or skewed questions and that the same are easily understood by any member of the sample and they should cover all the research objectives.

Sample design and survey technique are the two key issues in these researches  because therein lies the possibility of projecting the results to the entire studied universe.  We try to achieve the best compromise between the ideal method, what is actually possible and the available budget.  Read more on different kinds of samples

 

Personal surveys (face to face)

Up

These are applicable when the sample units are households, individuals or even companies and other organizations.

Advantages: samples can be designed so that any individual or household of the objective public will have the same chance to participate; visual stimuli can be used (pictures, printed material, etc.) and the survey can be longer.

Disadvantages: higher costs, the surveys are longer than those by phone and it is difficult to reach the higher socioeconomic tiers.

When the target consists of people, the personal interviews can be made according to the following formats:

  • At homes: the interviewer conducts the survey at the respondent’s home.
  • In a central location: respondents are summoned at one or several locations where they will be received by a surveyor who will conduct the interview.  This format is normally used for product testing (food or drinks) since a certain infrastructure is required (kitchen, refrigerator, etc.).
  • Street intercept: people are stopped in the street —pedestrians or motorists— at locations which have been previously agreed with the client.

 

Telephone surveys

Up

These are applicable when the sample units are households, individuals or even companies and other organizations.

In any case the informant should be previously defined with precision: householder, housewife, any other person who meets certain characteristics according to age, gender or other variables, purchasing decision maker, main user or consumer of a certain product or service, etc.

Advantages: less time is required than in a face to face survey, the cost is generally lower and it is easier to reach certain population segments (high socioeconomic tiers, company purchasing decision makers, etc.)

Disadvantages: phoneless homes are excluded, no visual stimuli can be used (cards, pictures, etc.) and maximum duration cannot exceed that of face to face interviews.

 

Self adiministered mail surveys

Up

These are generally applicable to very specific publics and when long and complicated forms are required for which the respondent has to previously collect information.  For example: physicians who have to gather information on their patients, businessmen who need to have at hand company information.

The survey is only delivered to the respondent once he has agreed to answer it and will then return it by mail or courier.

 

Self administered on-line surveys

Up

This format implies that an invitation is sent to an e-mail data base.  This invitation has a link to an on-line survey. The person fills in the survey which is stored in a data base which will be downloaded and processed.

The main advantage of this kind of survey is the speed with which one can survey a great amount of people at very low cost.  The main disadvantage is that the statistical representation of the sample cannot be guaranteed since the response rate can be very variable according to the target tiers, which is usually below 5%.

 

Mystery Shopping

Up

The “mystery shopping” or “mystery consumer” technique is to simulate to be a client and through direct observation to register a series of quality indicators.  This technique is particularly useful to detect whether the customer services rules defined by the company management are being followed or not.

For example, the following indicators can be surveyed at a gas station:

  • lapse between your arrival at the gas station and the moment the employee greets you
  • greeting
  • confirmation of type of gas and amount to be charged
  • product display at forecourt
  • gasoline pump cleanliness
  • illumination at forecourt
  • sign visibility

 

Some kinds of sampling

Up

Probability or random sampling: each one of the sample units has one known probability of being included in the sample.

  • Simple random: all the members of the universe to be studied are numerated and a random sample is drawn.  This sampling is the simplest one but it can only be applied when working with a data base provided by the client.  When the research is on a very large universe, this is inapplicable (for example public opinion national surveys).  A variation of a simple random sample is the systematic random: all the members of the universe are numerated but only one is chosen at random and all the rest are defined according to a certain algorithm (for example one every “n” elements).
  •  Stratified: this means you take into account typical categories different from one another (strata) but homogeneous regarding one characteristic (for example according to the size of the town).  This type of sample allows all the strata of interest to be adequately represented in the sample.  Each strata works independently and the simple random or the stratified random can be applied in order to choose the concrete elements which will be part of the sample.
  • In clusters: instead of selecting the elements directly from the universe, the sample is a group of elements of said universe which form a unit we call cluster.  A classic example is polling in a survey “exit polls.”  A number of clusters are randomized and then all the elements of the selected clusters are investigated.

Non probability: these are used when it is not possible to resort to the probability. One disadvantage is that it is not possible to calculate the sample error.

  • Quota sampling is generally based on a good knowledge of the population strata and/or the most “representative” or “adequate” individuals for the research purposes.  There is a certain resemblance to the stratified random sampling but it has not the random condition of the first.  In this type of sampling, quotas are fixed, which means a certain amount of individuals who have certain conditions, for example: twenty, 25-to-40-year-old-female-individuals, who live in Montevideo.  Once the quota is fixed, the first twenty found with these characteristics will be selected.
  • Convenience sampling: The individuals who form this sample are selected directly and intentionally depending on the ease with which they can be reached.

Snowball: some individuals who will lead us to other individuals are identified and these lead us to others and so on until we have a large enough sample.  This type is very frequently used for marginal population studies because they are difficult to reach or are too small to implement other sampling techniques.  For example: people with a particular disease or who practice a certain sport, or customers of very selective products.

©Grupo Radar