Week 5.3 Formal Research Design

Earlier in the previous lesson, we saw that marketing research consists of four basic stages. In this lesson, we are moving into the second stage, formal research design.

four-stages-marketing-research
Four Stages of Marketing Research. 

 

Formal Research Design

In the formal research design stage, the researcher produces a plan that outlines the method and procedures for collecting and analyzing the required information.

Remember that exploratory research is employed in the first stage (problem definition).  In the second stage (formal research design), descriptive and causal research are utilized. 

3 icons: survey, experiment, and observation.

 There are three methods researchers use in conducting descriptive and causal research:

  1. survey, 
  2. experiment, and
  3. observation.

 

survey icon Survey

Survey is known as the most common research method for generating primary data. Survey method generates data by asking people questions and recording their responses on a questionnaire. Surveys can be conducted by personal interview (face-to-face), by mail, by telephone, or online. There are advantages and disadvantages of conducting surveys with different types of mediums. The comparison of mail, telephone, personal, and online surveys are presented in the table below.

Basis of ComparisonMail SurveysTelephone SurveysPersonal InterviewsOnline Surveys
Cost per completed survey  Not very expensive Moderately expensive Most expensive Very expensive
Ability to probe and ask complex questions  Little to none Some, interviewer can probe and elaborate Much, interviewer is face-to-face Depends, can go back and ask respondent to clarify responses
Opportunity for interview to bias results None Some, because of voice and gender Significant, voice, appearance, gender present Little, if done correctly
Anonymity given to respondent Complete, unless coded instruments are used Some, because of telephone contact Little, because of face-to-face contact Some, e-mail/user name may be known
Response rate Poor or fair Fair, refusal rates are increasing Good Very good, if done correctly
Speed of data collection Poor Good Good Very good

As you see in the chart above, the personal interview method is the most expensive type of survey method per contact due to the high cost of reaching respondents individually at mutually agreed locations. In order to reduce the cost a little bit, researchers started to employ mall intercept interviews. These are personal interviews of consumers at shopping centres. While mall intercept interviews help to reduce the cost of personal interviews, they might bias the results. When the interviews are conducted at a mall, the people selected belong to the demographics of the neighbourhood served by that particular mall. They do not necessarily represent the general characteristics of the consumers targeted for the interviews, causing possible bias in results.

Even though cost is a disadvantage for the personal interview method, its biggest advantage is the ability to ask complex questions which allows for extracting additional information from the respondent. The highest speed in data collection is from online surveys. Also, online surveys are less expensive in comparison to the personal interview method.

four different shampoo bottles

In the survey method, sometimes panel data is collected. In the case of panel data, marketers survey people today and then after some time has passed they get back to the same people and survey them again. Panel data can reveal important information. For example, consumers indicate which brand of shampoo they use today. When they are surveyed again in a year, marketers can see if they switched from one brand to another and more importantly why. 

In the survey method, the results marketers obtain depend on the type of questions they include in the survey. The results might be biased if there are wording problems in the questions. The next table shows typical problems in wording questions. Marketers pay attention to such issues in order to obtain unbiased answers from respondents.

 

 

ProblemSample QuestionExplanation
 Leading question Why do you like Wendy's fresh meat hamburgers better than those of competitors?  The consumer is led to make statements favouring Wendy's hamburgers.
Ambiguous question Do you eat at fast-food restaurants regularly?  □Yes  □No What does "regularly" mean (e.g., once a day, once a week, or once a month)?
Unanswerable question What was the occasion when you ate your first hamburger? How can the consumer remember the answer? Does it matter?
Two questions in one  Do you eat Wendy's hamburgers and chili?  □Yes  □No How does the consumer answer if they eat Wendy's hamburgers but not chili? 
Nonexhaustive question Where do you live?  □At home  □In dormitory How does the consumer answer if they live in an apartment?
Non-mutually exclusive answers What is your age?  □Under 20  □20-40  □40 and over How does the consumer answer if they are 40 years old?
There are five different types of questions used in surveys:

 

  1. open-end question,
  2. closed-end or fixed alternative question,
  3. dichotomous question,
  4. semantic differential scale question, and
  5. Likert scale question.
Type of QuestionDescriptionExample
Open-end question  Respondent writes answer in his or her own words. What are your expectations from your real estate agent?          

______________________________________________________

______________________________________________________

Closed-end or fixed alternative question   Respondent checks one answer in a multiple choice question What do you like the most about taking an online course?
  1. gives me flexibility to study at my own pace.
  2. enables me to work while taking a course for credit.
  3. enables me to travel during the term.
  4. leaves free time in my daily schedule for other important things in my life.  
Dichotomous question Respondent answers only "yes" or "no." 

Would you paint your garage door red?

□ Yes

□ No

Semantic differential scale question

Respondent presented with a point scale in which opposite ends have one- or two-word adjectives with opposite meanings. 

Respondent marks an X on only one of the scales. 

Place an X in the box that best describes the quality of your experience. 

Cleanliness of room clean            dirty 
Room service prompt           late
Noise control quiet           loud
Pleasantness of staff polite            rude

 

Likert scale question Respondent indicates the extent to which they agree or disagree with a statement.

Check the box that best describes your reaction to each statement below.

StatementStrongly agree  AgreeDon't knowDisagreeStrongly disagree  
Processed meat should not be included in a child's diet.          
Children need one hour per day to play outdoors.          
You shouldn't discuss politics with colleagues from work.          

 

experiment icon Experiment

The method of experiment is especially important in establishing cause-and-effect relationships (causal research). It involves the manipulation of an independent variable (cause) and the measurement of its effect on the dependent variable (effect) under controlled conditions. Independent variables are the ones that are under the organization’s control.

Stop and Think Question: Can you think of an independent variable? Think about independent variables before clicking on reveal.

shopping cart in a grocery store aisle

Independent variables are under the control of the organization. They could be one or more of the marketing mix elements (4Ps) such as product features, price, place (distribution), or promotion.

 

An example of a dependent variable could be the change in sales. As we change an independent variable such as the product’s packaging, we measure its effect on a dependent variable, which would be the volume of sales. 

Experiments can be conducted in the field or in a laboratory. A field experiment is one in which the researcher goes to the real world, such as to a store or wherever the behaviour being studied occurs naturally.

observation icon Observation

Observation is the last method on the list of methods we use to obtain primary data. It involves watching behaviour either mechanically or in person. 

camera on ceiling of department store

Example: electronic cameras in the entrance of stores that record customers entering or leaving the store. Observation method is employed with the use of these cameras. The number of customers entering and leaving can be calculated which is useful in deciding the optimal opening and closing hours of the store.

many television screens showing variety of programs

Example: the Nielsen Media Research's people meter. It is used to measure the size of audiences watching TV programs delivered by the networks. These ratings determine the success or failure of TV shows and play an important role in setting advertising rates for these shows.

close-up view of computer screen with cursor at search field of browser

Example: electronic recordings to record Internet user behaviour. Companies track the actual mouse clicks made by users as they surf the Internet. These observations form a vital database for marketers in understanding user behavior. 

There are some creative methods for obtaining observational data such as ethnographic research and mystery shoppers.  

Ethnographic research is a specialized observational approach. Trained researchers observe consumers in their natural environments as they consume or use products. The goal is to discover subtle emotional reactions of consumers to the products. 

Deli Creations packages

Example: researchers from Kraft spent several months with consumers in their kitchens watching consumers prepare sandwiches. They discovered that consumers valued complete, ready-to-serve meals that are easy to prepare. Kraft launched Deli Creations as a result of the findings of this ethnographic research. 

A creative approach in obtaining observational data is the hiring of mystery shoppers. Businesses hire these people, who go to their stores pretending to be a customer. The individuals record their experience and prepare a report based on their observations. The key is to keep their identities secret from the staff working in the stores. 

Sun umbrella next to swimming pool with hat and towel on a table.

Example: a vacation resort hires a mystery shopper who is paid to travel to their location for a vacation. The mystery shopper stays at the resort, eats at the restaurants, plays golf, and swims at the pools/ocean. The experience of the mystery shopper reveals valuable information for the resort in evaluating how well they treat customers. 

Data can be directly obtained from brain activity. The concept of neuromarketing refers to collecting data based on the brain activity of consumers as they make purchase decisions. It measures the activity of certain centers of the brain as consumers are exposed to advertisements of products/services. 

In the formal research design stage we have seen survey, experiment, and observation methods to collect primary data. 

Stop and Think Question: Is there an optimal research design that is superior to other methods and should be given priority in research?

thinking icon

There is no optimal research design that would be always superior to other methods or that would fit into all research types. Researchers develop the ability to select the most appropriate research method as they work with data over time. There is usually more than one research method that could be utilized. An experienced researcher makes the decision on the best method depending on the specifics of the study in hand. Inexperienced researchers often employ survey method first. Experiment and observation methods require the eye of an experienced researcher in order to extract relevant and valuable information. 

 

We have defined primary and secondary data earlier. In collecting primary data, sampling plays an important role.  

Sampling is the process of gathering data from a subset of the total population rather than from all members (census) of the population. 

We are simply considering a subset from a larger population when collecting data with sampling. For sampling to be successful, the selected sample should be representative of the population as a whole. Before deciding on the subset of the population, we need to know the target market in mind for the study. Another important factor is the size of the sample. A larger sample size usually provides better representation than a smaller size. However, as long as the sample carries the main characteristics of the target population, small size samples can provide reliable measures for the study as well.

Two sampling techniques are applied: 

  1. probability sampling, and 
  2. non-probability sampling.

Probability Sampling

Each member of the target population has a specific known chance of being selected. 

Example: University of Waterloo wants to conduct a research study to identify the employment opportunities for the last year’s 1,000 graduates. Researchers decide to sample the 1,000 graduates based on the following rule: 100 graduates will be randomly selected. This is an example of probability sampling because each graduate has a known 100/1,000 or 0.10 chance of being contacted.

Non-probability Sampling

In non-probability sampling, researchers apply arbitrary judgment to select the sample. Each member of the target population has either unknown or zero chance of being selected.

Example: University of Waterloo wants to conduct a research study to identify the employment opportunities for the last year’s 1,000 graduates. Researchers decide to sample the 1,000 graduates based on the following rule: 100 graduates who live closest to the university will be contacted. In this case, many graduates will be eliminated due to the distance they live away from the university. This could introduce a bias, or possible lack of representativeness among the 1,000 graduates.  

Non-probability sampling provides advantages in terms of time and budget. It lowers the cost and speeds up the process of data collection. In many cases, non-probability sampling is applied for exploratory research purposes.

Concept Check Questions:

1. Which of the following is an example of a Likert scale?

 

2. _______surveys are somewhat expensive, fair in terms of response rate, and there is good speed with data collection.

 

3. What is the problem with asking the question, "Do you exercise regularly?"