Research Design

This chapter looks at the types of research designs that are utilized by marketers. A research design is an overall plan or structure for a research project. A research design will use different combinations of primary, secondary, qualitative, and quantitative data. Depending on the overall research questions, research designs in marketing may fall into one of the following three categories:

An exploratory research design is more informal and unstructured than the other two types of designs. Exploratory design is used to explore a situation, especially when the researcher is in unfamiliar territory. Most academic projects begin with exploratory research when researchers undertake desk research. Similarly, when marketers plan to venture into new markets, such as India, it is advisable to employ an exploratory research design. This could mean going through case studies, accessing published reports (i.e., secondary data) on the market, undertaking experience interviews with experts, and conducting focus groups – if needed. Exploratory design is useful in gaining background information and deciding about future research approaches. It may generate more questions, which need to be tackled with other research designs.

As the name suggests, descriptive research design is employed to describe the market or respondents’ characteristics. This type of research design generates quantitative information. Therefore, such a design would often involve surveys. Surveys are useful to measure the descriptive numbers relating to respondents’ age groups, income levels, expenditure patterns, and even attitudes. This can be done by employing relevant scales, such as, “On a scale of 1–5, how satisfied were you with this organisation’s service?”.

Physiological measurements, such as people’s involuntary responses (such as heart rate, skin changes, and eye movement) to marketing stimuli, such as an advertisement may also be categorised in descriptive research. While recording such information requires special instruments, it is nevertheless a type of descriptive information that can be useful for marketers.

Source: Meanthat and authentic data science^[1]

Causal research design (also known as experimental research design) examines cause-and-effect relationships. A well-designed experiment is the best way of understanding how one variable (e.g., advertisement) may influence another variable (e.g., sale of a product). An experiment involves one or more independent variables (for example, price level, and product features) which are manipulated to determine how they may impact one or more dependent variables (such as customer preference or customer satisfaction). Independent variables can be ‘manipulated’ by the researcher while the dependent variables are variables that get influenced due to changes in the independent variables.

Experiments can be categorised as field experiments or lab experiments.  When an experiment is conducted in a natural setting – such as in a retail store – it is referred to as a field experiment. On the other hand, experiments conducted in a researcher’s office or a university classroom are lab experiments. A lab experiment may be undertaken to measure the impact of an ad on the subject’s attitude.

A field experiment is often seen as providing researchers with reliable results as it is undertaken in the actual environment. However, it is not easy to implement a field experiment. It is difficult to control other factors – in the real environment – which may also impact the final results. Moreover, running an experiment requires expertise in the design of the experiment. It may also require financial and time resources. As an example, Mcdonald’s may wish to see if a drop in the price of its cheeseburger results in greater sales. The fast-food restaurant may drop the price in one locality, such as Fairfield. It may also tediously measure the sales around the time of a price reduction. While interpreting the results, it would need to be assured that any changes in sales figures can be attributed to the price change – and not to other environmental factors such as an increase in prices at Hungry Jacks or an overall increase in customers due to a local football match. Thus, controlling such extraneous variables – that is, all those variables besides the identified independent variables which may also affect the dependent variable.

One of the popular experimental research designs is the ‘Before-After’ Testing design. As the name suggests, it measures the dependent variable, before and after a change in the independent variable. An example will help clarify the concept:

Experimental Group       (R)                 O₁            X           O₂

Control Group                                          O₃                          O₄

Experimental Effect (E) =                     (O₂  – O₁ ) – (O₄  – O₃)

R = random allocation of subjects to experimental or control group

O = observation

X = treatment or manipulation of the independent variable, such as a change in price

Explanation of the above example:

1. Subjects or research participants are randomly allocated to one of the groups
2. The experimental group is the one in which the independent variable (e.g., price) is manipulated (e.g., reduced)
3. The control group is not exposed to any changes in the independent variable
4. Measurements (i.e., O) for the dependent variable (e.g., sales) – for both groups – are taken before the change in price (X) – PRE-TEST
5. Measurements (i..e., O)  for the dependent variable are taken after the change in price – POST-TEST
6. The difference between O₂ and O₁ demonstrates the change in the dependent variable due to the change in X (independent variable)
7. The difference between O₄ and O₃ demonstrate any change in the dependent variable due to factors other than X
8. Therefore, the difference between the two groups  (O₂  – O₁ ) – (O₄  – O₃) demonstrates the true effect of the independent variable as it removes any influence of extraneous variables.