All other research that is not exploratory in nature and that is aimed at evaluating alternative courses of action or measuring and monitoring the organization’s performance is described as conclusive research. Conclusive research can pro- vide information that is inherently descriptive in nature or causal in nature.
Looking first at descriptive information about a market, examples are:
the proportion of the population buying a particular brand of toothpaste;
146 The Marketing Book
Research design Conclusive research design Exploratory research design Causal research Descriptive research Cross-sectional research Longitudinal research
Figure 8.2 Types of research design
the customers’ attitudes towards an organization’s new products;
the level of awareness of a particular sponsorship activity;
the extent to which customers are satisfied with the service they receive.
Descriptive information of this type is gathered using descriptive research. It provides the answers to the who, what, where, how and when of mar- keting research. The research outcomes describe what is happening; they generally do not explain why it is happening. Descriptive research is appro- priate when the research objectives include the description of the characteristics, frequency or prediction of marketing phenomena.
If descriptive research of this type is under- taken in relation to one point in time, it is called a cross-sectional study and involves the research being undertaken once to explore what is happening at that single point in time. Alterna- tively, if research intends to measure trends in awareness, attitudes or behaviour over time, it is called a longitudinal study. This may involve ask- ing the same questions on a number of occasions of either the same respondents or of respondents with similar characteristics. The cross-sectional survey is aimed at obtaining a picture of a mar- ketplace at one point in time. However, longitu- dinal studies, such as panel surveys, do enable researchers to go back to the same respondents again and again to detect ongoing changes in behaviour, awareness and attitudes in response to advertising campaigns, pricing changes or even through changes in the economy. The cost and complexity of maintaining a group of respond- ents for a longitudinal study tends to be greater than for one-off cross-sectional projects.
Descriptive research can tell us that two vari- ables seem to be somehow associated, such as dis- counting and sales, but cannot provide reasonable proof that high levels of discounting result in high sales. Research that does examine information on relationships and the impact of one variable on another is called causal research. This addresses research questions such as:
the relationship between disposable income and expenditure on luxuries;
the relationship between time listening to radio and expenditure on music downloads;
the relationship between awareness of sponsorship activity and purchasing behaviour.
Causal research provides the type of evidence necessary for making inferences about the rela- tionships between variables; for example, whether one variable causes or determines the value of another variable.
Causal research and descriptive research should not be seen as mutually exclusive, as some studies will incorporate elements of both causal and descriptive research. Conclusive projects should be seen as falling along a research con- tinuum with ‘purely descriptive studies’ at one extreme and ‘purely causal with strict manipula- tion and testing of relationships’ at the other. Virtually all marketing research projects fall some- where along this continuum, although the point where descriptive ends and causal begins is subjective and somewhat arbitrary.
Studies that are nearer to the causal end of the continuum will require an experimental research design in comparison to more standard data gathering for a descriptive study. Experimental research allows changes in behaviour or attitudes to be measured while systematically manipulat- ing one marketing variable (e.g. price) and holding all other variables constant.
Although in pure science a completely con- trolled experiment can indicate for sure whether something is caused by something else, in market- ing practice complete control is rarely possible. Therefore seldom can causation be conclusively established in practical settings, although most experimental research designs will provide an indicative set of findings.
In addition to the determination of the data collection method, the research design phase will also consider the nature and number of respond- ents (sampling) to be included in the research as well as aspects such as the timing and scheduling of the research.
Stage 4: collection of secondary data
Secondary data is information that has been previ-ously gathered for some purpose other than the current research project. There are two main sources of secondary data: data available within the organ- ization (internal data) and information available from published and electronic sources originating outside the organization (external data). Internal data may include information from customer loy- alty cards, sales reports and information in the internal marketing information system. External data may include government reports, news- papers, published research reports and the Internet.
Secondary data is used in many studies because it can be obtained at a fraction of the cost and time involved in primary data collection. It is most com- monly obtained prior to the primary research as it:
can help to clarify or redefine the research requirements. For example, internal data held on customers may provide more information on the detail of customer behaviour and may therefore clarify which customers or which product offerings should be researched further;
may actually satisfy the research needs without the requirement for further primary research;
may alert the marketing researcher to potential problems or difficulties. Secondary information about previous research studies may identify difficulties in accessing respondents through, for example, the Internet and therefore may persuade a researcher to switch to some other data collection method.
Stage 5: collection of primary data
Primary data will be collected by a programme of observation, qualitative or quantitative research, either separately or in combination. Each of these types of research and their application will be dis- cussed in detail in later chapters. In this chapter, it is sufficient to signpost their existence and define them (Wilson, 2006):
Observation research is a data gathering
approach where information is collected on the behaviour of people, objects and organizations without any questions being asked of the partici- pants. The researcher becomes the witness of behaviour and events rather than the collector of information second-hand from others about their perceptions and recollections of behaviour and events. Events may be witnessed by human observers or using equipment such as security cameras, cookies on websites, electronic scanners, automatic counting devices at the doors of shops or buildings or other forms of recording device.
Observation only measures behaviour – it can- not investigate reasons behind behaviour, it cannot assess the participant’s attitudes towards the behaviour and it cannot measure the likelihood of the participant repeating the behaviour. Also, only public behaviour is observed; private behaviour in people’s homes or offices is generally beyond the scope of observation studies unless permission has been obtained to install recording devices in an individual’s home. The types of behaviour that can be measured are physical actions such as shopping
patterns or television viewing, spatial patterns such as traffic flows on roads and in-store, temporal pat- terns such as the amount of time spent queuing, verbal behaviour such as conversations with cus- tomer service personnel and physical objects such as the brands of products in a consumer’s kitchen cupboard.
The major advantage of observation research over surveys of respondents is that the data that is collected does not have distortions or inaccura- cies as a result of memory error or social desir- ability bias. The data that is recorded is the actual behaviour that took place. Often facts are only brought to light by means of natural settings only. The interviewee is not conscious of them and they are therefore not easy to get at by question- ing. People may not remember which shelves they looked at when going around a supermar- ket, however, a video camera can capture every movement.
Observation overcomes the high refusal rates that may exist for some survey research. Respondents may be willing to provide their atti- tudes in a survey but may be less willing to spend a large amount of time listing their behaviours and purchases. Observation can gather this behav- ioural information without inconveniencing the participant and in certain cases without the partici- pant even being aware that it is happening.
Qualitative research uses an unstructured
research approach with a small number of care- fully selected individuals to produce non- quantifiable insights into behaviour, motivations and attitudes. It is a mistake to consider qualita- tive and quantitative research as two distinctly separate bodies of research – many studies encom- pass both approaches, with qualitative research being used to explore and understand attitudes and behaviour, and quantitative research being used to measure how widespread these attitudes and behaviours are. However, in comparison to quantitative research:
the data gathering process in qualitative research is less structured and more flexible than
quantitative research and does not rely on the predefined question and answer format associated with questionnaires;
qualitative research involves small samples of individuals who are not necessarily representative of larger populations, however, great care is taken
in the selection of respondents owing to the time
and effort that will be spent on researching the views of each of them;
the qualitative researcher obtains deeper and more
penetrating insights into topics than would be the case
with a questionnaire or a more structured interview;
the qualitative data produced is not quantifiable and is not statistically valid – qualitative research is concerned with understanding things rather than with measuring them.
The most commonly used forms of qualitative research are the individual depth interview and the group discussion.