Research Methodology - Qualitative Approaches

RESEARCH METHODOLOGY Welcome to research methodologies. This video will give you an overview of qualitative research. According to Denzin Lincoln and Creswell, qualitative research is a situated activity which locates the observer in the world. It consists of a set of interpretive material practices...

RESEARCH METHODOLOGY Welcome to research methodologies. This video will give you an overview of qualitative research. According to Denzin Lincoln and Creswell, qualitative research is a situated activity which locates the observer in the world. It consists of a set of interpretive material practices that make the world visible. Such practices transformed the world. They turn the world into a series of representations including field notes, interviews, conversations, photographs, recordings and memos to the self. At this level, qualitative research involves an interpretive, naturalistic approach to the world. This means that qualitative researchers study things in their natural settings, attempting to make sense of, or interpret phenomena in terms of the meanings people bring to them. The definition of qualitative research highlights its objective of transforming the world. Qualitative research begins with assumptions and proceeds with the examination of research problems and may refer to groups or individuals or organizations. To study the research problem qualitative researchers adopt an emerging qualitative approach to inquiry and data is collected in a natural setting. The collected qualitative data is analysed to identify patterns or themes. The key characteristics of qualitative research are the following. The natural setting is where qualitative researchers collect their data, for instance, observing the purchasing behaviour of customers in the supermarket. And it's where key informants experienced the issue or problem under study. Within the natural setting of the key informants, the researchers have direct contact with the participants of the research. The second characteristic is multiple sources of data such as interviews, observations and documents. The qualitative researchers review all the collected data and group them into themes or categories that emerge from all the different data sources. The third characteristic is key informant meanings. This refers to ensuring the researcher remains true to what the participants’ means when discussing a problem or issue and not interpreting their own meaning of the data collected. The next characteristic is emergent design. The initial plan of research cannot be tightly outlined and must be able to adapt. Throughout the different stages of the process the research plan changes and can shift when researchers enter the data collection state. For instance, the questions may be adjusted, the forms of data collection may change and some of the individuals to participate in the research project may be replaced. Interpretive inquiry is also a characteristic of qualitative research, in the sense that researchers interpret what they can see, hear and understand. And qualitative research all interested parties, including the readers, key informants and researchers form their own interpretation. Therefore, multiple viewpoints about the same problem may emerge. Final characteristic is a holistic view. Qualitative researchers strive to draft a complex picture of the issue under study. This entails reporting of multiple perspectives, identifying the various factors involved in a situation and identifying the broader picture of the research problem. IMPORTANCE OF QUALITATIVE RESEARCH Qualitative research can be used in reverse order, as projects can start with qualitative or quantitative research approaches, depending on the nature of the research project. The findings, which can emerge from quantitative research, can be enhanced and refined by adding qualitative information, which pertains to customer motivations and perceptions. Market research is used to offer insights about the behaviour of customers and a thorough understanding of customer demand. Qualitative research is used when a research team wishes to explore an issue within a certain group of the population, identifying variables that should be measured (e.g., purchasing behaviour, consumer loyalty) or observe behaviours. Qualitative research can offer valuable insights into a complex and detailed understanding of the issue under investigation. Such detail in qualitative research can be attained by interacting directly with people, visiting their homes or working environments and giving them the space and flexibility to tell their point of view or stories about the issue under investigation. Qualitative research allows individuals to share stories and to express their opinion in a comfortable environment, using open and straightforward streams of communication. Qualitative research is useful when the research team wishes to approach the research problem through narratives or stories, which do not necessitate a structured and formal way of tackling the research problem. In qualitative research, the research team cannot separate what individuals say from the context, this may be their home or work or family, amongst others. As we have already mentioned, qualitative research can be used to complement quantitative research, as adopting mixed methodologies in a set context (e.g., qualitative and quantitative) allows researchers to identify the thoughts and behaviours behind the responses of participants. Qualitative research requires researchers to commit to focusing on the problem and investing time and resources. The qualitative inquiry is for researchers who have the time to commit to the field, to collect detailed data, and who strive to build trust and commitment with the research participants to gain an insider perspective. Through a qualitative study, the researchers are involved in complex and time-consuming procedures of data analysis with the objective of reducing them and organising them into themes. Focus groups consist of 8 to 12 participants who are involved in an in-depth discussion about a topic and are coordinated by a moderator. The interactive nature of the focus group and the group dynamics are essential for the success of the focus group discussion. The dynamic and interactive nature of the focus group differentiates it from the interviews as a data collection of broads. In a focus group, the response of a person may provoke response and reflection from other individuals in the focus group. As a result, an interplay of responses will produce more information than participants may have contributed independently. The research idea of group dynamics comes from the field of social psychology where research findings reveal the individuals who come from all walks of life and from different occupations. I would like to talk about a topic in more depth instead of responding to questions. In a group dynamic direct questions are avoided and are replaced by free and flexible spontaneous discussions and interactions. Due to their nature focus groups are very rich resources of information. Focus groups are an approach to collecting data in qualitative research. They are used extensively in consumer goods companies compared to industrial organizations. Focus groups usually take place in a focus group facility. The location maybe a conference room or a living room setting. The setting is important and is usually chosen to make the participants feel as comfortable as possible. The participants of focus groups are recruited by a number of sources and criteria are established for the group participants. For instance, focus groups recruiters go to the market to find focus group participants. Recruiters strive to avoid repeat or too formal or professional respondents for focus groups. Professional respondents perform like actors and are not considered candid participants. Companies often find it easier to recruit repeat participants than to source a new group of individuals. Focus groups usually lasts around one hour and a half. The initial ten minutes are reserved for introductions and to introduce processes. The moderator also takes around 20 minutes of the session to explain the process and manage the process with the rest of the time left to the focus group members. But why would someone participate in a focus group? This can be for several reasons, such as genuine interest to the topic, knowledge of the topic of discussion, curiosity, or to express viewpoint. Focus group participants can participate for monetary compensation, but these people are generally considered less committed to the objective and the topic of the focus group. Personal in-depth interviews are unstructured interviews that take place through one-to- one interactions. The interviewer should be trained to elicit in-depth and comprehensive answers to each of the questions. In-depth interviews can reveal hidden motivations and are considered one of the most popular data collection approaches in qualitative research. The in-depth interview is guided by the responses of the interviewee. As the interview progresses, the interviewer probes each answer and uses the answers for further questioning. For instance, an in-depth interview may commence with a discussion on purchasing intention for a certain product. The interviewer may follow each answer by asking the participant to elaborate on their previous answer. ETHNOGRAPHY Ethnographic research helps brands to learn about their consumers, it offers valuable insights about the topic under discussion. Ethnography collects qualitative data, as customers offer insights regarding a digital advertising campaign, or a certain product or service. An ethnographic study focuses on collecting information and data with respect to, how the facts are presented. The main attribute of ethnography, is the richness of description of the topic it focuses on. Researchers using an ethnography take into account, and observe the daily life of participants for extended periods of time. This will reveal written behaviours, habits, and attitudes. As a consequence, participants, observation and qualitative interviews are considered the main approaches to collecting data in ethnographic research. Ethnographic studies are broad and become more and more narrow, as the research progresses. The researcher needs to identify associated concept to focus on, as part of the ethnographic research. This requires a significant amount of time, and substantial experience in the field. Ethnographic studies should be undertaken by experienced qualitative researchers. In ethnographic market research, researchers observed the participants using a product or service. Ethnographic research, can reveal gaps in the market, and can help with product or service improvements and refinements. An example of ethnographic market research includes, research undertaken by Panasonic. The company adopted ethnographic research, to observe the behaviour of women in the USA, while grooming. With ethnography, Panasonic gained insights with respect to the colours and designs, which benefit their customers according to the observations. As an outcome of ethnography, Panasonic designed the Lady Shaver, a product that achieved a differentiation advantage compared to other competitive products. In ethnographic research, researchers often visit people in their homes, or in their working environments to observe how a new product could be integrated into their lives in a non- direct informal way. Ethnographic research, allows researchers and companies to collect information about, how a product could impact the daily lives of people without disrupting their daily routines. Ethnography allows companies to offer products or services that best fit the needs, and requirements of their customers. Case studies are a popular qualitative tool in market research because the approach looks in depth at one or a limited number of firms, events, and individuals generally over time. Case studies are often criticised as they do not have the rigour and reliability of scientific research, and they also limit the generalisation of their findings to the whole population. To address the criticisms, it is essential for case studies to have a clear design before any data is collected. These designs should cover the key questions or propositions, as well as processes of interpreting the collected data. The unit of analysis should be identified, as well as the links between data and propositions. Case studies can offer a rich picture regarding life, attitudes and behaviours in groups and organisations. Case studies can assist in developing general principles or identifying attitudes and behaviours over several years. Investigating attitudes and behaviours through different points of time refers to a longitudinal case study methodology. Case studies represent a detailed investigation of a phenomenon that is investigated in- depth. The research question(s) should be clearly designed from the beginning of the research project. The researcher should identify the case(s) that will be examined and the data analysis techniques that will be adopted. The way in which data will be collected should be specified from the beginning of the research project. Clear questions should be designed and used to collect the data. Data should be collected in the field or lab and should be well organised through a clearly structured database or records. Clear questions and clear data collection methods contribute to the development of a robust research report. Granting consumers access to the results of the case studies of a service, product or product line motivates them to purchase the product or service again and again. Case studies are therefore considered valuable go-to-sales tools. The question in market research is: There are a number of ways for market researchers to present case studies that have an effect on consumers, highlight the firms’ solutions to the customer-related problems and help the company or brand to differentiate against the competition. The case study must be blended and presented with other content on the company’s website. Video can be used to communicate how the company addressed the problems which arose from the study. The company can create a section for case studies on the website by showing in detail the challenges which customers encountered regarding the use of products or services and how they coped with them. Each different story must offer a unique solution to the customer and relate to the target audience. ACTION RESEARCH Action research assumes that social phenomena are constantly changing rather than being static. By adopting action research, researchers are part of the change process. Action research designs follow the philosophy that the best way of learning about a company or a social system is to try to change it. This should be facilitated through the participation of an action learning researcher. The individuals who will be most likely affected or participate in implementing the changes should be involved in the research process. Some approaches to action research attempt to implement the changes from outside the organisation and then they attempt to measure the results. The origins of action research come from Kurt Lewin (1948) who tried to change the habits of housewives regarding nutrition, through the use of experimental designs to investigate efficacy. Kurt Lewin’s research was different from the traditional experimental research because he devoted attention to changes to an individual's behaviour. The housewives were active participants in identifying what should change. The limitation with adopting experimental design was that Lewin could not identify the causes of the housewives’ changing behaviour. This necessitated more experiments and qualitative research to identify the reasons that led to the change of behaviour under various circumstances. As the research process went further, it became a cooperative inquiry which was designed to examine individual or community behaviour. It commences with an idea, which the individuals may have, their capability to be self directed, to choose how they will react and offer meaning to their own experiences. The action research approach is different from examining the experiences and perceptions of the key informants. In the case of action research, the key informants become active partners in the research process instead of subjects. Kurt Lewin introduced the action research cycles in his work, which consist of: Pre-Step The pre-step specifies general objectives and establishes an understanding of the context of the project. At this stage, it is essential to address why the project is desirable, what are the internal and external forces that drive change? The future steps of the process should be defined at this stage. The learning groups are identified and formed at this stage as well. These groups are going to offer valuable insight to the action research project. Planning Planning involves a complete plan and a decision on beginning an action research project. The planning step focusses on exploring the context and the purpose of the project. Constructing the issue should align with the previous step. In the planning action stage, collaboration is important, as the construction of a plan should be a joint venture where the researcher engages others in the process. Act Action involves plans being implemented and interventions made collaboratively between the researcher and the research participants. Observe Once the action step is completed, it is important to stop and observe the outcomes and results. By observing what has been learned, it is possible to set the basis for future adjustments to refine the approach. Reflect The final stage is to reflect on and evaluate the outcomes of the action against the objective to see if the planned method was effective, and whether the actions taken matched the plan. The reflection stage will then feed back into the plan stage as the approach is adjusted and developed based to enhance the methodology going forward. In action research projects, there are multiple action research cycles that operate concurrently. The action research takes place by individuals, by teams, within teams, within departmental or interdepartmental groups and between organisations. The rigour of the action research process is based on how the various activities are exposed to critique and how the conclusions are supported by actionable knowledge. The general empirical approach in action research includes:  Experience that is demonstrated in activities such as attending, sensing, and imaging.  Understanding is shown through inquiring and understanding.  Judgement that is depicted through reflecting, weighing up evidence and judging.  Decision is depicted through deliberating, deciding and acting. Experiments are studies which involved the researcher intervening. The usual intervention is to manipulate a variable in a setting and observe how it influences the subjects being investigated. The researcher manipulates the independent variable, and soon after checks whether the dependent variable is influenced by the intervention. In experiments, there is at least one independent variable and one dependent variable in a causal relationship. The independent or explanatory variable, is the cause of influences the dependent variable that is the effect. In a well-executed experiment, researchers should complete a series of activities to carry out their craft with success. T he experiment is a method that establishes causation. However, the creativeness of the researcher is required for the experiment to reach its potential. In an experimental approach, it is essential to select the number of variables and to design the relevant measures of them. The researchers, should select variables which best measured the concepts considered in the research. Further, it is important that all participants in the experiment of the same written instruction. The arrangement of the room, the time of administration, and the experiment contact with the subject should be all online at the time of the experiment. We will now discuss an example of an experiment in market research. A product manager for a line of Sousse wishes to launch a new product line in a department store, to increase sales and improve the brand self-presence. Unfortunately, the department is not convinced that the new shoe line will increase sales. And the owner believes that it will take sales from the existing shoe line. As a solution to this issue, an experiment can take place by conducting an in store test in some stores that are representative of the market. In order to compare sales and revenue before and after the launch of the new shoe line. The experiment can last for a period of six weeks. And will demonstrate if the new line of shoes will bring additional revenue and profits. Survey approaches are based on the design and development of well-structured questionnaires. Key informants may be asked several questions about their intentions, motivations, attitudes, demographic and lifestyle characteristics among others. The questions may be asked and collected verbally, in writing or through a remote device (e.g., laptop). Structure refers to the level of standardisation used during the data collection procedure. In structured data collection, a formal research instrument is developed and the different questions are offered in a pre-arranged order. The research can be classified as direct or indirect based on whether the objective of the research is known to the respondents. In a survey questionnaire, the vast majority of questions are fixed responses that require the respondent to select the answer from a prearranged set of responses. For instance, a question designed to operationalize consumer perceptions of a product may be read as follows: The nutrition value of product X is high and the answers are anchored in a 5 point Likert Scale ranging from strongly disagree (1) to strongly agree (5). Advantage of Surveys Survey methodology has a number of advantages:  It is easy to administer a questionnaire.  The data collected is consistent as the responses are restricted to the alternatives available.  The fixed-response option eliminates the variation of the findings which may be caused by different interviewers.  Coding, analysis and interpretation are technical and straightforward. Disadvantage of Surveys The disadvantages of the survey methodology may include: An unwillingness to offer the necessary information. If the questions refer to the beliefs or motivations of consumers when selecting specific brands or purchasing certain products, the consumers may not provide accurate answers, especially when the information required is sensitive or personal. Questions, which require choosing from a predetermined set of answers, may result in a loss of validity for some types of data including beliefs and motivations. The language and the wording used in the questionnaire needs to be meaningful to the respondents. Despite the limitations, the survey methodology is a widely used approach for collecting primary data in market research. Types of Survey Administrations Survey questionnaires can be administered in different ways, including, telephone interviews, personal interviews, mail interviews and online interviews.  Telephone interviews are classified as traditional or computer assisted (CATI) where the interview takes place through a computerised questionnaire over the telephone.  Personal interviews may take place at home, in the office or in the street. In traditional mail surveys, the questionnaires are posted to the potential respondents.  Finally, electronic surveys take place through email or can be administered on the Internet.  In Internet surveys, respondents are recruited online from different respondent databases. Respondents are required to visit a certain web location in order to complete the survey. KEY OBJECTIVE The key objective of market research is to provide information about the attributes of a certain type of population. A population in this context is defined as the aggregation of all elements that share the same common characteristics. The shared characteristics formed the base to address the market research problem. The population parameters are mainly numbers, including for example, the proportion of consumers who are satisfied with a certain brand of clothes. Thus the population is a group of individuals of interests from whom the researcher needs to collect information. The population may be the target market of a product or a product line. Defining the population of interest is a key step in this sampling process. There are no certain rules for researcher to follow when defining the population of interest. To identify population, the researchers should address the following questions. Whose opinions are needed to meet the aim of the research? The answer is based on the attributes of the target consumers. Information and insights about the population parameters can be taken through either as census or a sample. A census involves a complete enumeration of all the elements of a population. For instance, all consumers purchasing a certain product. The parameters of the population can be calculated straightforwardly soon after the census is enumerated. In a census, data is obtained from or about any member of the population of interest. A sample is defined as a subgroup of the population selected for participation in the study. The characteristics of the sample are known as statistics and are used to make inferences regarding the population parameters. The research results based upon a sample can be generalized to the whole population. The key to making precise predictions about the attributes or behavior of a large population using a small sample has to do with the individuals included in the sample. A sample should be representative of the population. All different types of individuals who make up the population of interests should be represented in the sample in similar proportions into which they exist in the population. Glivenko and Cantelli, two mathematicians who studied probability in the early 1900. Found out that several observations randomly selected from a population will naturally take the shape of the population distribution. This implies that the researchers randomly sample from a population and obtain a sufficient sized sample. Then the sample contains attributes that represent the characteristics of the population. A sample may be more effective than a census due to budget and time constraints. A census is very time consuming and costly. A census is unrealistic If the population is large in business to business markets. For example, a census is visible due to the small population. The Sampling Design Process: Target Population, Sampling Frame, Sampling Techniques The sampling design procedure begins with a number of interrelated steps. These steps include defining the target population, the sampling frame, and selecting sampling approaches. Defining the target population focuses on collecting elements that have the information sought by the researcher. The target population should be defined precisely and accurately. Wrongly specifying the population may result in the failure of the research and produce misleading research results. Accurately defining the target population involves translating the problem into an accurate statement about the individuals and their eligibility to be included in the sample. The target population should be specified in terms of elements, sampling units, extent and time. Element An element is defined as the object from which the information is desired. In survey methodology, the element is the respondent. Sampling Unit A sampling unit is defined as an element that is available to select at a stage during the sampling process. For instance, a company specialising in cosmetics may wish to sample women over 30 years old to gather their opinion of a new line of lipsticks. If the women above 30 are sampled directly, the sampling unit will be the same as an element. The sampling unit could also be households, for example, to identify women who wish to participate. As a result of this, households will be sampled and all females in each household will be considered and interviewed. Sampling Frame A sampling frame refers to representing the various elements of the target population. It consists of a set of guidelines specifying the target population. An example of a sampling frame concerns a customer database or a mailing list. If a population cannot be identified, some guidelines for identifying the target population can be specified, such as random digit dialling in telephone surveys. If the sampling frame is not accurately specified, there is the risk of sampling frame error. To account for sampling frame error, research participants must be screened during the data collection stage. The respondents should be screened in regards to their demographics, familiarity, and usage of the product to ensure they fulfil the criteria to be included in the target population. Screening can limit the unnecessary elements, which are included in the sampling frame, but it cannot account for those elements that have been omitted. Sampling Approach Selecting a sampling approach requires the researcher to decide whether to use a Bayesian sampling approach, to sample with or without replacement and to employ a probability or non-probability sample. Under the spectrum of a Bayesian approach, all elements are selected sequentially. Soon after each element is added to the sample, data is collected and sample statistics are computed. The Bayesian approach considers previous information about population parameters as well as the probabilities of making wrong decisions. In sampling with replacement, an element is selected from the sampling frame and relevant data are produced. Then, the element is put back in the sampling frame. As a consequence, it is possible for an element to be included in the sample for more than once. In sampling without replacement, once the element is selected for inclusion in the sample, it is removed from the sampling frame and cannot be used again. One of the most important decisions when choosing sampling techniques is the adoption of non-probability or probability sampling. Non-probability sampling depends on the discretion of the researcher and probability sampling relies on chance. The sampling Design Process: Sampling Process, Sample Validation, Non Response Issues in Sampling The sampling design procedure begins with a number of interrelated steps. These steps include defining the target population, the sampling frame, and the selection of sampling approaches. It continues with the sampling process, sample validation and non-response issues in sampling. Execution of the sampling process requires a detailed and insightful specification of how the sampling design issues will be implemented. The specification for the implementation of the sampling process is essential to ensure that the process takes place in a consistent manner. For instance, if the households are the sampling unit, it is important to specify what is meant by household. Once data is collected using a sample, comparisons between the structure of the sample and the target population should be made. Weighting is a process of statistically adjusting the collected data. Such adjustment is not always necessary, but it can increase the quality of data analysis. Weighting is a statistical process where each case of the respondent in the database is assigned a weight to reflect its importance over other cases of respondents. The value of 1.0 demonstrates the unweighted case. The impact of weighting is to enhance the number of cases in the sample which hold specific attributes. Weighting is adopted to make the collected data more representative of the target population regarding certain attributes. One possible use of weighting is to adjust the sample, so that importance is attached to respondents who possess specific attributes. For instance, if a research project takes place to identify the type of modification which should be made to a product or service, the researcher may wish to attach great weight to frequent users of the product or service. This can be attained by assigning weights of 3.0 to heavy users, 2.0 to medium users, 1.0 to light users and 0 to non-users. Weighting should be used with caution however, and it depends on the researcher’s experience how it will be used. If adopted, it should be part of the research report and should be very well documented. Considering that low response rates enhance the probability of non-response bias, an effort should be made to enhance the response rate. This can be attained through the use of incentives, well-designed questionnaires, and follow up. In a research project, increasing the response rate may not reduce the non-response bias if the additional respondents are no different from those who have responded but do differ from those who still do not respond. Types of Probability Sampling: Single Random Sampling, Systematic Sampling, Stratified and Cluster Each of the elements of the population carries an equal chance of being chosen for a probability sample. There are four broad types of probability sample, namely single random sampling, systematic sampling, stratified sampling and cluster sampling. Single Random Sampling In single random sampling, the known and equal probability is calculated as follows: Probability of selection=sample size/population size. For instance, if the population size is 8000 and the sample size is 400, the probability of selection is 5%, 400/10000= 0.5. If a sampling frame is available, the market researcher can select a simple random sample by adopting the following: Assign a number to each element of the population. For instance, a population of 8000 can be assigned numbers from 1 to 8000. Develop a table of random numbers, start at some arbitrary point and move down, up and across until 400 moving across equal digit numbers (that is the actual sample size) to be reached. The selected numbers from the table depict the certain population items which will be considered in the research project. A simple random sample ensures that each element of the population has an equal chance to be included in the sample. Systematic Sampling Systematic sampling is sometimes used as a substitute for simple random sampling. It generates samples similar to the ones produced from simple random sampling. To generate a systematic sample, the researcher should first assign numbers to the whole population as in simple random sampling. Next, the researcher determines a skip interval and chooses names based upon the selected interval. The skip interval can be calculated using the following formula: skip interval=population size/sample size. For instance, if the researcher uses a database and has computed a skip interval of 50, every 50th name will be selected for the sample. Adopting a formula ensures that the list is covered. Systematic sampling is considered simple and less time consuming than simple random sampling. Stratified Sampling Stratified sampling is a probability sample that is distinguished by the following processes: the original population is divided into two or more mutually exclusive and exhaustive subsets (e.g., males and females) and simple random samples of elements from two or more subsets are selected independently of each other. Thus, stratified sampling is a two-step approach where the population is divided into subpopulations or strata. Each element of the population should be allocated to one and only one stratum and no population elements should be omitted. As a next step, elements are chosen from each stratum using a random process. A stratified sample ensures that all the key subpopulations are included in the sample. Cluster Sampling In cluster sampling, the target population is initially divided into mutually exclusive and collectively exhaustive subpopulations. These subpopulations or clusters are considered to have the diversity of respondents held in the population. A random sample of clusters is chosen on the basis of probability sampling. For each selected cluster, either all elements are considered in the sample or a sample of elements is selected probabilistically. If all elements in each selected cluster are considered in the sample, the process is called one-stage cluster sampling. If a sample of elements is selected probabilistically from each selected cluster, the procedure is two-stage cluster sampling. The main difference between cluster sampling and stratified sampling is that in cluster sampling only a sample of subpopulations (clusters) is selected, whereas in stratified sampling, all the subpopulations (strata) are chosen for further sampling. We discuss non probability sampling, with a focus on convenience sampling and judgmental sampling. Convenient sampling, focuses on generating a sample that consists of convenience ingredients. In particular the interviewer identifies and select the sampling units, for convenient sampling respondents are selected if they are in the right place at the right time. For instance, convenient sampling may include a group of customers stopping in a certain supermarket. Or interviews of customers in the street without any checking whether they meet the eligibility criteria, for inclusion in the research, and as an outcome to qualify the respondents. A convenience sample can be also emerged by sending email surveys, to a sample of customers or inviting some customers to take part in an internet survey. Convenience sampling is the least expensive and least time consuming way of collecting data and survey research. The research participants are easily accessible and it is fairly easy to make them respond and collaborate. Although it has a number of advantages, convenience sampling is characterized by a number of limitations, including selection bias and lack of representativeness of the whole population. Selection bias appears when the sample is elected in a manner, where it does not represent the whole population. As a consequence, it is not easy to generalize the results across the whole population. Convenient sampling can offer some insights about the topic under investigation, and they can be used in protesting questioners or pilot studies. To do the limitations of convenience sampling, findings should be used without caution that mental sampling is a type of sampling. Where the population elements are selected based on the judgment of the researcher who exercised Jasmine or expertise. Select the elements considered in the sample based on whether they think the elements represent the whole population. Examples of judgmental sampling include testing a target market to identify the potential of a new product or service or product, testing without people who have high expectations of the product. Judgmental sampling is considered inexpensive and quick, but it is not indicative of the whole population. Dismantle sampling is subjective and its value depends on the researchers expertise and creativity. That's mental sampling can be used when the population is small. Quota sampling is considered a two-stage restricted judgemental sampling that is used extensively in street interviewing. Stage 1 The first stage includes the development of control attributes or quotas of population elements, including age and gender. In order for these quotas to be developed, the researcher should list pertinent control attributes and determine the distribution of such characteristics in the target population. For instance, 56% of females and 44% of males stemming from a sample of 1000 where 560 women and 440 men were selected. The quotas are sometimes allocated so that the proportion of the sample items containing the control characteristics is the same as the proportion of the population elements with these characteristics. The quotas make sure that the composition of the sample is the same as the composition of the population regarding the characteristics of interest. Stage 2 In the second stage, the sample elements are chosen based on convenience or judgement. Once the quotas are allocated, there is considerable freedom in identifying and selecting the elements to be included in the sample. The only requirement is that the elements chosen should fit the control characteristics. Even if the sample composition mirrors that of the population regarding the control attributes, there is no assurance that the sample is representative of the population. If a characteristic pertaining to the problem is overlooked, the quota sampling will be not representative. Pertinent control characteristics are often omitted due to the fact that there are practical difficulties relevant to the inclusion of specific control characteristics. For example, a sample representative of the various strata of economic and social classes in a population. In this vein, street interviewers approach potential respondents who believe that they fit into the quota they have been set. Could an interviewer guess which respondents fit within the various classes in the same manner, which they may guess the gender and age of respondents? The introductory questions of a street interview may focus on the establishment of attributes of potential research participants in order to identify the extent to which they fit a set quota. The levels of non-response and ineligibility levels identified by such a procedure, it is not considered an optimal solution. Due to the fact that the elements selected for each quota are chosen based upon convenience or judgement, several sources of selection bias may appear. The interviewers may go to selected areas where eligible research participants are likely to be identified. They may also avoid people who seem unfriendly, not well dressed or people who are based in undesirable locations. Quota sampling does leave space for sampling error. The main objective of quota sampling is to offer representative samples at a relatively low cost. Its key advantage is the lower cost and the greater convenience to the interviewers in choosing elements for each quota. Quota sampling can offer results similar to conventional probability sampling. SNOWBALL SAMPLING Market researchers can have difficulties in recruiting research participants to contribute to their research projects. In this case, market researchers can start considering snowball sampling. Snowball sampling is a type of non-probability sampling, also known as chain- referral sampling. Step 1 The first step in snowball sampling is for the market researcher to form a first group of research participants by identifying the individuals who will participate in the systematic investigation. Step 2 For the next step, the market researcher asks the research participants to recommend similar individuals who wish to participate in the market research project. As the existing research participants continue to refer appropriate research participants for inclusion, the number of participants or the ‘snowball’ continues to grow. Therefore, the market researcher has more data to work with. Snowball sampling is usually adopted to recruit research participants from rare or niche populations, which also represent a small percentage of the whole population. Due to the cost of identifying and recruiting such populations, the market researcher adopts snowball sampling where additional research participants are identified based upon referrals from initial research participants. The key advantage of snowball sampling is the elimination of search costs, but the market researcher may instead experience a drop in sample quality. The total sample may be biassed as the individuals whose names were obtained from the individuals sampled in the first stage may be too similar to the ones initially sampled. As a consequence, the sample may not be a good representation of the whole population. It is good practice to set some limits on recommending individuals from the initial participants but, it is not clear what rules should be set. The initial respondents may not always wish to obtain referrals. Type of Snowball Sampling There are three types of snowball sampling, those being, linear snowball sampling, exponential non-discriminative snowball sampling and exponential discriminative snowball sampling. Linear Snowball Sampling In linear snowball sampling, the researcher depends on a straight line referral sequence that commences with a single variable. The variable offers information regarding one other potential research subject and continues until the point where the desired sample population was attained. Under this type of snowball sampling, the market researcher does not need to commit considerable amounts of time, however, the chain can be interrupted if any of the existing individuals fail to obtain a referral. Exponential Non-Discriminative Snowball Sampling In exponential non-discriminative snowball sampling, the referrals are recruited in a geometric chain sampling sequence where the snowball sampling begins with the suggestion of one referral who in turn suggests multiple referrals. The process ends when the desired number of research participants is attained. Under this type of snowball sampling, the market researcher may identify the sample quickly, although individuals who are not a good fit may be recruited. Exponential Discriminative Snowball Sampling In exponential discriminative snowball sampling, the market researcher carefully screens the individuals before accepting them in the final sample. Therefore, only individuals who meet the screening criteria are permitted to be included in the final sample. This snowball sampling approach ensures that only the best research participants will be recruited to take part in the market research project. Simple size considerations. Simple size focuses on the number of elements included in the research study. For instance, 1,000 customers were included in the survey out of the population of 10,000. Determining the sample size and tells careful consideration of the following aspects. The quality issue to be considered when determining the sample size include, the nature of the research and the research context. For instance, exploratory research design, collecting responses from vegan customers. The number of variables of topics to be investigated, such as purchasing behavior, brand image, customer satisfaction, and e-word of mouth. The type of analysis or the size of the sample sizes used in similar studies, and resource constraints, including financial resources.This calls for large samples, which increases the cost of the research.The nature of research determines the sample size. For instance, in exploratory research using qualitative research, the sample size is small, in descriptive research, larger samples are necessary. Additionally, if a survey contains a lot of questions, a large sample is required. With large samples, the effects of sampling errors across variables are reduced. A sampling error occurs when the researchers select a sample that is not representative of the whole population. Using larger samples, requires the adoption of advanced statistical approaches to analyze the collected data, as well as the data to be analyzed in detail. Through the use of multi variant data analysis approaches using statistical packages such as SPSS or stadium. The size of the sample can be determined by the sample sizes used in similar types of research. Deciding the sample size is directed by resource constraints. It is essential for researchers to validate their sample that entails to screen respondents, to identify if they meet the eligibility criteria for inclusion in the sample. Respondents may be screened based on their demographics, including age, gender, experience. And other associated attributes to ensure that they meet the criteria for inclusion in the sample. Screening can limit inappropriate elements which are included in the sampling frame. The success of the selection and validation process of the sample should align with the attributes, which describe the structure and composition of the target population.. Non Response Aspects in Sampling It is important when sampling to improve the response rate, because when some of the potential respondents do not respond we get a non-response error. Non-respondents may be different from respondents in terms of certain demographics such as personality and perceptions, this in itself can provide useful insights. Where the non-respondents differ significantly with respondents, it means that the sample is significantly biased. Increased response rates usually imply lower rates of non-response bias. Ways to Increase Response Rates Response rates can be improved if the market researcher manages to cope with refusals and non-at-homes. Refusals stem from individuals’, who were considered in the sample, unwillingness to participate in the study. Market researchers should aim to reduce the refusal rates. This can be attained through prior notification, incentives, questionnaire design, administration and follow-up. Prior Notification Potential research participants are contacted in advance to be informed of the survey and to express their willingness or not to take part. Prior notification increases the response rate as the key informants are aware of the objective of the research and the possible benefits. How market research projects are introduced and communicated to respondents is crucial to encourage them to participate in the study. Incentives Response rates can be also increased by offering incentives to research participants. In this respect, monetary and non-monetary incentives can be offered to potential research participants. Monetary incentives can be prepaid or promised. For example, participants in the research study could receive a voucher from Amazon. The prepaid incentive should be included or attached with the survey or the research instrument. The promised incentive is sent to research participants who complete the survey or questionnaire. The most common non-monetary incentives include books, pens, pencils or a copy of the research results, for example. Questionnaire Design A well-developed research instrument can enhance the response rate and reduce refusals to participate in the research study. If the research instrument and the experience completing the questionnaire is pleasant for the research participants, the response rate may improve. Administration Well-trained researchers can enhance response rates and increase the conversion of key informants. Follow-Up Follow-up or contacting the non-respondents periodically works well to eliminate refusals and convince individuals to participate in the research project. A reminder could also be sent to non-respondents to complete the questionnaire. Follow-ups can be sent through emails. Another cause of non-response refers to the not-at-homes or people who are busy when the invitation to participate in the research project arrives. Consumers are usually more at home at the weekend instead of during the week. Therefore, before sending a questionnaire, the market researcher should have already contacted the key informant to ensure they are available to participate in the survey, online/face-to-face interview or focus group. Low response rates increase the potential of non-response bias. As a result, the impact of non-response should be estimated. This can be obtained by connecting the non-response rate to estimated differences between respondents and non-respondents. Combining Various Types of Data Different types of secondary data can be merged and presented in a way which depicts and communicates customer characteristics. Gathering customer data from different sources of information helps with market segmentation analyses. Focusing on the variables, which can be used as market segmentation variables, such as demographics, psychographic and geographic can offer a clear picture of target markets’ profile. The combination of survey data and databases helps the organisation to identify the nature and scope of different customer segments, understand the nature of the factors, which influence the needs of customer groups, and reflect on best practices regarding target market decisions. The actual combination of the various sources of information supporting decision-makers takes place through a number of steps that assist the integration. It is essential to analyse the current customer databases, focusing on data that considers the routine operational transactions or inquiries that are made to an organisation. Customer databases are considered as internal secondary data and represent readily available data offering the organisation easily accessible customer insights data. Additionally, it is important to focus on existing geodemographic profiles, which can be used to classify customers on the basis of their neighbourhood. If a set of neighbourhoods are similar across a broad range of demographic measures, they can obtain potential across products, brands, services and media. Maps can also be developed to illustrate different types of customer profiles and assess the potential in different locations. Further, it is essential to use surveys carried out by the organisation and analyse the data in line with the geographic data available. The combination of different sources of data can assist in developing the profiles of the various customers. In this respect, data from the geodemographic profiles of customers can be used in combination with recently collected data of customer perceptions and attitudes. The use of data warehouse analytics provides access to data from different sources of information and captures qualitative data that may emerge from intelligence and primary sources of information. The insights emerging from data warehouse analytics would address issues arising from target markets. Data warehouses are extensively used in the banking sector. A data warehouse helps to collect and integrate separate data and transforms it into a consistent format, which can facilitate decision making by offering information. As a consequence, a data warehouse includes a collection of different and integrated databases developed to support decision- making and problem-solving. The data included in the data warehouses can be accessed through data mining approaches. Data mining is defined as the process of identifying constructive associations, patterns and trends emerging from accumulated amounts of collected data stored in repositories, through the adoption of pattern recognition and statistical and mathematical approaches. Examples of data mining have the objective of classifying customers into certain categories, considered meaningful to decision-makers, for example: Identifying potential target markets with the attributes which decision-makers are looking for. Identifying the types of services or products or product lines that are purchased together. Data mining is an approach that helps to exploit accumulated data stored by organisations and helps the researcher to identify certain types of information, which help with more efficient decision making. EXTERNAL SECONDARY DATA: SYNDICATED SOURCES OF SECONDARY DATA In addition to published data and information available in computerised databases, syndicated sources are considered a key source of external secondary data. Syndicated sources refer to organisations that gather and sell pools of data and information developed to address information requirements and needs. This data is collected without a focus on specific market research related problems, however, the data and information gathered is offered to client organisations and can be used to serve certain needs and requirements. For instance, documents can be developed and organised with a focus on customer sales districts, locations, certain products or product lines. The use of syndicated services is less expensive than collecting primary data tailored to certain market research objectives. Syndicated services or sources can be divided based on the unit of measurement, which may be households or consumers or organisations. Consumer data can be gathered from surveys. The information refers to data including values and lifestyles, or purchasing patterns and behaviour. Apart from survey data, other types of syndicated sources are mail diary panels, scanner volume tracking data, audit services and industrial firms. Mail diary panels refer to information coming from households. This information is offered over an extended period of time. The key informants are required to record the behaviour as it takes place. The information can be used by the organisation to predict sales revenue, and market share as well as creating consumer profiles and creating brand loyalty. Scanner volume tracking data focuses on household purchases recorded through electronic scanners in supermarkets. This information helps to track prices. Audit services focus on verifying product movement by investigating physical records or performing inventory analysis. The data is used to measure sales, assess market share and track the performance of new products or services. Industrial firms focus on data banks of industrial establishments developed through the direct inquiries of companies, clipping services and corporate reports. The information and data can be used to assess market potential around geographic areas, defining sales territories and deciding on the advertising budget. Secondary data refers to information that has already been collected and may be relevant to the problem at hand. Data collected for a similar market research problem that has been investigated before can serve as a source of information for the research problem under investigation as a result, secondary data by nature can offer cost effective solutions and insights and is therefore considered as an effective means of obtaining information. There are two broad sources of secondary data. These being external organizations or individuals and internal databases. External sources of secondary data referred to government departments and agencies that collect and publish evidence regarding the consumer trends and market and industry data, amongst others. Internal secondary information, which originates from within the organization refers to annual reports, sales data profiles of customers, purchase patterns and in house magazines compiled by the company. All these internal publications and data are included in the company's internal database. Secondary data has a number of advantages for example, it helps to redefine the research problem during the exploratory research stage. Secondary data may flag potential problems and difficulties with the data collection. This may include difficulty in collecting the relevant responses or hostility of the target population. Secondary data may offer credibility regarding the research project and offer background information about it. Despite the advantages secondary data has a number of limitations, including lack of availability, lack of relevance and inaccuracy. For some research questions there is no available secondary data. In some instances, secondary data cannot be used because it is represented in units or measures that cannot be used by the researchers. Secondary data is sometimes inaccurate. This is due to the fact that potential sources of error exists when the researcher collects codes, analyzes and presents data. In order for researcher to assess the accuracy of secondary data, it is essential to identify who collected the data, what the objective of the research study was, what kind of information was collected, when the information was collected and how the information was collected. Secondary data can be collected through the use of a number of data collection tools. They can be collected through the use of bots, where programs have created bots to do automatic web scrapping. Identifying data relevant to the research. Through the use of mobile devices, including a laptop or a mobile phone, researchers can gather relevant to the research projects secondary data such as journals, articles, websites, blogs, social media. Libraries constitute a traditional secondary data collection tool where researchers can collect secondary data from reports, journals, magazines and books. The nature and scope of secondary data. Secondary data refers to information that has already been collected and may be relevant to the problem at hand. Data collected for a similar market research problem that has been investigated before can serve as a source of information for the research problem under investigation. As a result, secondary data by nature, can offer a cost-effective solutions and insights and is therefore considered as an effective means of obtaining information. There are two broad sources of secondary data, these being external organizations or individuals, and internal databases. External sources of secondary data refer to government departments and agencies that collect and publish evidence regarding consumer trends and market, and industry data amongst others. Internal secondary information which originate from within the organization, refers to an annual reports, self data, profiles of customers, purchase patterns, and in-house magazines compiled by the company. All these internal publications and data are included in the company's internal database. Secondary data has a number of advantages, for example, it helps to redefine the research problem during the exploratory research stage. Secondary data may flag potential problems and difficulties with data collection. These may include difficulty in collecting the relevant responses or hostility of the target population. Secondary data may offer credibility regarding the research project and offer background information about it. Despite the advantages, secondary data has a number of limitations, including lack of availability, lack of relevance, and inaccuracy. For some research questions, there is no available secondary data. In some instances, secondary data cannot be used because it is represented in units or measures that cannot be used by the researchers. Secondary data is sometimes inaccurate, this is due to the fact that the potential sources of error exists when the researcher collects, codes, analyzes, and presents data. In order for a researcher to assess the accuracy of secondary data, it is essential to identify who collected the data, what the objective of the research study was, what information was collected, when the information was collected, and how the information was collected. Secondary data can be collected through the use of a number of data collection tools, they can be collected through the use of bots, where programs have created bots to do automatic web scraping. Identifying data relevant to the research. Through the use of mobile devices, including a laptop or a mobile phone, researchers can gather relevant to the research project secondary data, such as journals, articles, websites, blogs, social media. Libraries constitute a traditional secondary data collection tool where researchers can collect secondary data from the reports, journals, magazines, and books. Research Drives Business Strategy A company keeping up with its customers needs and requirements, as well as market trends is a matter, not only of market share and profitability, but also of survival and long-term sustainability and viability. Market research helps companies to progress in the right direction. Companies can use the collected information to improve strategic decision-making and customer satisfaction. Market related information and data can assist in designing and developing better products and achieving a competitive advantage. The organisation should reflect on how collected market research data will be used to enhance value. Gathered market research data should be integrated with other research data coming from both internal and external sources, to make the most of their use and to make strategic decisions. The research budget should also be considered. If there are limited research resources, more cost efficient ways of collecting data should be considered, including automated data collection approaches. Accurate market research data helps companies to better connect with their existing customers, as well as attracting new customers. Both primary and secondary research data helps companies to improve processes, products and services. Exploratory research is the starting point for innovation. The combination of primary and secondary data allows effective decision-making and is a perfect starting point in realising past and existing performance trends. Additionally, it is essential for a company to link its existing knowledge of research with future research efforts, to offer a comprehensive and holistic approach to problems and address each of them. When starting new research projects, companies must clearly define the research goals and objectives. The collected information should be specified by how it will be used and how it will inform effective decision making. It is also important to prioritise focus on the research. In this respect, the focus of the research project should be specified, as well as the sources of the data and information collected. These could be social media platforms, trade associations, government data, or existing company data. If data is gathered through an external market research company, it is essential to understand the ways in which the company operates and the research needs. The key objective of research is to bring data and information that will bring actionable knowledge and insights within the company and will improve decision making and organisational outcomes. Research that adapts experimentation requires the researcher to be an active participant in the research project. In experiments, the research approach focuses on one variable that is manipulated and the impact on another variable is observed. Experiments can take place in labs or in the field. Laboratory experiments offer several advantages. The most important is the capability to control extraneous casual factors. And focus on the impact of a change on the development on the dependent variables. In the lab, a researcher can efficiently deal with additional independent variables. On the other hand, the full controlled environment of the lab may not be as good as setting as the market plays. Due to this, the findings of the laboratory experiments are a bit problematic when they are transferred to the market. Field experiments take place outside the lab, in a real market setting. Failed experiments address the problem of the dynamism of the environment, but they can set the basis of new problems. In the field, the researcher cannot control all factors which may affect the dependent variable. Including the actions of the competitors, the economy or the legal environment. As a result, the field experiments have more problems concerning their internal validity. Whereas lab experiments have more issues regarding their external validity. Validity is defined as the degree to which an experiment measures what the researcher is attempting to measure. The validity of a measure is dependent on the degree to which the measure is free from systematic and random error. Two different types of validity are associated with experimentation. Those being internal and external validity. Internal validity focuses on the degree to which connecting explanations regarding the experimental results observed can be ruled out. If the researcher can show that the experimental variable produced the differences observed in the dependent variable, then the experiment is considered to be internally valid. The internal validity requires evidence that demonstrates that variation in the dependent variable is caused by exposure to the treatment variable. And no other possible factors. External validity focuses on the degree to which the causal relationships that are operationalized in an experiment can be generalized to external people, environments and settings. Focus is devoted to ensure that the subjects and setting adopted in the experiment, a representative of the population. And environment the researcher would like to project the result. Field experiments are characterized by an increased degree of internal validity compared to lab experiments. MEASUREMENT Measurement focuses on assigning numbers or other symbols to attributes of objects to align with pre-specified rules. Measurement concentrates on or captures some characteristics of the object. As a result, market researchers do not measure consumers but their attitudes, beliefs and perceptions. In market research, numbers are assigned to allow statistical analysis of the collected data. Numbers make consistent communication of measurement rules and properties easier. The allocation procedure should be isomorphic, for example, the same dollar figures should be allocated to households with similar annual incomes. As an outcome, the assigned numbers will be related to certain attributes of the measured object and the opposite. Moreover, the rules for assigning numbers should be standardised and applied in a uniform manner. Consider a scale of measurement to determine consumer attitudes towards a certain product line. Each key informant is assigned a number that depicts an unfavourable attitude (measured as 1) a neutral attitude (measured as 2)or a favourable attitude (measured as 3). Measurement is considered the allocation of 1, 2 or 3 to each key informant. Scaling Scaling can be considered as an extension of measurement and focuses on creating a continuum on which measured objects can be placed. Scaling is defined as the procedure of placing key informants on a continuum to demonstrate their attitude towards a certain product or product line. Scaling determines how each respondent would be classified, whether they have an unfavourable, neutral or positive attitude to the product. Key Scales of Measurement There are four key scales of measurement namely, nominal scale, ordinal scale, interval scale and ratio scale. Nominal Scale A nominal scale is a figurative labelling scheme in which numbers serve as labels or tags to identify and classify objects. The numbers assigned to the key informants in a study constitute a nominal scale. For example, a male research participant may be allocated number 1 and a female participant number 2. When a nominal scale is adopted with the objective of identification, there is a one-to-one correspondence between the numbers and the objects. Each object only has a number assigned to it. Examples of nominal scale are student registration numbers at a university. Ordinal Scale An ordinal scale is a ranking scale in which numbers are allocated to objects according to their attributes and their relative position in the attributes compared to another object. An example of an ordinal scale is the ranking of service quality offered by a number of banks. Interval Scale An interval scale is a scale where the numbers used demonstrate equal distances in the attribute being measured. For instance, the distance between the scale values is considered the same, thus the distance between 1 and 2 is the same as the distance between 5 and 6. A typical example of an interval scale is the temperature scale. Ratio Scale A ratio scale has all the characteristics of a nominal, ordinal, and interval scale and also has an absolute zero point. As a consequence, ratio scales can rank objects, and compare intervals or differences. Examples of ratio scales are age and weight. The different scaling approaches that are adopted in market research can be classified as comparative or non-comparative scales. Comparative scales entail the comparison of different objects. For instance, key informants may be asked to indicate if they prefer Coca Cola or Pepsi. Collected comparative data have ordinal or rank order characteristics. Due to this, comparative scaling is also referred to as non-metric scaling. Comparative scaling includes paired comparison, rank order, constant sum scales and Q-sort. Comparative scales compare stimulus objects, and key informants are forced to choose between them. Paired Comparison Scaling Paired comparison scaling focuses on two objects, on which the respondents are asked to select one. Collected data are ordinal. A key informant can express that they prefer Gucci compared to Louis Vuitton, for example. Paired comparison is the most popular paired comparison scale. Paired comparison data can be converted to rank order. Transitivity of preference assumes that if brand A is preferred to brand B, and brand B is preferred to C, as an outcome brand A is preferred to brand C. Paired comparison scaling is useful when the number of brands is limited because it necessitates direct comparison. An example of paired comparison scale is the following: Question: Please find below a pair of brands. Please indicate which brand you prefer from each pair. Rank Order Scaling Rank order scaling is a comparative scale in which respondents are presented with a number of items at the same time and are asked to order or rank them according to some criterion. For example, key informants may be asked to rank them with respect to their overall preferences. Such rankings typically emerge by requesting respondents to assign a rank of 1 to the most preferred brand, 2 to the second most preferred, and so on until the rank number is allocated to the least preferred brand. Constant Sum Scale A constant sum scale is a comparative scale in which key informants are asked to assign a constant sum of units such as points, euros, or dollars to a set of stimulus objects according to some criterion. For instance, respondents may be asked to allocate 50 points to characteristics of a product to demonstrate the importance which they attribute to each characteristic. If an attribute is not important to the respondent, they assign the number 0. If a characteristic of the product is very important to the respondent, they can assign double points. The number of the assigned points should be 50. Q-Sort Scaling Q-sort scaling is a comparative scaling approach that adopts a rank order process to sort objects on the basis of similarity against some criterion (e.g., best to worse). For instance, key informants are given 50 attitude statements (e.g., brand X is of high quality or brand X is the best in the market for clothing) asking them to place them into 10 categories, following a range from ‘most highly agreed with’ to ‘least highly agreed with’. The number of objects sorted should be between 60 and 90. The number of objects in each category is pre-specified. This ensures a normal distribution of objects on the whole set. we will discover the role of internet in market research, focusing on survey research on the internet. The internet has transformed the way in which survey research takes place. Nowadays, the majority of surveys take place online. Because the number of internet users is constantly increasing, online surveys have become an increasingly popular primary data collection tool. There are different ways to conduct online survey research. This includes web survey software, survey design websites, and web hosting. Web survey software refers to software systems that are developed for questionnaire construction and delivery. The survey instruments are developed with an easy-to-use edit characteristic within a virtual interface and then automatically transferred to a Web server system. The Web server distributes the research instrument and responses in a database. The researcher can have access to data including questionnaire, completion data, and descriptive statistics on responses at any time. An example of survey research software includes SPSS conquest. To encourage participants to complete the online surveys, the researcher should offer a good experience to keep people engaged. The more engaged participants are, the better quality the insights will be. To provide a pleasant experience, the language which will be adopted in the questionnaire should be simple and understandable to the participants. Additionally, the researcher should be honest about the time duration required to complete a survey. Opportunities for open-ended answers should be offered to participants as well as the chance of expressing themselves and their perceptions in the best possible manner. In this respect, space should be offered to them in the research instrument to express their point of view and to contribute effectively to the objectives of the research project. The surveys should be kept less than 20 minutes and offer the participants progress information and time remaining to complete the online survey. It is also advisable to use online graphics in order to make the online surveys experience more appealing. Interaction between the researcher and the participant will enhance the response rate of the survey. It is also important to make the online survey more informative, and this will encourage the research participants to acquire new knowledge and valuable information regarding the various sections of the online survey. Incentives can also stimulate engagement and participation in online research. Many websites permit the researcher to develop an online survey without the adoption of online software. The survey is then administered on the design of the site's server. Well-known sites are Survey Monkey and Qualtrics. Such services also offer options of exporting the collected data directly to data analysis software packages such as SPSS and STATA. Non-comparative scaling techniques include the semantic differential scale, the Stapel scale and the Likert scale. The semantic differential scale was developed with the objective of measuring the meaning of an object from the consumer’s point of view. For example, the object could be a person and the meaning of the perceptions from a group of customers. Developing a semantic differential scale begins with a concept that should be rated, such as the image of a brand or the quality of service. The market researcher chooses opposite pairs of words or phrases (for instance, reliable (1) to unreliable (7)) that capture the concept. As a next step, respondents rate the concept on a scale which is usually anchored from 1 to 7. The semantic differential scale is an effective way to measure the strengths and weaknesses of a product or brand image against competitors. The semantic differential scale is reliable and valid for decision-making as well as prediction in market research. It can be also used to generalise findings from the sample to the whole population. The Stapel scale is an adjustment of the semantic differential scale. In the Stapel scale, a single adjective (e.g., user friendly, high quality) is allocated to the centre of the scale. The scale is usually a 10 point scale that ranges from +5 to -5. The approach is designed to operationalize the direction and the intensity of the attitudes at the same time. The Likert scale consists of a number of statements that communicate a favourable or unfavourable attitude towards the concept under investigation. The key informant is asked to indicate their level of agreement or disagreement to each statement by assigning a numerical score. The scores are used in total in order to measure the key informants’ attitudes. The Likert scale requires the respondent to take one statement into consideration at a time. The scale runs from one extreme to the other anchored from strongly disagree (1) to strongly agree (5). A number of statements are available to the respondent and there is a set of standard replies for the key informant to choose from. Likert developed this scale to operationalise an individual’s attitude for concepts and assess their perceptions towards the point of reference. An example of Likert type scale is depicted below: Question: Please indicate your level of agreement of disagreement with the following statements, by ticking the number that best reflects your point of view: Strongly disagree Strongly agree Brand X procuses high quality cars 1 2 3 4 5 Brand X offers an excellent after sale service 1 2 3 4 5 Brand X has reasonable prices for its cars 1 2 3 4 5 Likert considered the following when constructing the scale: First, identify the concept or activity which will be scaled. Second, a large number of statements can be used (around 100) about the concept or the activity under investigation. Third, a pre-test takes place, this includes the whole set of statements and a certain number of respondents. Fourth, during the pre-test the key informant indicates their agreement with each of the statements or items selecting one of the following descriptors, strongly agree, agree, undecided, disagree and strongly disagree. Five, each answer is assigned a numerical weight (for instance, 1, 2, 3, 4, 5). Six, the respondent’s total attitude score is depicted by the summation of weights related to the items checked. For instance, if 5 was assigned to ‘strongly agree’ for favourable items, 5 should be allocated to ‘strongly disagree’ for unfavourable items. Seven, after the pretesting the researcher selects the items that appear to discriminate between high and low total scores (for instance strongly disagree 1 to strongly agree 5). The final 20 to 25 items selected depict the greatest differences in mean values. Steps 3 and 5 are repeated in the main study. COLLECTING DATA THROUGH OBSERVATION Observational research focuses on the various types of non-experimental research in which human behavior is observed and recorded. Observational research aims to observe a single object or a set of objects. Observational research differs from the experimental approach because the observation takes place without any manipulation, intervention, or control from the market research team. The data that is collected through observational research is usually qualitative, but it can be also quantitative, or mixed method. There are different approaches to observational research including naturalistic observation, participant observation, and structured observation. Naturalistic observation is a type of observation where the market researcher observes the behavior of people in their natural environment. This makes naturalistic observation a type of field research. Naturalistic observation includes, for example, observing customers in a retail outlet, or tourists in a hotel. Naturalistic observation is often undertaken without the individuals being aware they are being observed by researchers. It is called distinguished naturalistic observation. This approach is acceptable as long as the research participants are anonymous, and their study behavior occurs in a public place where individuals do not have expectations of privacy. For example, loyalty cards can be used to monitor the behavior of retail outlet shoppers. Consumer behavior can be monitored by both retail employees and researchers. Although distinguished, naturalistic observation is considered acceptable. It is not always practical or feasible to observe research participants without their permission. When participants are aware of the researcher or that their behavior is being observed, this is called undistinguished naturalistic observation. This main issue with undistinguished naturalistic observation is the concern that individuals do not act naturally if they are being observed. For instance, customers may react differently if they're aware that they are observed and their behavior and reactions are recorded. Another way to collect data in observational research is through participant observation. For this approach, researchers take part as active participants in the research project. Data collected under the approach includes interviews, as well as notes based on observations and interactions. Participant observation can also be classified into distinguished and undistinguished. In distinguished participation observation, researchers act as members of the group they are investigating while concealing their role as researchers. In undistinguished participant observation, the researchers are part of the group they investigate, however, they reveal their real role to the group members. Ethical issues may arise when the researcher becomes a member of the group under study without the group members being aware of it. Another observation method is structured observation. Under this observation method, the researcher makes careful observations of certain behaviors in a specific setting. The setting is more structured than both the naturalistic and participant observation settings. Usually, the setting used under structured observation is a laboratory. Alternatively, the researcher may observe the individuals in a natural setting that is characterized by some structure, such as a certain task in which research participants have to be involved. The structured observational approach focuses on collecting quantitative rather than qualitative data. The researchers under this approach are focused on a limited number of behaviors to ensure its behavior is effectively monitored Data Collection: Case Study The objective of ethnography is to find out how the culture works, rather than gain insights on an issue or a problem using the case as a reference point. The case study entails studying an issue that is investigated through either one or multiple cases within the realms of a system, that may be a setting or a context. A case can be considered a strategy of inquiry, methods approach or a comprehensive research strategy. A case study is a qualitative approach where the investigator investigates a case, or several bounded systems simultaneously (cases) through in-depth data and insights using different and multiple sources of information including, observations, interviews, audiovisual, and documents and reports. The case study reports a case description and associated case-based themes and issues. The different types of case studies are characterised by the size of the bounded case, whether the case considers one individual, a number of individuals, an entire programme or an activity. They may be classified according to the intent of the case analysis. In terms of intent, a case study can be classified as a single instrumental case study, the collective or multiple case study and the intrinsic case study. Single Instrumental Case Study For a single instrumental case study, the researcher concentrates on an issue or concern (e.g., the digital strategy of a brand) before choosing a case study to depict the case. Collective Case Study In a collective case study or multiple case studies, the issue for investigation is chosen but the researcher chooses a number of case studies to demonstrate the issue. This helps to investigate different aspects of the issue. Qualitative researchers are reluctant to generalise the cases from one context to another because the contexts of the various cases are different. If the intention is to generalise, the qualitative researcher should select representative cases to be considered in the qualitative study. Intrinsic Case Study The intrinsic case study focuses on the case itself if the case presents an unusual or unique situation. This situation may be to study a business customer who is unique or evaluate a marketing programme targeted to mass markets (e.g., social media campaign). For instance, the purchasing behaviour of another organisation that is a large customer. The researcher should offer a detailed description of the case following the case study analytic procedures. When conducting a case study, the researcher should determine the most appropriate research approach to address the problem. A case study is an appropriate approach to research when: Cases have been identified When the aim is to understand the case in-depth compared to other cases. Researchers should identify the case(s) which they will investigate. This may include an individual, a number of individuals, a programme, an incident or an activity. The researchers should decide on the type of case that may be collective or single, across several sites or within a site. After the case study is chosen, judgemental sampling can be used. Collecting data under a case study methodology, is an extensive process and is based on multiple sources of information including interviews, observations, documentation and audiovisual. Such data are assessed and analysed through a holistic analysis of the case or an embedded analysis of a certain aspect of the case. Through the data collection, a thorough presentation of the case communicates the different aspects of the history of the case, the chronology of the events or a day-by-day description of the events or activities in the case. The researcher can analyse the different themes that have emerged and to better understand the complexity of the case and to strive and identify similar themes that are encountered across different cases. When multiple cases are selected, it is important to offer a detailed description of each case and the themes within each case. Such an action is called within- case analysis and as a next step a thematic analysis across the cases should take place. This is called cross-case analysis of the meaning of the case. The final stage reports the meaning of the case, whether the meaning comes from learning regarding the issue of the case or learning regarding an unusual situation. The final stage in a case includes the lessons learned from the case. DATA COLLECTION: CONTENT ANALYSIS Content analysis is an approach which is based on the manual or automated coding of transcripts, documents, articles, radio, video and newspaper content and material. It can be to collect and analyse all transcripts from qualitative interviews but it can be used to analyse documents or any other type of textual data. The aim of content analysis is to reduce the amount of collected information and data into a manageable level. The textual information and insights can be transformed into manageable data and in this vein they can be used for additional statistical analysis. The two key aspects to reduce the amount of collected information are meaning condensation and categorisation. Content analysis is an approach which can assist in systematically analysing and investigating media coverage regarding a product. As a first step the researcher can simply count how many times or how often the word that characterises the product appears in articles, magazines or any type of text or audiovisual material. A simple form of content analysis would count the word which refers to the product under study through the analysis of how often the product name appears across different types of media. A more advanced approach of content analysis can investigate the different contexts where the word relevant to the product appears. The collected information can address different types of questions including media coverage such as: Why are some media positive towards the product? Why are some media negative regarding the product? How can the diffusion of information be attained through media? Why do some media publish information earlier than others? Which are the attributes of media coverage? What are the factors affecting media coverage? The sources that can be used in content analysis are archival material, recordings of conversations, responses from unstructured interviews. The researcher should define the sources which will be used and the type of information which is being looked for. The important aspect of qualitative data analysis is the coding of the interview transcripts or any other kind of documentation. Coding is construed as the process of categorising and organising collected data around themes and ideas and soon after marking similar passages of text with a code label. All fragments which have the same code, share a similar theme or idea. In a prescriptive analysis there is need to define words and phrases that are searched through the different texts apriori whereas in an open analysis, an attempt is made to distil the broad message of the text. In prescriptive analysis the codes are defined in advance (e.g., predefined), whereas in open analysis the codes come from the search process. In particular in open analysis the researcher should develop codes in the sourced materials along with the collection and analysis. There are a number of software packages which can assist with coding including ATLAS and NVivo DATA COLLECTION: PHENOMENOLOGY Phenomenological research describes the meaning for a number of individuals regarding their lived experiences of a concept or a phenomenon. Phenomenologists focus on the description of what all research participants have in common during their experience of a phenomenon. The key objective of phenomenology is to limit the individual experiences regarding a phenomenon to a description of the essence of the phenomenon. In this respect, phenomenologists identify a phenomenon, which serves as an object or the unit of analysis and is related to an individual experience. The researcher gathers data from individuals who have experienced the phenomenon, and as a next step constructs a description of the essence of the experience for all of the individuals who participated in the research study. The description includes what the individuals experienced and how they experienced it. There are two approaches to phenomenology namely, hermeneutic MEMOLOGYphenomenology, and empirical, transcendental or psychological phenomenology. Hermeneutical Phenomenology Research Hermeneutical phenomenology focuses on lived experiences and interprets the texts of life or hermeneutics. Under the spectrum of hermeneutical phenomenology research is a dynamic interplay of different research actions. The research commences when researchers turn to a phenomenon which is og interest to them (e.g., consumer experiences with a product). During the process the research participants reflect on key themes which constitute the lived experience. The phenomenologists develop a description of the phenomenon keeping a strong association to the topic of inquiry and keep a balance of the various sections to the issue under study. Phenomenology constitutes an interpretive process where the researcher makes an interpretation regarding the meaning of the lived experience of the research participants. Transcendental or Psychological Phenomenology Transcendental or psychological phenomenology focuses less on interpretations of researchers and has a more focused orientation to the experiences of the participants. Additionally, under the spectrum of transcendental phenomenology, everything is perceived as ‘fresh’ and the experiences of the researchers are left aside. In a number of times researchers start by describing their experiences with the phenomenon and in this respect they bracket their views before moving forward with the experiences of the research participants. In transcendental phenomenology the research effort commences with the identification of the phenomenon to study, bracketing out the experiences of an individual and as a next step gathering data from a number of people who have experienced the same phenomenon. As a next step the researcher analyses the data by eliminating the collected information to key statements or quotes and combines the statements into different themes. As a subsequent step the researcher constructs a textural description with respect to the experiences of different individuals (e.g., what consumers experienced regarding a certain product). Additionally, the researcher develops a structural description of the individual experiences (e.g., how they experienced the phenomenon in terms of conditions, situations, or context). The combination of both the textural and structural descriptions offer the whole essence of the experience with the phenomenon. The process of conducting a phenomenology study commences with the identification of the issue and that the use of phenomenology is the best way to investigate the selected research issue. Then the broad assumptions of phenomenology are specified. A researcher can fully describe how the research participants view the phenomenon whereas researchers should bracket out their own experiences. Data are gathered from research participants who have experienced the phenomenon. Data can be collected through interviews in-depth and multiple interviews with research participants. A number between 5 to 25 participants who have experienced the phenomenon are selected to participate in the interviews. Alternative data collection tools can be observations, journals, reports, and music. The research participants are asked two general questions, those being: what have you experienced in terms of the phenomenon? What contexts or situations have typically influenced or affected your experiences of the phenomenon? Other open ended questions can be asked which will assist in collecting the data. Building on the data from the first and second questions, data analysts go through the data (e.g., interview transcripts) and highlight important statements, sentences or quotes which offer an understanding on how the research participants experienced the phenomenon. This is called horizontalization. As a next step the researcher develops clusters of meaning from the key statements into themes. The significant statements and themes are used to write a description regarding the experiences of the participants. This is the textural description. The researcher also uses the statements to write a description in the context or setting that influenced how the participants experienced the phenomenon. This is called imaginative or structural description. +-] In the methods section the role of the researcher should be also discussed. Using the structural and textural descriptions the researcher then can develop a composite description where communicates the essence of the phenomenon. This is called essential, invariant structure or essence. An action research project focuses on real time change through the action research cycles, including the general objective pre step and the four key stages of constructing, planning, action and fact-finding. Such stages describe how the project is conceived, the objectives of it, the cycles of action and the results. The action research project focuses on observable behavior. Framing an action research project focuses on defining an issue that includes components of a solution. The issue may be observing the purchasing behavior of a group of customers about a product or a service. The content of the selected issues should be defined, for example, observing a group of customers in a retail outlet. Reframing a process, where a researcher questions a current frame and potentially discards it to adopt another one. Through the action research project, it may be necessary to reframe and make changes to affect the purchasing behavior of a group of customers in a positive direction for the brand or a company. When attempting to frame the action research project, it is essential to identify a wide and diverse array of issues and to understand that any issue selected for investigation may be embedded in a set of related issues