The sampling design process

Get Complete Project Material File(s) Now! »

Method

In this chapter the methodology for the thesis is presented regarding the choice of research approach. This is followed by the sampling design process and the information source which includes search parameters and search strategy. Furthermore methods of data collection and analysis of data will be presented. The last part of this section includes reliability, validity and generalization. Together all sections will provide the reader with an overview of the methodology of the thesis were each approach is presented and argued for.

Research approach

Research is defined as”…a systematic investigation to find answers to a problem.” (Burns 1990, p 1) When conducting a research, the literature distinguishes between two different approaches, qualitative and quantitative. Qualitative approaches are primarily but not necessary related with interpretivism and quantitative approaches are mostly related to positivism. Regardless of approach the fundamental of each approach is a problem which will be investigated but through different mode of procedure (Williamson, 2002). In the following sections we will present both the qualitative and the quantitative approach.

Qualitative and quantitative approach

Qualitative data is collected in form of words. To collect qualitative data an interpretive approach has to be taken which includes techniques such as interviews and observations to collect meanings and thoughts created by people (Williamson, 2002). When collecting qualitative data the answers could not be restricted of that many limits. The answers should instead be seen with a lot of transparency. A qualitative method puts effort into details and the uniqueness from the answers (Jacobsen, 2002). Collecting qualitative data could be done through for example an open-ended questionnaire or longer interviews trying to analyse the answers quite thoroughly. Qualitative data is often analysed through the use of conceptualisation (Saunders, Lewis & Thornhill, 2009).
According to Jacobsen (2002) the qualitative methods could contribute the correct understanding of a situation since it is those who are investigated or interviewed that have the possibility to define what the correct situation is. When investigating it often occur a kind of proximity because of the one who investigate another person tries to get a deepened view of the other person’s thoughts to get a good and reliable result of the research. Jacobsen also mentions that a qualitative method is flexible because the problem that we want to find answers of could change with time. This is due to the received answers can for example give a new view of the problem that not was the original thought. Some disadvantages with qualitative methods are that it often take long time, the investigated group of people is small compared to quantitative method and the answers are often relative hard to interpret and could easily been understood wrong.
Quantitative data is primarily collected through methods such as questionnaires which simplifies the results in term of numbers and is analysed by using statistics and diagrams (Saunders et al., 2009). The use of quantitative method provides the advantage and the ability to reach a high number of respondents. From this it is possible to relatively easy standardise and generalise the information from the respondents and get a clear overview of a wide range of data. The quantitative method is assumed to describe the scope and scale of a phenomenon and different correlations relatively exact with high extern validity. The distribution of answers to a specific question can be presented accurately by number or percent given a known degree of uncertainty (Jacobsen, 2002).
Possible disadvantages with the quantitative approach may be the fact that it could give the research a shallow character. Since the method comprise a high number of respondents in the research it is hard to explore different correlations deeply. Therefore it is needed to keep the measured correlations relatively simple. Another risk with the method is that the researcher in advance defines what is relevant to answer in the investigation. A questionnaire consists of questions and answering alternatives which the researcher consider relevant. This means there is no room for anything that is beyond what is covered in the questionnaire. To some extent the quantitative method therefore only provides answers to what the researcher actually asking for, which sometimes may affect the internal validity (Jacobsen, 2002). Even if a research approach is mainly quantitative it is important to emphasize that it may still not meet all criteria of positivism. This may for example be the case with some surveys where the collected data is quantitative but the samples are not selected randomly (Williamson, 2002).
The research of this thesis is mainly based on a quantitative approach since one important part of the aim with this study was to follow up and analyse the effects of the campaign Vi saknar dig by Destination Jönköping. As the sample consisted of 103 persons, the assumption was that the quantitative approach was most appropriate to answer the purpose based on the limited resources that were available. The choice of method made it possible to standardise and generalise the collected data from a high number of respondents in an easy and structured way. On the other hand it was not possible to get a deeper insight in the context, attitudes and motives through the quantitative approach from each and other of the respondents of the study.

The sampling design process

Since quantitative methods mainly are analysed through statistics, there are two important concepts to be aware of, population and sample which is clarified in figure 3.1. Many research projects with a quantitative approach aim to collect characteristics and information of a population (Malhotra, 2004). A population consist of the total number of elements that share common characteristics in which the investigator is interested in. To be able to draw conclusions from the population a sample which is a selected subset from all elements within the population is used (Aczel & Sounderpandian, 2009). In some cases it is needed to use a sample since it is not possible to investigate the whole population. The major reasons to use a sample and not the whole population is that it would be too costly, impractical and very time consuming (Saunders et al., 2009).
In the following section we will clarify and explain the chosen sampling design process of this thesis, which mainly is based on a five step model presented by Malhotra (2004). In figure 3.2 is the five different steps presented in a clear manner to get an overview of the sampling design process. Furthermore is each step presented more in depth in its own section.

Define the target population

To be able to conduct the research through a quantitative approach a target population has to be defined. From all elements that share the same characteristics of the population a sample is selected in order to be able draw conclusions and generalise from the whole population (Aczel & Sounderpandian, 2009). In many cases it is necessary to use a sample since it might not be possible to investigate the whole population due to that it would be too expensive, impractical and require too much time (Saunders et al., 2009). Our underlying criteria for the target population of this thesis was all people that have been participating in one of the different parts of the campaign Vi saknar dig by Destination Jönköping. The campaign was launched in November 2010 and has been implemented in different steps. Within the different steps of the campaign that we have investigated, 185 persons have been participating in different age groups from 18 years up to 50 years and older. To summarize, the target population consist of 185 elements where each element represent one of the persons that have been participating in the campaign Vi saknar dig from November 2010 and forward.

Determine the sampling frame

The purpose of a sampling frame is to facilitate and make it possible to identify the different elements of the target population. Examples of different types of sampling frames could be telephone lists, mailing lists or a map to mention a few (Malhotra, 2004). The sampling frame of this research is based on a telephone list provided by Destination Jönköping. The list contained names and contact details of all persons that have been participating during the campaign. From the telephone list we sorted the names by age to be able to choose the most relevant elements for the research which was people between 25 and 50 years.

Select a sampling technique

It exist different types of sampling techniques and they are primarily divided in two different kinds of categories, probability and non probability techniques. Characteristic for probability techniques such as simple random, systematic, stratified and cluster sampling are that the probability is equal for each element of the population to be selected. Convenience, judgmental, quota and snowball sampling are examples of other techniques that instead are related to non-probability sampling techniques (Malhotra, 2004; Collis & Hussey, 2003). Characteristic for these types of techniques are that the probability is not known for each element of the total population to be selected in contrast to probability sampling techniques. Instead it is possible for the researcher to do personal judgements. This means that the elements to be included or not included in the sample can be decided by the researcher without any known probability as the basis (Malhotra, 2004). There is of course different advantages and disadvantages connected to each type of sampling technique. Since the probability is known for each element being selected from the total population when using probability techniques it is possible to estimate characteristics from the whole population by using the sample. This is not possible for non-probability techniques. This is due to the fact that it is not known what the probability is to be selected for each element of the total population. Even if it not is based on statistical grounds it is still possible to use the samples from non-probability techniques to generalise about the population (Saunders et al., 2009).
When the elements of the population are selected by criteria’s and judgment of the researcher the sampling technique is called judgemental sampling. Since the elements to be included in the sample are chosen by the researchers it is a non-probability sampling technique. This type of sampling technique is often used in commercial marketing research projects. One potential disadvantage with judgemental sampling is that it is not possible to make direct generalisations to a specific population. On the other hand, the technique provides advantages such as that the sampling is quick, require low cost and is convenient (Malhotra, 2004).
Based on the circumstances that we in cooperation with Destination Jönköping decided to investigate their primary and secondary target groups which cover persons between 25 to 50 years, we had to select persons from the telephone list within the correct age groups. Due to this and our limited resources of time and finance the assumptions was that judgemental sampling was the most appropriate sampling technique for this thesis.

Determine the sample size

The size of the sample is a complex but important part of the sampling design process. This means that the researcher needs to make a decision of how many elements from the population to be included in the sample, which is mentioned as determining the sample size. Different kind of studies and the constraints of resources such as time and financing available are two factors that affect the sample size. Often could similar studies be used as rough guidelines (Malhotra, 2004). It is likely that the trustworthiness of the conclusions of the population will increase in line with the sample size. To make a statistical analysis a rule of thumb is that if the sample size is 30 or greater it is sufficient to assume that it is normally distributed (Saunders et al., 2009). Since the population consist of 185 elements we assumed that a sample of 103 elements would be sufficient to receive reliable data and to be able to draw conclusions from the population. The main reason that we used a sample and not the whole population was mainly because it would require too much time to conduct 185 telephone questionnaires due to our available resources.

Execute the sampling process

To be able to execute the sampling process the previous parts of the sampling design process model must be specified such as population, sampling frame, sampling technique and sample size. It is also important at this stage to make clear how the researcher will act and proceed if the respondents refuse to respond (Malhotra, 2004). The respondents of this study will be selected by two main criteria’s to qualify to answer the telephone questionnaire. The respondents should both have been a part of the campaign Vi saknar dig and in the age groups between 25 years to 50. This specific age-group was investigated since it is within Destination Jönköping’s primary and secondary target groups. If the respondents were unable to reach we have made three additional trials to reach them at different times of the day. If it still was unable to make contact with the respondents they have been excluded from the research.

Information sources

Information about Destination Jönköping, Welcome2work and the campaign within the project was essential in order to perform this study. It was of importance to have the knowledge concerning the organisation of Destination Jönköping, their operation areas, the background of the project Welcome2work and the purpose of the campaign. The information was primarily received through two major sources, public published corporate information and internal information by the marketing coordinator at Destination Jönköping Maria Lindén. The corporate information consisted of official documents as for instance brochures, operating statement and annual reports. The marketing coordinator at Destination Jönköping contributed information and knowledge about their daily operations, projects and strategies through several interviews. Different kinds of literature such as articles, books and journals were the major source of information about city branding and additional. The assumption was that it was necessary to use literature to get a broad perspective of information within the specific field of the thesis. Without literature it had probably been hard to obtain enough and accurate information for the study to understand and get a good insight of previous research and to be able to write the thesis in a professional way.
To make the information search of what previously had been published within the field of this thesis more effective we formed a search strategy to try to provide up to date information, limit the information search and to avoid information overload. In table 3.1 we categorise and clarify the parameters of the search strategy of this thesis. Since we had the aim of making a broad information search we used several different databases and search engines to increase the chances of finding the most appropriate publications. Even if we through this thesis focus and use the definition city branding we have still used destination branding and place branding as search words due to the fact that all three concepts are adjacent and close related and may therefore be relevant.
The publications that are used are mainly published in English, with some exceptions for Swedish publications. The reason for using mainly publications in English depends on the fact that the available range of publications is much greater compared to Swedish publications. Another motive for this was to decrease the risk of possible mistranslation of terminology by using other languages. Although city branding still is a relative new phenomenon we restricted our main information search to publication between years 2000-2012 to as much as possible base the thesis on up to date information due to our available recourses. A summary of the search parameters and strategy could be seen in table 3.1.

Methods of data collection

When collecting data, it is divided into two different types of data, primary and secondary data. Primary data is: “Data collected specifically for the research project being undertaken.” (Saunders et al., 2009, p.598) This could for instance be done through for observations, questionnaires and interviews. Secondary data is data has been collected earlier by someone else for another purpose than fore this thesis. Some examples of what secondary data could be are articles, books and websites (Saunders et al., 2009).
To be able to perform a reliable and interesting thesis both primary and secondary data will be used. The primary data have been collected through a telephone questionnaire which is a technique where the respondents of the survey have answered several questions and where all persons are answering the same questions (Saunders et al., 2009). Due to the choice of making a quantitative research the possible answers have to be clear and distinct. Because of that, more detailed answers that are more complicated to interpret were avoided. Detailed answers would give a more qualitative approach to the survey which not was the intention. When making a quantitative questionnaire it must be decided what kind of answers each question would have. It could be simple yes or no or it could be a question where the respondent has to rank some answers with numbers. There are several different kinds of example and therefore a thoroughly planning has to be done to create a questionnaire that will give good answers that could make the research reliable (Jacobsen, 2002).
When the selected respondents are answering the questionnaire it could be made in different ways which are divided into two groups, self-administered and interviewer-administered questionnaires. Self-administered questionnaires are sent out to the respondents who will fill in the answers themselves. This group includes three different alternatives to send out the questionnaires, through internet, by mail and delivered by hand to collect later. Interviewer-administered questionnaires mean that a person interviews the respondents either by phone or face to face (Saunders et al., 2009). One advantage with self-administered questionnaires is that it require less time than using interviewer-administered questionnaires since it is possible to send out several questionnaires at the same time, but making interviews has generally a better response-rate (Williamson, 2002).
The assumption was that two different alternatives of methods to collect the data were most appropriate for the thesis. Since the research aims to reach a special target group requested by Destination Jönköping, the sample of the respondents was to some extent preselected which made their e-mails and phone numbers available from the beginning. With this in mind a questionnaire sent by e-mail or conducted by telephone was assumed to be suitable methods for this purpose. After comparing these two methods against each other, making a telephone questionnaire seemed to be the best method for this research. The key factors in this decision were the fact that an interview by phone would probably have a better response rate and that it would require less time to complete the survey by phone than sending it by e-mail and maybe wait quite long time to receive answers. The advantages with sending the questionnaire by e-mail are that it is cheaper and that the respondents could answer the questionnaire when he or she prefers to do it. Overall the assumption was that making the questionnaire through an interview by phone would be the best choice in this case due to the available resources (Zikmund & Babin, 2007).

Rejected methods

As indicated and discussed in the previous section the chosen and most appropriate data collection method was assumed to be telephone questionnaire. Even though it is important to clarify and explain why other methods have been rejected in this paper.
Qualitative methods such as depth interviews and focus groups have been rejected, due to the fact that the sample of 103 persons was in some extent preselected by Destination Jönköping live all over Sweden in different cities. This makes interviews and focus groups on face to face basis difficult to perform due to the distance, cost and length of time it would require given the limited resources of the thesis (Saunders et al., 2009). Focus groups also need a moderator or facilitator to be able to run the session, this may be difficult to accomplish without leading the focus group into certain opinions as well as keep an objective approach. Another possible disadvantage could be that some members may dominate the focus group session which could limit other respondents’ opinions. It exist different types of questionnaire techniques such as internet-mediate questionnaire, postal questionnaire and telephone questionnaire. The major reason for the choice of telephone questionnaire is based on the fact that this is the method that is assumed to result in the possible highest response rate (Zikmund & Babin, 2007).

Structure of questionnaire

Within our telephone questionnaire we decided to have structured questions with a set of different answer alternatives. There are different kinds of structured question and we have used all three of them. We used multiple choice questions which provide three or more different answers where the respondents have to fill in one or more of the given alternatives. When having a multiple choice question it is important to include all possible alternatives that may be used or include an alternative that provides the ability to answer other which could cover the more unusual possible answers. The dichotomous questions were also used in our questionnaire and it has only two alternatives such as yes or no (Malhotra, 2004; Oakshott, 2012). Often a neutral alternative such as no opinion could also be available. If this neutral alternative not is included in the questionnaire the respondents has to choose between the alternatives. This could be negative if the respondents really do not have an opinion but also positive since otherwise the respondents may take the easy way and avoid answering what they really think. The last kind of structured question is scales where the respondents have to fill in the scale which alternative that is most suitable for the respondent. When making a question with a scale it has to be decided whether an odd or even number of categories will be included. If using an odd number of categories the respondents have the possibility to take a neutral approach of the question. With an even number of categories the respondents are forced to choose where they stand (Malhotra, 2004). When using questions with a decided set of answer alternatives it will take less time to make the interviews and it is easier for the respondents to answer since they do not have to come up with own alternatives. At the same time using open questions, where the respondents are free to answer whatever they feel, could give a more exact and detailed answer since the respondents are able to choose their own word to answer with. When having open-ended questions the job that has to be done are increasing due to extensive coding and analysing the amount of data that often is quite larger than within questions with fixed alternatives (Zikmund & Babin, 2007).
The telephone questionnaire consisted of 13 questions with different approaches even if all of the questions are within the category of structured questions with several answer alternatives instead of open-ended questions. The complete questionnaire that we have used to our respondents is presented in appendix 2 and the English translation in appendix
3. The two first questions concerned what gender the respondent had and which age group they belonged. These are general questions that we need to get answers from to receive a better and more detailed overview of the respondents to make it easier to analyse. The following two questions was multiple choice questions where several different answers of the question are available to choose. In the first one of these we wanted to find out what education the respondent had. This information was important since Destination Jönköping’s aim with the campaign was to attract people with an academic background. It means that this question could help us sort out which of the respondents that have an academic background and which have not. In the fourth question we thought it was interesting to see what kind of earlier relation the respondent had to Jönköping. We thought it could be a great difference between different answers depending on if they have grown up in Jönköping and lived most of their life in the city or for example if they have only studied there for about three years. Then we asked a dichotomous question if they lived in Jönköping or not. The answer of this question determined which way of this survey the respondents will take. If they answered that they had moved to Jönköping the respondent continued with questions regarding the relocation and if they answered no they continued to question ten which gave a hint if relocation to Jönköping within a period of five years was likely or not. We also included an answer alternative for those who not had relocated to Jönköping but worked in the city here. Since this question was easy to answer if the respondents lived here or not we chosen not to include a neutral alternative to this question.
Question six to nine was only be answered if the respondents had answered yes at question five because it covered the relocation. Question six and seven was ordered by Destination Jönköping because they wanted to know who the respondents moved together with and what the reason behind the relocation was. A relocation together with a family or at least a partner would give Jönköping more than if they just moved by themselves. The seventh question determined the reason behind the choice to relocate. Were they moving because of a new job, a relationship or did they want to move to the city where family and friends lives? These two questions were also multiple choice questions and did not concern any special theories from our framework but Destination Jönköping thought they would be interesting to include in the survey. Question eight investigated how great influence Destination Jönköping had within the choice of moving. We thought this question was important and interesting since we needed to know if Destination Jönköping had any impact of the decision to relocate to provide an indication if the city branding works. To get a useful answer of this question we thought that it would be most suitable to use a scale where the respondents filled in how great impact they thought Destination Jönköping have had in their decision. Furthermore we asked if the respondent had got a job in Jönköping and if the respondent in that case were hired by a company within Welcome2work. The ninth question had a great importance for Destination Jönköping when they tried to involve new companies in the project. If a lot of respondents answered they had been hired through Welcome2work, more companies would probably be interested to join the project. Question nine was a kind of a dichotomous question since it was a yes or no question but if the respondents’ answered no they had several different alternatives of no related alternatives to choose between. Those no answers that we provided were that they worked for another employer, that they searched for a job or that they were studying.
The respondents who answered that they not had relocated to Jönköping were the only respondents that answered question ten which concerned how great the probability was that he or she would relocate to Jönköping within a period of five years. This question where measured by a scale and gave us an indication of how interested the respondent were to relocate to Jönköping. To investigate in which kind of media the respondents had noticed the marketing communication that Destination Jönköping had used, we asked question eleven where the respondent could provide several answers. This question gave us a hint if the IMC had worked out well and in which kinds of media Destination Jönköping had most effective reached out to the target group. It indicated if some of the different media had a greater risk to lose attention due to the existing noise. Question twelve concerned if the respondents had been affected of the way Destination Jönköping marketed themselves through several different kinds of media. This question was measured with a scale and was also asked to see if the work with IMC had succeeded. The last question of the telephone questionnaire measured if the respondent’s image of Jönköping had been changed due to the campaign Vi saknar dig and in which direction, positive or negative. Our assumption was that this last question involved both the communication process and city branding. Communication process because we wanted to investigate the response and the feedback that the respondents could give and city branding were involved since we were given a overview of how successful Destination Jönköping’s work with city branding had been.
In all questions where a scale was provided to answer with we decided to use an odd number of alternatives. This means we gave the respondents an opportunity to take a neutral approach in all questions. The reason to have an odd number of the scale was because we did not want to force the respondents to take a more positive or negative approach in relation to their actual view of the questions.

Table of Contents
1 Introduction
1.1 Background
1.2 Problem discussion
1.3 Purpose
1.4 Delimitation
1.5 Contribution
1.6 Perspective
1.7 Definitions
1.8 Structure of thesis
2 Theoretical framework
2.1 Choice of theory
2.2 Communication in general
2.3 Branding
2.4 Summary of theoretical framework
3 Method
3.1 Research approach
3.2 The sampling design process
3.3 Information sources
3.4 Methods of data collection
3.5 Analysis of data
3.6 Reliability, validity and generalisability
4 Empirical findings
4.1 Structure
5 Analysis
5.1 Communication
5.2 Branding
5.3 Comparison of variables
5.4 Summary of analysis
6 Conclusions
7 Discussions
7.1 Critique of the study
7.2 Suggestions for further research
Reference list
GET THE COMPLETE PROJECT