• No results found

4. RESEARCH METHOD

4.5 D ATA COLLECTION QUALITY EVALUATION

Data collection quality can be evaluated through assessing its reliability, validity and generalizability. Reliability concerns the consistency of the measurement and refers to whether the data collection techniques will yield consistent findings, meaning similar

observations or conclusions reached by other researchers performing the same study (Adams et al., 2007, p. 236; Saunders, Lewis & Thornhill, 2007, p. 609). The reliability of a research project can be threatened through participant error or bias, or observer error or bias.

Participant error occurs if the response is collected at a time where the participant is not neutral to the subject in question. Participant bias occurs if the respondents´ answers are biased by a belief of how they should answer. Observer error occurs if the researcher

influences the participant through the procedures of the data collection. Observer bias occurs if the researchers´ interpretation of the data collected effect the conclusions being drawn.

(Saunders, Lewis & Thornhill, 2007, p. 150) All these situations result in the collected data not showing the respondents´ true opinions of the subject in matter. Validity refers to the extent to which the data collection methods accurately measure what they attend to measure, which influences the strength of conclusions, inference or propositions of the study (Adams et al., 2007, p. 237; Saunders, Lewis & Thornhill, 2007, p. 614). Generalizability refers to

the extent to which the research findings are applicable to other settings (Saunders, Lewis &

Thornhill, 2007, p. 598).

4.5.1 Reliability and validity

The reliability and validity of secondary data is a function of the source of the information and the method used when collecting the data. The source should have high authority and a good reputation of an objective source of data to secure its quality and suitability to the research. Moreover, as the secondary data is collected and often interpreted for other

purposes, it might be too old or not be objective sources of information. (Saunders, Lewis &

Thornhill, 2007, p. 257-262, 265-268) The secondary data used in this thesis is produced by sources with high authority such as government funded statistical databases, official coffee organizations and serious newspapers, that all have a good reputation as objective sources of data. This minimizes the risk of the data material not being objective sources of information.

Moreover, the collection and publish dates of the data was carefully evaluated to secure that it was not too old to be relevant. However, as some of the sources of secondary data were reports and articles, these data may inevitably be subject to the author´s opinions and therefore to some degree subjective even as a source of high authority and reputation.

For semi-structured interviews, the lack of standardization may affect the reliability and validity due to the interviewer affecting the respondent or the respondent adapting the answers to their perception of the interviewer (Saunders, Lewis & Thornhill, 2007, p. 318).

To reduce these threats, the questions asked were formulated in a neutral tone to obtain the respondents true opinions. Moreover, the respondents were carefully informed that their answers would be made anonymous. This also included interview quotes used to underlie arguments of the analyses of the thesis. Furthermore, as the subject of the interview was the overall market structure and competition of the coffee bar market in Oslo, the participants were not subjected to revealing sensitive information about their own company or

themselves. This further reduces the treats to reliability influencing the data collection quality.

To increase reliability of the survey, the respondents were recruited through different channels to gather a heterogeneous respondent group. However, the respondents accessed did not consist of an even group of gender, age or occupation as described in chapter 5.2.1.

Moreover, the sample consisting of self-selected respondents might lead to the participants

being especially passionate about coffee or coffee bars. This may bias the results so that they are less valuable as foundation for a generic coffee bar strategy in the Oslo market. The consumers who did not want to participate might therefore have been valuable to the study, as their responses might have yielded other results for the analyses. However, these might not be as relevant to the coffee bar strategy if their lack of coffee or coffee bar passion results in them rarely frequenting coffee bars.

Another important aspect of reliability lies in the definitions of the variables being measured (Adams et al., 2007, p. 235). To minimize the problems of such misunderstandings, the questionnaire design aimed to ask clear questions and included explanations of relevant concepts such as the definition of a coffee bar. The questionnaire was also tested and re-tested in the development phase to increase the reliability and validity of the data collected and conclusions drawn.

4.5.2 Generalizability

As the focus of this thesis is limited to the coffee bar market in Oslo, the generalizability of the findings may be limited to other geographical areas or markets. However, the aim of the research is not to produce results that are generalizable to all markets. Still, as the analyses are based on general economic mechanisms, some of the conclusions drawn might be valid to other geographical markets in Norway with a similar competitive environment.

The respondents of the survey were mainly students in the age between 20-30 years, and their response may therefore not be general for the larger population of coffee bar customers in Oslo as these belong to a number of age segments and occupations. This may therefore reduce the generalizability of the research to all populations of coffee bar customers in Oslo.