The approach of this study is deductive. This means that the research project goes from theory to empirics (Jacobsen, 2002). The researcher reviews data that has already been col-lected, but for another purpose, to get his/her own expectations about the reviewed sub-ject and then collects data from the reality to check whether the expectations and reality are consistent with each other (Saunders, Lewis, & Thornhill, 2009; Jacobsen, 2002). From the Swedish auditing standards and previous studies around the subject of implementing the international standards in the EU for example, we got expectations. Thereafter, we checked whether the expectations from the reviewed data were consistent with the data collected in our study.
The purpose of this study is of an explanatory nature. An explanatory purpose is used when the researcher wishes to explain the relationship between variables in a certain situa-tion or problem (Saunders et al., 2009). Our purpose is to study the situation when ISA was implemented in Sweden in order to explain the relationship between the variable “changed auditing standards” and the variables: “impact on the audit process”, “audit quality” and “costs”.
The research strategy is the general plan for how the research question shall be answered. This means that the research question will guide which strategy to use (Yin, 2009). This study aims at answering the questions of how the implementation of the clarified ISAs has affected the audit process for Swedish auditors, the effects from this and also to compare the affect on the audit process at small versus large audit firms. Some different research strategies are experiments, survey and case study (Saunders et al., 2009). For this research, we found that the case study was the most appropriate strategy. A case study is used to in-vestigate a contemporary phenomenon in its real life context. Using case study as research strategy means that many sources of evidence are used to investigate a phenomenon. This strategy is suitable if the researcher wants to get a deep understanding of the context, it is particularly useful in answering the question “why” but also “what” and “how” to some degree. This makes case study appropriate when the purpose is of an explanatory and/or exploratory nature (Saunders et al., 2009).
To be able to differentiate both data collection techniques and data analysis procedures, the terms quantitative and qualitative are used (Saunders et al., 2009). A quantitative study pro-vides data in numeric (numbers) forms while a qualitative study instead focuses on non-numeric (words) data (Saunders et al., 2009; Jacobsen, 2002). Since we are interested in get-ting data in the form of words rather than in numbers, a qualitative study is most appropri-ate.
It is common to combine a variety of different techniques when collecting data in a case study. Examples of different data collection techniques are interviews, observations and questionnaires. When using the case study as research strategy, it is common to triangulate the different sources of data used. Triangulation is a technique of making sure that the data collected is true and interpreted correctly by using multiple data collection techniques (Saunders et al., 2009). Since the data collected in this study is very subjective, we want to have as many opinions as possible to be able to draw reasonable conclusions. Therefore, questionnaires were constructed and sent out to auditors across the country. However, be-fore sending out the questionnaires, interviews were conducted in order to make sure that the questions asked in the questionnaire were relevant and to get a deeper insight about the subject.
There are four types of method approaches; mono method, multi-method, mixed method and mixed-model (Saunders et al., 2009). The one used in this study is the multi-method. The multi-method means that the researcher has either a qualitative or quantitative study, but uses several data collection techniques and/or analytical procedures (Saunders et al., 2009). We have used two different qualitative data collection techniques; interviews and questionnaires and two qualitative analytical procedures; categorization and testing relation-ships.
According to Saunders et al. (2009), researchers can conduct either cross-sectional studies or longitudinal studies. Performing cross-sectional studies means that the researcher exam-ines a situation or problem at several points in time and longitudinal studies means that the researcher only examines a situation at one point in time (Saunders et al., 2005). There is a special version of a cross-sectional study that will be most efficient for the purpose of this study. This version of a cross-sectional study means that the researcher asks the respond-ents to remember the past to be able to compare a situation at one point in time to another without having to ask at several points in time (Jacobsen, 2002). This method will be used in this study since the change in auditing standards already has occurred. In order to be able to compare the situation before and after this change, we will have to ask the respond-ents to remember. This method will be used both in the interviews and in the question-naires. Jacobsen (2002) addresses a problem with this method. He states that it can be diffi-cult to get reliable answers from the respondents if they do not remember clearly. This problem will not be significant in this study. There are two reasons for this. The first is that the changes in the auditing standards occurred recently. The second reason is that the au-diting standards is something really important for auditors to know by heart and is some-thing that they work with every day. We will therefore assume that the auditors will re-member the time before the auditing standards were changed in 2011 since the questions will be all about the important auditing standards.
When conducting a research, both primary and secondary data can be used. These two ways of collecting data can often complement each other so the ideal thing is to use both (Jacobsen, 2002). In this study, both primary and secondary data have been used to be able to answer the research questions.
Primary data is collected by the researcher for the specific purpose of the study. This type of data is collected directly from the source (Jacobsen, 2002). In this study, primary data have been collected through semi-structured interviews and questionnaires. The reason for why both interviews and a questionnaire were used was to be able to fulfill the purpose of this study. The purpose with the questionnaire was to get as many opinions about the im-plementation about ISA as possible, and also to get opinions from auditors across the whole country. The purpose with the interviews was to test whether the questions in the questionnaire were relevant and also to get a deep insight about the subject. Because of this, the questionnaire was constructed as a start and thereafter the interviews were con-ducted before the questionnaire was sent out in order to be able to make adjustments in the formulations and adding new variables and alternative answers. Due to the importance of how to construct interviews and questionnaires, these are explained later in this section.
Secondary data is already existing data within the specific subject but that may have been collected for another purpose (Saunders et al., 2009). This type of data is not collected by the researcher directly from the source and can be both raw data and published summaries (Jacobsen, 2002; Saunders et al., 2009). For this study, we have used secondary data to gain a deeper understanding about the subject as well as finding relevant theories. Summaries and studies about different EC projects have been collected from the EC website. The main sources of secondary data are academic journals and other relevant literature. Due to lack of information about Swedish conditions in academic journals, different publications from FAR have been used for this matter. Legislative publications have been another im-portant source in gaining an understanding about the audit process.
When conducting a research, it must be considered whether or not sampling is needed. It is often impossible to collect data from the entire population due to time and money re-strictions. Therefore, a sample is often necessary to draw (Saunders et al., 2009). Collecting data from all auditors in Sweden would be impossible in this study due to these restrictions and that is why sampling is used.
The sample can be either a probability sample or a non-probability sample. The type of sampling technique used depends on whether the researcher wants to make statistical gen-eralizations or not. The probability sampling technique is used when the research question needs to be answered on statistical grounds and when the researcher wants to make gener-alizations. Using a non-probability sampling technique makes it possible to make generali-zations but not on statistical grounds (Saunders et al., 2009). In this study, non-probability sampling techniques are used since the goal is not to make statistical generalizations. This means that the probability of each case being selected is not known (Saunders et al., 2009).
To be able to draw a sample, a sample frame is needed. A sample frame is a complete list of all cases in a population where the sample is going to be drawn from (Saunders et al., 2009). In this study, the sample frame was collected from a list of all qualified audit firms in Sweden found in an electronic phonebook called Eniro. This sample frame was used since the differentiations of auditors from small respectively large audit firms were needed to be able to answer the research questions. When the audit firms were drawn from the sample frame, qualified auditors at that firm were asked to be our respondents.
The only characteristic required of the respondents in this study was high experience of auditing. Therefore, only qualified auditors were asked to be our respondents for the inter-views and questionnaires. This ensured that the respondents had been in the industry for at least three years.
The non-probability sampling technique convenience sampling has been used when draw-ing the sample for the interviews. This means that the cases chosen are the ones who are the easiest to obtain (Saunders et al., 2009). We have chosen to conduct interviews with auditors located in Jönköping since it would not be convenient to travel across the country to conduct interviews. With the convenience sampling technique, six different auditors in Jönköping have been chosen for the interviews; three working at the “big four” audit firms and three working at smaller audit firms. For the questionnaire, the non-probability sam-pling technique self-selection sampling has been used to select cases. According to Saun-ders et al. (2009) self-selection sampling means that the researcher publicizes the need for cases or asking them to take part and then collect data from those who respond. This has been done in this study where we have sent out the questionnaires to qualified auditors across Sweden and collected the data from those who chose to take part in our survey.
When using these two non-probability sampling techniques there are no rules regarding the sample size (Saunders et al., 2009). Patton (as cited in Saunders et al., 2009) states that the sample size should be dependent on your research questions and objectives, for example what the researcher needs to find out. In addition, Saunders et al. (2009) give the simple recommendation to collect qualitative data until saturation is reached. As mentioned above, the purpose with the interviews was to get a deeper understanding about the implementa-tion of ISA in Sweden and to test the relevance of the questions in the questionnaire. Hav-ing six interviews gave the information needed and to conduct additional interviews would not have provided any new insights. The questionnaire has been sent out to 300 qualified auditors spread across Sweden; 150 working at the “big four” audit firms and 150 working at smaller audit firms. When using a questionnaire, the researcher has to consider the re-sponse rate. This is likely to depend on how interested the respondents are in the subject. If the respondents are interested in the subject, i.e. they are considered an “elite”, and if they realize the importance of responding, they are more likely to respond. Auditors are considered to be such an “elite” (Jacobsen, 2002). Therefore, the problem of lack of inter-est in the subject does not exist. However, before sending out the questionnaire we were aware that the auditors have a high workload during the spring. Because of this, the questionnaire was sent out to many auditors in order to get as many opinions about the imple-mentation of ISA as possible. The response rate for the questionnaire was 18,3 % (55 out of 300). We got the problem of the high workload confirmed when many auditors an-swered that they did not have the time to help us fill in the questionnaire. This is one rea-son for why the response rate is relatively low, but this number of respondents and opin-ions about the implementation of ISA will be enough for us to be able to draw conclusions.
There are different types of interviews; structured, semi-structured and unstructured inter-views. The different types of interviews goes from standardized interviews with predeter-mined questions, i.e. structured interviews, to informal interviews with no predetermined questions and with the respondent guiding the interview, i.e. unstructured interviews. In between these two extremes we find the semi-structured interview (Saunders et al., 2009. This is the type of interview we have chosen for this study.
Both semi-structured and unstructured interviews are often used in qualitative research. The semi-structured interview is suitable for both exploratory and explanatory purposes. This type of interview will be helpful when the researcher wants to know the reason behind the respondent’s decisions or attitudes but also when relationships between different varia-bles are interesting. Themes and questions will guide a semi-structured interview. The structure may, however, vary from interview to interview. Which questions that will be asked in each interview and in which order they will be asked will be determined as the conversation goes on (Saunders et al., 2009). When conducting the interviews for this study, guidelines were used for collecting the data (Appendix 1 and 2). These were only guidelines and the order of the questions was not followed precisely during the interview. However, every interview started with insensitive and general questions in order to get the information the respondents wanted to tell and not what we wanted to hear. However, we realized a pattern quickly about how the respondents felt about the implementation of ISA. At first, they said that they had not realized any differences from the implementation but when we asked about details and discussed the changes in the standards they said that there were some changes.
Before the interviews, we had reviewed the literature and were well prepared on the sub-ject. To enable the interviewees to be prepared and to be able to give us all relevant infor-mation, the guidelines for the interviews were sent out to the interviewees before the interview. In addition to the guidelines, a short explanation about our study, a guarantee for an-onymity and a question about their permission for us to record our interview was sent. The only purpose with the recording was to avoid noting everything they said and thereby being able to be active during the interview. We could listen carefully to what they said, get a good understanding and therefore also got the possibility to ask follow-up questions. After the interviews, the transcription of the audio recording was sent to the interviewees to ena-ble them to correct misunderstandings.
The length of the interviews varied where the first interviews were longer than the last. In the beginning, the interviews were almost an hour long, and in the end they were only around 30 minutes. The reason for this is that during the first interviews, there were a lot more discussion around the subject. As our understanding and knowledge increased, there were more opinions and experiences that were discussed rather than the subject of the study.
The reliability in a research refers to the extent to which another researcher would yield similar results using the same techniques for collecting or analyzing data. Two different types of biases that will affect the reliability of a research may occur when conducting in-terviews; interviewer bias and interviewee bias. Interviewer bias may occur due to the fact people ask questions in different ways, i.e. they express themselves differently. Interviewers may also, intentionally or unintentionally, direct the respondents towards an answer that the interviewers wish to receive. Finally, interviewer bias may occur because of peoples’ dif-ferent ways of interpreting the responses. Interviewee bias on the other hand may occur if the respondent does not feel comfortable to give the true and whole picture. The respond-ent may not have the power or the desire to reveal certain information or he/she may only answer what they think is expected. The risk of interviewer bias can be reduced by making sure that the interviews are well-structured and the risk of interviewee bias can be reduced by ensuring the respondents anonymity (Saunders et al., 2009).
1.2 Problem discussion
1.4 Thesis outline
2 Frame of reference
2.1 The need for auditing
2.2 The audit process
2.3 The change of auditing standards
2.4 Possible effects of changing standards
2.5 Agency theory
3.1 Data collection
3.2 Data analysis
4 Empirical findings
7.1 Critiques of the study
7.2 Suggestions of future studies
GET THE COMPLETE PROJECT
The implementation of the clarified International Standards on Auditing