Survey research is the collection of a small amount of data in standardized form from a large number of individuals (Robson, 2002). A survey research is different than experimental research because it measures many variables, test multiple hypotheses, and infer temporal order from questions about past behavior, experiences, or characteristics.

Robson (2002) states that surveys are used in descriptive or explanatory research. In this context, Neuman (2003) indicates that surveys with open-ended questions can be used for exploratory purposes. For instance, a researcher can use open-ended questions in pilot tests and then develop closed question responses.

 

When conducting a survey research, one has to consider both internal validity and external validity. If questions are ambiguous, the exercise will be waste of time. Therefore, a researcher uses pilot tests to ensure that the intended meaning in the questions are clear. The selection of representative samples from known population is critical in survey research because if sampling is wrong, a researcher will not be able to generalize the findings. When generalizing the findings to the population from which the sample is drawn, a researcher should also consider the homogeneity of the population.

 

Among the surveys that I have filled out for my friends, I faced the following problems defined by Neuman:

• Noattitude,
• Lack of partially open questions, and
• Length of Survey.

 

In some surveys, I had no opinion about the issue and I felt that the researcher was forcing me to make a selection. Therefore, I selected the neutral option. I am not sure whether my answers for some surveys were helpful or not. I beleive that our friends who conduct surveys should be careful about this issue because it will impact the generalizability of their findings. 

 

Another problem that I noticed (especially in the survey we filled out in our class) is the lack of partially open questions. Recently I got a high cholesterol problem and in some of close-ended question I felt that the researcher could not provide an answer that captured my true feelings.

 

Finally, the length of the surveys that our friends prepare is the scariest part for me. In my previous school, I conducted an online survey to determine students’ feelings toward online classes. In this study, my department gave me about 20 questions and I put only 5 questions to a page. The online survey had 5 pages including the “thank you” page. With this format, I got a very high return rate on the Christmas break from the students.

After reading this chapter and the article.I think I am little confused about the difference between pilot test and pretest. On one hand, pretest is taken before services or activities begin. It is compared with the results of a posttest to show evidence of the effects of the services or activities being evaluated. On the other hand, pilot study is a mini version of a study conducted before committing to the big one. In the article, it seems to me that they did a pilot study to ensure that there is a significant difference between uncertainty levels. They used on 43 students not all students. In order to consider that a pretest, I think all participants need to take the test. Furthermore, all participants need to take the posttest too. However, the authors used only about 30 percent of the sample for what they referred as pretest. In this context, I believe that they used the term wrong.

 This chapter focuses on the fundamentals of how sampling works, the differences between good and bad samples, how to draw a sample, and basic principles of sampling in social research. According to Neuman, sampling (collecting specific cases, events, or actions that deepen understanding) is an important issue when conducting a research because it influences research design, measurement of variables, and data collection strategies.  In this context, qualitative and quantitative researchers approach sampling differently.

 

On one hand, quantitative researchers prefer probability sampling (each person has equal chance of being included in the sample) because it produces more accurate sampling by eliminating bias. In probability sampling, the central limit theorem (with a huge number of random samples, the sampling distribution forms a normal curve) is important because it allows a researcher to generalize from one sample to the population. In the “An Empirical Study of the Preferences of Student Internet Users in Malaysia for Online Products and Services” article, the authors used this theorem to generalize the findings. Furthermore, this study also resembles the characteristics of stratified sampling because the authors first divided population into subpopulations (different institutions) and then the authors drew a sample from each institution. However, the authors did not indicate whether they selected those schools and the students in those schools randomly or not. Dividing a population into subpopulations and then randomly sampling from each population resembles parametric design under true experimental design that we discussed in the IS:360 class. I assume that researchers use this sampling when they need further confidence.

 

On the other hand, qualitative researchers tend to use the following nonrandom samples.

 

Haphazard Sampling: The most basic nonprobability sampling and it raises questions of generalizability because the people walking past a television studio in the middle of the day do not represent everyone.

 

Quata sampling: Ensures that there are some differences in the sample because researcher first identifies relevant categories of people. A purpose of the

Judgmental Sampling: A purpose of this sampling is

 

“A researcher wants to identify particular types of cases for in-depth investigation. The purpose is less generalize to a larger population than it is to gain a deeper understanding of types” (p. 213).

 

In this context, I am curious about the knowledge management systems continuance article because this paper had a generalization problem and I wonder if it would be better for the authors to follow this approach and gain a deeper understanding.

 

Snowball Sampling: The sociogram may also provide further understanding on how social relationships affect attitude toward knowledge management systems continuance.

 

Deviant Case Sampling: Aims to learn more about social life by focusing on peculiar cases that are not representative of the whole.

 

Sequential Sampling: Sounds similar to Robson’s idea of “keep going until data collection adds little or nothing you have already learned”.

 

Theoretical Sampling: Goes with grounded theory. From my point of view, grounded theory is the most difficult research strategy.  

The link below describes a process that the American Society of Cytopathology suggests when writing and submitting a research paper for publication.Although the focus is on medical research papers, the steps defined in the process can be used in any other fied. There are four main sections in this process and each section is divided into sub-sections. Hence, one can evaluate a research paper based on how the authors followed the process.

http://www.cytopathology.org/website/download.asp?id=55   

According to Reynolds (1971), the purpose of science is to provide typologies, explanations, predictions, and a sense of understanding. Furthermore, Reynolds (1971) indicates that abstractness, intersubjectivity, and empirical relevance are the desirable characteristics of scientific knowledge. In this context, Neuman focuses on empirical relevance and states that the goal of the researcher is to establish unambiguous links between abstract ideas and empirical data.  

 

Both quantitative and qualitative (fixed or flexible) designs use conceptualization and operationalization in measurement. Since there is substantial amount of pre-specification in quantitative design, the measurement process is a straightforward process sequence:

 

Conceptualization

Operationalization

Applying the operational definition or measuring to collect data

 

The social relationships in knowledge management systems continuance in organizations: A case study in China article followed the process above. In this article, conceptualization was the refinement of the construct (knowledge system continuance) by social capital and change management theories. Furthermore, operationalization was the link between the conceptual definition of knowledge system continuance and social relationships (trust, shared norms, tie strength).

 

However, in qualitative research, conceptualization and operationalization are different. On one hand, a researcher can not refine abstract ideas into theoretical definitions (conceptualization) early in the research process. Hence conceptualization is determined by the data. On the other hand, operationalization describes how a researcher developed working ideas while making observations and collecting data.

 

Reliability and validity are important issues to consider when evaluating the trustworthiness of a study. According to Neuman (2003), reliability means dependability or consistency (p.178). The articles that we have read so far used the survey method to collect data. In a survey, if questions are ambiguous, respondents will have difficulty to understand the questions and the whole exercise may end up a waste of time. In this context, pretests and pilot studies can be used to ensure that survey questions mean the same thing to all respondents. However, the two articles we read did not mention a pre-test or a pilot study. Therefore, they are susceptible to reliability.

 

Finally, I am a little confused about semantic scale. Based on my understanding, semantic differential scales are used to asses the subjective meaning of a concept to a respondent. In the knowledge management systems continuance in organizations: A case study in China article, the concepts social relationships (trust) and perceived usefulness seem subjective to me. Furthermore, semantic scale is also appropriate for marketing research. However, an empirical study of the preferences of student internet users in Malaysia for online products and services article preferred likert scale to semantic differential scale.        

It is my good fortune that I had an opportunity to meet Prof. Ikujiro Nonaka and Prof. Hirotaka Takeuchi, who are pioneers in the knowledge management field. I took knowledge management class last semester and I saw a ton of papers that cited their names. They came to Drucker School on February 12, 2007 to give a presentation on topic “Strategy as Distributed Phronesis.” They also mentioned that Prof. Drucker is an important pioneer in the knowledge-based society.

 

It was very impressing to attend their lecture to hear them discuss their thoughts and experiences. Prof. Nonaka talked about leadership in a knowledge-creating company and explained that “it is a dynamic process of synthesizing the vision, Ba, dialogue, practice, knowledge assets, and ecosystem of knowledge to create knowledge.” The basis of leadership in his perspective is phronesis.

 

Prof. Nonaka doesn’t like the term “knowledge management” but rather “knowledge-based management.” He explained that we can’t manage knowledge directly. We need to understand this concept before managing knowledge.

 

Prof. Nonaka and Prof. Takeuchi were humorous. I very much enjoyed their lecture. Prof. Nonaka talked about Aristotle’s three types of knowledge: Episteme (scientific knowledge), Techne (skills and crafts knowledge), and Phronesis (prudence, practical wisdom, practical reasoning). He focused on six abilities that constitute phronesis which now he and Prof. Takeuchi are working on.  It’s very interesting to me that they said knowledge is also related to philosophy because this semester I’m taking happiness class which talks about the Nicomachean Ethics of Aristotle. Prof. Nonaka said “you can’t apply a universal rule when you’re dealing with human beings.” and explained that “knowledge is created and refined to be wisdom.” This semester, I’m also taking IS362 that focuses on social research methods. Thus, his ideas are very attractive to me. In the qualitative method, wisdom of the researcher is very important. His lecture helped me gain insight from this perspective.

Based on Neuman, the unit of analysis refers to “the type of unit a researcher uses in measurement" (p.156). For instance, individuals, groups, organizations, society…In this context, (similar to triangulation) considering both holistic and individualistic views provides a way to complement the findings of a study by broadening the results. However, I have two concerns here. The first concern is the ambiguity about what counts for individual or unit. Based on Hollis, some economics treat firms as individuals and others regard them as organizations. The second concern is how to conduct a research while analyzing both units. For instance, should units be analyzed at the same time, in a sequential or random order?

    Based on the Social Relationships in Knowledge Management Systems Continuance in Organizations: A Case Study in China article, the figure below analyzes the reason why sales representatives use the knowledge portal to exchange best practices from systems, games, agents, and actors perspectives.

In this chapter, Neuman uses two new terms (technocratic and transcendent perspectives) to distinguish quantitative and qualitative styles of research (fixed and flexible design). On one hand, in technocratic perspective (fits with positivism), research questions often generate with the sponsors of the research. On the other hand, in transcendent perspective research questions originate with standpoint of the people being studied, not that of outsiders. From my point of view, this definition is more proper for clarifying the difference between real world and laboratory studies and it makes me think that quantitative research is more client-oriented than qualitative research. For instance, a company who wants to improve the customer satisfaction on a certain product would likely to see the product from customers’ perspective. In this context, the company may interview its top customers in an unstructured way to develop research questions based on their standpoint. Therefore, I believe that the sponsor of the study does not have to be the source of the research questions.

In our last class, Kate, Rosemary, and I pointed out the difficulties of interpretive and critical social science over positivism. In this context, Neuman states that “qualitative research relies on the informal wisdom that has developed from the experiences of researchers” (p. 140). Therefore, inexperienced researchers who can record information without interpreting it can not conduct a credible interpretive research. Neuman’s sentence also supports researcher as an instrument approach followed by Robson. However, researcher as an instrument emphasizes the potential for bias and this is where triangulation is used to deal with data overload, first impression, inconsistency, uneven reliability, and missing information problems.

The major difference between quantitative and qualitative styles of research (fixed and flexible design) is the pre-specification. In quantitative research, the researcher begins the study with a conceptual framework or theory that provides substantial amount of pre-specification. The pre-specification also resolves the threats to validity. However, qualitative research has much less pre-specification and the design evolves in time. In this context, Neuman indicates that “qualitative researchers remain open to the unexpected, are willing to change the direction or focus of a research project, and may abandon their research question in the middle of a project” (p.146). Although I believe that a researcher should be adaptive and flexible, I do not think that there is a free range exploring is possible due to time and cost constraints. From my point of view, a research should look for something specific while being open to unexpected discoveries in flexible design.  

As reviewers, our responsibility is to indicate whether we would like to publish a certain paper in our journal. Based on Jianmin, we can use the following criteria when making a decision:

·        Importance of the topic being studied: Is it a hot, new, or controversial topic?

·        Significance of the contribution:

1.      Significant substantive findings (new insights & knowledge)

2.      Meaningful theoretical developments and generalizations

3.      Important methodological advances

4.      Tests of existing theories (surveys and experiments)

5.      Comparison of methods (which one is the best)

6.      Empirical investigations (Why some firms perform better)

7.      Cross functional research (interdisciplinary approaches)

·        Readability of the paper (English, logic, presentation, and positioning)

 

In this context, Jianmin relates high rejection rates of top journals to the following reasons:

·        Editors are extremely risk averse: Editors are more concerned with the risk of accepting low-quality papers than rejecting good articles.

·        Acceptance rate is less than 1/9

1.      Associate editor and at least two referees.

2.      Research interest conflict.

3.      A well-known referee weigh more heavily

4.      Negative comments influence more

·        Other factors: New author without experience

 

Based on MIS Quarterly’s review process from 1995 to 2001, the journal's acceptance statistics are summarized below.

 Average number of papers on which decisions were made each year: 236 

Average number of papers on which final disposition decisions were made each year: 122 

Average acceptance rate based on final disposition decision: 13.6% 

Average number of revisions required before acceptance: 2.66 

Average percentage of rejected papers rejected after first review: 82% 

Average percentage of rejected papers rejected after second review: 13% 

Average percentage of rejected papers rejected after third review: 4% 

Average review turnaround time for first review round: 67 days SD = 67 days 

Average review turnaround time for second review round: 89 days SD = 63 days 

Average review turnaround time for third review round: 74 days SD = 63 day 

 

References 

http://www.misq.org/archivist/vol/no26/Issue2/EdComments.pdf

 

http://som.xjtu.edu.cn/shuyanban/xiazai/Issues%20in%20Publishi Criteria  

Different approaches represent different ways of looking at the world when conducting a research. Each approach (positivism, interpretative social science, and critical social science) has its own strengths and weaknesses. However, they have one thing in common and that is scientific attitude. In order to consider a research “scientific”, it must be carried out in systematic, skeptical, and ethical way. Based on positivist approach, reality can only be known imperfectly and probabilistically because of the researcher’s limitations. Interpretative research resembles relativist approach defined by Robson because this approach indicates that everything is relative and nothing is absolute. For instance, “the fall of a tree in a forest where there is no one to hear it. Does it make sound?” Finally, critical social science seems like the combination of realist approach and action research. Two aspects of chapter 4 were confusing to me. The first one is the term “social positivist science”. Based on Robson (2002),” positivism is not proper for social science because it has not produced any scientific laws yet” (p. 21). However, Neuman (2003) indicates that “positivism is associated with many specific social theories. Based on my understanding from the IS:360 class, positivism is proper for natural sciences where we fist generate pure research findings and then apply them.

In chapter 5, Neuman talks about the literature review which I believe the authors of the article that we are reviewing would benefit. For instance, in knowledge management, there are lots of definitions and some of them contradict such as traditional data, information, knowledge, wisdom (DIKW) model versus Firestone approach. In the article, the authors define the term knowledge under the heading definition of concepts (page 3), but they do not state whose definition is that. However, in organizational factors sub-section (page 17) the authors get back to that definition and state whose definition they used. Nevertheless, the author did not state why they chose that definition. For instance, why Davenport instead of Senge or Firestone. Based on the classification of literature review, I would say that the authors of the article made a theoretical review because they presented different theories (TAM, Expectation Confirmation, Social Capital, Change Management) to explain knowledge sharing and then they evaluated how well Social Capital and Change Management accounted for the findings.