Kenneth J. Petersen

1
Lessons Learned in Publishing and the Review
Process

• Kenneth J. Petersen
• Associate Professor of Management
• Colorado State University

2
A Pile of Stones

• Science is facts, just as houses are made of
  stoneBut a pile of stones is not a house, and a
  collection of facts is not necessarily science
  (Poincaré (1983) in Whetton (1989).

3
Introduction

• Lessons learned from my publishing experiences
  (1994-2006)
• Discussion based on actual reviewers comments
  that I have received on my submissions (all
  comments relate to work that is now in press or
  in print)
• Suggestions for how to avoid some of the
  stumbling blocks that I have encountered

4
Program of Research
5
Published Work
6
Work-In-Progress
7
Writing Style

• (IJPMM) - Please eliminate the use of the first
  person in the paper, e.g. on page 10 We may
  therefore
• (JSCM) - Write the paper in the 3rd person.
  Delete references to we/our, etc.
• (JOM) - Although it is fairly well-written, there
  are a number of punctuation and grammatical
  errors that detract somewhat from the readability
  of the manuscript. Have the paper professionally
  edited.
• (JOM) - References in the body of the manuscript
  are inconsistently written.
• (JSCM) - Some of the manuscripts citations are
  missing from the Bibliography.
• (JBR) - Claims are made throughout the paragraph
  without supporting citations.
• (JOM) - What is obvious to one discipline may not
  be necessarily obvious to the other.

8
Writing Style

• (JOM) - Pages are not numbered.
• (JOM) - First a technical remark It is rather
  hopeless to give detailed remarks on a manuscript
  without pagination.
• (JPIM) - The authors seem to identify themselves
  (which is fine but I thought the reviewing
  process was supposed to be blind)

9
Writing Style

• (JSCM) Your tables are not consistent in
  format. Center and bold the headings using the
  same format throughout. Either keep or lose the
  shadowing effect. The main thing is to make them
  consistent.
• (JOM) - The authors spent much effort in bringing
  related literatures together, which makes the
  first three sections of the paper easy to read.
  Yet several typographical errors still exist.
• (JOM) - Consider carefully the value of using
  figures and tables in lieu of text, especially if
  text repeats information in figures and tables

10
Writing Style

• (JOM) - It would help if the text could be
  organized in more distinctive sections of
  traditional separation. Problem, literature
  mapping, aim, method, theoretical/model
  development, data, analysis, synthesis, and
  conclusion or something like that.

11
Writing Style - Recommendations

• Academic writing is a craft (a creative process)
• Self-plagiarism
• Have colleagues critique your work.
• Use a professional editor.
• Use Endnote.

12
Writing Style - Recommendations

• Your manuscript MUST tell a story.
• Your manuscript MUST have a strong and compelling
  organizational structure.
• Dont submit your first draft!
• If the reviewer (or reader) has to work very hard
  to understand your work, they will not read it.

13
Literature Review (Theory Development)

• (DSJ) The theory development section of this
  manuscript is disjointed.
• (JSCM) - I have concerns regarding the relatively
  weak theory, lack of hypotheses to support the
  papers empirical models, and failure to
  adequately examine reliability and validity,
  which preclude me from recommending publication
  of the paper in its current form.
• (JSCM) - The authors need to review and
  incorporate this literature into the development
  of their empirical model.

(What How)
14
Literature Review (Theory Development)

• (JSCM) - Perhaps most disconcertingly, the
  authors fail to develop formal hypotheses, or to
  even explicitly rationalize the structural
  relationships displayed in .
• (JOM) Hypotheses appear abruptly without
  justification.
• (JSCM) The literature review in the Prior
  Research section of the paper must include the
  theoretical development and introduction of the
  hypotheses which are implicitly represented by
  the models in these exhibits.
• (JPIM) - It would help if the general discussions
  before the hypotheses were organized with clearer
  logic. It should be possible to conclude by
  something like "because so and so as discussed
  above we hypothesize so and so". Many hypotheses
  arrive quite abrupt.

(What How)
15
Literature Review (Theory Development)

• (JSCM) - The bibliography is far too shallow and
  superficial. Given the quantity and quality of
  literature on supply integration, supply chain
  performance, and related areas, this section
  should be strengthened substantially.
• (JPIM) - It may be difficult for the authors to
  have seen or included at a late stage but
  articles in JPIM vol 20 no 4 and 5 should be
  included in a revision of this manuscript.

(What How)
16
Literature Review - Recommendations

• The literature review has to develop the theory
  and provide the basis for the hypotheses you will
  subsequently test.
• Weak theoretical support will lead to rejection.
• Your work must make a theoretical contribution
  (What constitutes a theoretical contribution,
  Academy of Management Review, 14(4), pp. 490-5).

(What How)
17
Literature Review - Recommendations

• The editorial team for a given journal tends to
  know the literature published in that journal the
  best.

(What How)
18
Measurement

• (JOM) Constructs not well developed
• (JSCM) - The authors fail to explicitly define
  most of the papers constructs (with the
  exception of trust), and do not refer to the
  extant literature regarding the
  operationalization of these constructs.
• Provide nominal definitions and justification for
  the various constructs.

(What How)
19
Measurement - Recommendations

• Constructs must be well-supported by the extant
  literature.
• Ideally, measures will have been previously used
  and validated.

(What How)
20
Methodology

• (JOM) The wave analysis for non-response bias
  is a somewhat weak approach to addressing this
  issue.
• (IJPMM) Confirmatory Factor Analysis Much of
  this discussion is too technical for our
  practitioner readers. Please consider moving
  some or all of this to an appendix and including
  in the body of the paper a brief discussion that
  practitioners could easily understand.
• (IJPMM) Please justify using P. 1 as indicating
  statistical significance.
• (JPIM) - The most worrying is the selection of
  data points or respondents. Purchasing or
  sourcing (what is that) managers may have one
  view - product and process developers another.

(What How)
21
Methodology, Contd

• (JBR) (Paragraph 1) refers to how missing data
  values were handled. This description is somewhat
  vague and needs to be expanded upon.
• (JBR) (Paragraph 2) states that the chi square
  value for the model is 60.06 and that p 0.16.
  This high a value of p would tend to indicate
  that the model is not a good fit, contrary to the
  claim made in the paper. This discrepancy needs
  to be explained. (SEM)
• (JBR) How specifically was the model empirically
  tested and shown to be at least identified? Also,
  please clarify what is meant by at least
  identified.
• (JPIM) - Earlier in the paper, you state that you
  developed a path model. Then you use regression
  to test the model. Why not use path analysis,
  covariance (e.g. LISREL, AMOS, etc.) analysis, or
  another form of SEM? Wouldnt they be more
  appropriate?

(What How)
22
Methodology, Contd

• (JSCM) - How did you select your sample?
• (JSCM) - Is your sample representative of the
  population?
• (JSCM) - Structural equation modeling would be a
  far better way to examine these relationships
  particularly Exhibits 2 and 3, which indicate
  substantial mediation.
• (JSCM) - The unit of analysis seems to be vague.
  What is a case? A product, a component, or ? Or
  product development activities in general?

(What How)
23
Methodology, Contd

• (DSJ) The authors need to state the actual
  number of questionnaires sent out and response
  rate. Also, the authors need to present an
  industry-wise break-down of the data.

(What How)
24
Methodology - Recommendations

• Non-response bias dont use wave analysis
• Adjust your treatment of the methods to the
  specific journal
• Is your chosen level of statistical significance
  adequate? If not, power?
• Discuss missing data treatment explicitly
• Discuss sampling technique explicitly
• Include response rate and sample demographics
• State your unit of analysis.

(What How)
25
Methodology - Recommendations

• EFA is not the best method for validating a
  measurement model. CFA is the preferred
  approach.
• Dont run a CFA and then test your theoretical
  model using regression (unless you have a very
  good reason to do so)
• Multiple Moderated Regression

(What How)
26
Discussion

• (JOM) with all this distracting (and largely
  tangential) literature being introduced here, it
  is difficult to tell what the implications are
  for this study.
• (IJPMM) Discussion Much of this section is
  unnecessary. Most of the arguments in this
  section have been, or should have been, made in
  the discussions of the literature and the
  construction of the model.

(Why)
27
Discussion

• (JSCM) - Your final conclusion and discussion of
  Exhibit 5 are too abrupt and leave the reader
  hanging. Discuss Exhibit 5 in greater detail.
  Explain the importance of why it is included.
  This appears to be the crux of your findings.
  Yet, you devote very little effort to explaining
  why it is important. Go into detail, and show
  the reader how he/she can use it effectively.
  This is perhaps the most important change you
  should make for this manuscript.

(Why)
28
Discussion - Recommendations

• The discussion and conclusion section of a
  manuscript are typically the most difficult for a
  writer to craft.
• Often, writers get through the theory and methods
  sections and seem to forget about the importance
  of writing about the contribution!
• Dont introduce new literature in the discussion
  section unless it is unavoidable.

(Why)
29
Conclusion

• (JOM) The conclusion section does not fit with
  the paper
• (JOM) - The conclusion section introduces new
  literature
• (JOM) - The manuscript clearly identifies an
  issue and a gap in the literature. There is a
  clear aim and intended contribution. The
  structure is not so well organized with different
  pieces of theory and data mixed to make analysis
  difficult to follow. When it comes to conclusion
  the contribution is not so clear any more.

(Why)
30
Conclusion - Recommendations

• This section is even more important that the
  discussion section. Readers will often read the
  abstract and the conclusion and make a decision
  about whether to read your work.
• A poor discussion section leads to a really poor
  conclusion.

(Why)
31
Appendices

• (JOM) The survey instrument is not included in
  the manuscript or not reference to the survey
  instrument is made in the manuscript
• (JBR) Please include as an appendix, the specific
  questions asked.
• (DSJ) There would be great clarity if a copy of
  the questionnaire or the exact form of the
  questions were listed in the manuscript.

(What How)
32
Appendices - Recommendations

• ALWAYS include your questions. If your survey is
  complicated, use the exact survey as an appendix.
• It is good practice to include a correlation
  matrix and a vector of standard deviations

(What How)
33
Limitations

• (JOM) The limitations of this study should be
  provided
• (JSCM) What are the limitations of this study?
  Suggest a path for future research in this area.

(Who, Where When)
34
Limitations - Recommendations

• Write this section with an eye towards the
  effective limitations of your work. Often,
  writers include the universally recognized
  limitations and perhaps skip more relevant
  limitations.

(Who, Where When)