These guidelines are for students who want to write a quantitative research paper with me. These guidelines are in English because my expectations grow out of the Anglo-American scholarly publishing tradition and because some of the concepts that I use in these guidelines do not have exact German analogs. By writing in English I hope to avoid any misunderstanding.

Wenn Sie Englisch nicht gut lesen können, schlage ich vor, dass Sie einen anderen Gutachter aussuchen.

In these guidelines I generally refer generically to research papers. The expectations for BA theses are different than for MA theses, which are different than for doctorates. Nonetheless all should follow the same format.

Digital Copy

I will only read a digital copy of your research paper. If you want me to read and grade your work, I must receive it in digital form in the following formats:

Docx, Doc, or ODF fomat; or
PDF with a minimum 12 point font; or
ePUB format for eReaders.

According to university regulations, you must also provide paper copies of to the appropriate office.

Expectations

This section explains what to expect in applying and in working with me.

Research colloquium

Those who live in Berlin will be required to attend the Research Colloquium.

Those not in Berlin should plan on communicating with me at least once per semester via video-conference, skype, phone, or email.
Participation in the Research Colloquium via videoconference may also be possible.

Term paper (Hausarbeit), BA or MA Thesis

Application

For a BA or MA student, please send an email with a few paragraphs explaining what you want to write about and how you intend to do it. If you are still unclear about what or how, just say so and we can discuss it. The second step is to make an appointment to talk with me in person or (if necessary) by phone or skype. Please make the appointment with my secretary via: office (at) ibi.hu-berlin.de.

Term Paper or BA expectations

A term paper or BA thesis can focus more on summarizing the existing literature and authors will not be held to rigorous standards for research questions or data gathering. It is important that such papers show some understanding of research methodologies.

MA Thesis expectations

An MA thesis should represent a minor contribution to the scholarly literature. The scope cannot be as broad as with a doctoral dissertation, but the research question should be carefully designed and the methodology carried out correctly. Limits on data gathering should be explained, but narrow limits are normal and acceptable.

Doctoral Dissertation

Application

I will accept only a limited number of new doctoral students. In order to apply to have me as your PhD advisor please:

Step 1: Send an email explaining what you want to write about and what scholarly methodology you want to use. If you are unsure about what or how, we need to discuss this.

Step 2: After we agree on these basics, you should write a proposal of 5 - 10 pages that explains your research methodology, where and how you will gather your data, and what you want to study. If you have an idea about your research questions, you should include that too.

Step 3: I will review proposals and let you know if I can accept you as a doctoral student.

Step 4: Please sign the doctoral research agreement form and return it to me.

Doctoral Dissertation expectations

A doctoral dissertation should make a genuine contribution to the literature. Authors should not expect to create a revolution in the field, but should offer solid evidence and argumentation that provides new knowledge.

Quantitative Research Topics

There are three broad areas in which I will accept quantitative topics:

public dataset analysis (including infometrics)
survey research
experiments (using econometric or psych-based methods)

Please be aware that I may well NOT accept quantitative research topics outside these areas. If you want to do qualitative research or computer science-based research, please look at the appropriate guidelines.

Basics

I care less about the topic than the research methodology and the methods that you plan to use. The standard elements of quantitative research in our field include:

Research question

The research question is essential and should be stated explicitly, preferably near the beginning. The question itself need not be unique, but either new data (new information) should be used to answer it or a research method should be used that has not been applied to the question before. The research question should ideally grow out of the scholarly context (ie out of the library and information science literature), rather than merely be something that seems interesting to the author. For quantitative research, a good research question is one that the author can answer specifically with a yes or no, with numbers or percentages, or by demonstrating that the answer is one out of a previously described set of options.

WARNING: any paper that does not include an explicit research question will be marked down as much as a full grade.

Quantitative Research methodology

Anyone doing quantitative research needs to consider the following issues. Examples will be taken mainly from survey research, since that is so common in the LIS field.

Population

The first issue for any quantitative research is to determine the population. For survey research this means determining the group about whom the researcher wants to generalize and for whom or from whom the researcher can get data. The population of all professors in the world is problematic, because the definition of professor varies widely from culture to culture and because gathering data would be difficult.

I strongly discourage the use of a publicly available online survey because it is hard to determine who has participated and what their motivation for participation is. By and large people will take part in a public survey only if they have strong feelings about the subject and if they have free time -- in other words, people who are atypical.

For any survey, the researcher needs enough demographic information about the population to determine how representative the sample is.

Representativeness

For a survey to produce statistically meaningful results, it must have a reasonable claim to being representative. Convenience samples are common in research about human populations. Students in a class are a good example, because the researcher knows to what degree the students mirror the target population. For example, MBA students are often used to represent managers in business, since there is a reasonable expectation that they will soon be part of that group.

Demographic information is one way to tell whether a convenience sample represents the population as a whole.

Random Selection

For most inferential statistics (including simple percentages that are meant apply to the population as a whole) random selection is necessary. If the selection is not random, then the results may well have a built-in bias.

Meaning

Many surveys fail to make the meaning of their questions clear. It is dangerous to assume that a population uses the same definition of key words that the researcher does, without explicitly defining them.

Data Model

A data model assists the researcher in evaluating the meaning of the responses. In voter surveys, for example, it is important to have a model to determine whether persons who claim they will vote will actually go to the polls.

Tests

Researchers need to consider how well their data match the requirements for various statistical tests. Parametric statistics assume, for example, make more assumptions about the distribution than non-parametric tests.

WARNING: any paper that does not include a discussion of an explicit research method will be marked down as much as a full grade.

Research Methodology (other)

If you want to combine a quantitative methodology with a qualitative one, you should need to talk with me. I will not accept research papers that do not involve empirical research.

For quantitative methodologies or for computer science-based research, please see my other guidelines.

Permission and Anonymization

In general students should ask permission to interview people whom they are studying. A written or recorded permission is best, but occasionally an informal verbal permission is all that is possible. The date and time should be recorded, in addition to the language used in so far as it can be remembered.

Permissions for observations are good to have as well, but need not be taken to extremes. Actions in public places are certainly legitimate to observe and report on. Behavior in large workgroups is a middle ground between the public and private. In general it suffices to tell the group that you may include observations of their work in a research paper, and members should have a chance to raise questions or concerns. Individual permissions are not needed, unless a particular person becomes the focus of a study.

The need for a permission depends in part on how easy it is to make the observations anonymous. Changing names and personal descriptions is standard, but in smaller groups particular persons may stand out more. It is worth remembering also that people will recognize themselves more readily than others will. The more negative observations are, the more important it is to anonymize or to have a permission. Positive or patently neutral observations can be made more freely.

The Philosophische Fakultät 1 generally follows the ethical guidelines of the American Anthropological Association. The Fakultät has an ethics review board, for times explicit ethics determinations are needed (as is often the case in the US).

Scholarly context

The scholarly context is sometimes called the literature review. Students sometimes add this as an afterthought and merely do a database search to gather references. The scholarly context should in fact be the first step in preparing a research paper, because it is primarily out of our scholarly knowledge about a subject that a good research question should emerge.

The discussion of the scholarly context traditionally takes place in a separate section, but that need not be the case. It can, for example, very reasonably be integrated into the discussion of why a particular research question makes sense.

WARNING: any paper that does not include a discussion of the scholarly context (literature review) will be marked down as much as a full grade. Claiming that no literature can be found is NOT acceptable.

Data gathering and Analysis

The data gathering lies at the heart of a scholarly work in quantitative LIS research. The data need not be new. An existing data set may be reexamined with different statistics, selection methods or ideas. Authors should explicitly discuss their sources of data and comment on its qualities, which might include accuracy, representativeness, or uniqueness.

What data is relevant for quantitative research? It includes all forms of externally gathered information: formal numeric data sets, text-based sources -- essentially anything that can be counted and numerically processed.

Data analysis is the process of drawing conclusions on the basis of data in a research paper. The effectiveness and credibility of the analysis may determine whether the paper makes any contribution to knowledge or not.

Outline

The internal structure of a dissertation should reflect standard social science practice. In general this means an outline that has roughly the following structure:

research question
research method
scholarly context
data gathering
data analysis
conclusions

Some variation in the order is certainly possible, but having these sections and labeling them as such aids the reader in understanding how the research was done.

Grading

Research papers will receive grades and feedback on the following topics:

Is the research question well-designed?
Can the research methodology answer the research question?
Does the literature review explain why the research question is of interest and the method appropriate?
Are the data adequate to answer the research question? Were the data analyzed properly? Does special consideration need to be given to the difficulties of getting appropriate data?

Background Reading

Authors of a term paper (Hausarbeit), a thesis, or a dissertation are writing in some sense for a very narrow audience: two or three readers (Gutachter) who will give the work a grade. Under these circumstances students should take the trouble to read what their readers have written on their subject. I have no problem with intelligent disagreement, but blithely repeating opinions that I patently reject will not improve my opinion of a work.

Copying and Quoting

Plagiarism is of course not allowed, but questions often arise about legitimate copying. The basic principle is that any time researchers quote an exact expression (literally another person’s words), the expression should be in quotation marks and should cite the original source.

If researchers want to paraphrase what another person says using their own words, then a proper scholarly reference is still needed.

If ideas are borrowed, but not the expression of them, then a reference is still appropriate, but may need to be general. For example, if I say “Clifford Geertz speaks repeatedly in his books about detailed observation.” then I cannot refer to a specific book as a reference and it is not currently normal to cite all of his works to support such a statement. Be aware, however, that expectations change over time.

Paper-writing Mechanics

The mechanics here refer to the details of how a research paper should look, not to the content.

References

Your references should be in the British (EU) version of Harvard format or in University of Chicago (Turabian) style. Systems like Citavi, Zotero, Refworks and other citation management systems are able to generate either either Harvard or Chicago (Turabian) references automatically.

Footnotes

I prefer endnotes to footnotes. If you feel strongly about using footnotes, then they should only be used for URLs. Footnotes in modern research works should never be used for text. If the comment is important, then include it in the body of the text. If it is not important enough for the body of the text, then leave it out.

Style

A research paper ought to have clear simple prose that relies on nouns and verbs to make its points, rather than adjectives and adverbs. A research paper is not a work of literature, but a scholarly argument.

Language

I am willing to read a research paper in either German or English and in some cases might agree to reading one in French (ask first, please). If you want to write in English and it is not your native language, I will be quite flexible about minor grammar, syntax, or word-choice errors – probably more so than most German teachers. For those who are ambitious and hope for future publication to a larger audience, writing in English means a larger audience.

Quotations in the body of the text that are in German, English, or French need not be translated for me. Short Dutch, Swedish, Danish, Norwegian, Spanish and Italian quotations may also be left in the original as long as the context is clear.

References

Below are two standard works that can be helpful for quantitative research. Unfortunately neither is in German. These are only examples, though Dillman is very widely used as the “Bible” for survey research.

Dillman, D.A., 2007. Mail and Internet Surveys: the tailored design method, 2nd ed. John Wiley & Sons, Hoboken, N. J.

Hinton, P.R., 1995. Statistics Explained: A Guide for Social Science Students. Routledge, New York.

version 2 Sept 2013