# Research methods

Knowledge, insight and results depend on methods. This applies equally in Business, as well as in Academics. To be able to validate knowledge, you have to know in what way this knowledge has been achieved and assess, whether the methods used are capable to justify this knowledge. The complexity of research methods is underestimated often enough.

Research methods require in-depth knowledge in the fields of science and object theory, measurement and testing theory and technology, experimental design, data management, statistics, syntax programming, or the algorithms of various statistical programs etc. We have aquired this experience in years of researchpractice and are offering with pleasure.

But everything starts with an assumption, a theory...

## Theory

People say theory is no use without practise. On the contrary: Research without theory is no science. Theories are the bases of action and decision in everyday life and in research. Theories are present on many levels of the research process: Theories of measurement or testing, statistics etc. Hypotheses or scientific laws are deducted from theories about the research object (object theory) and reflected again in form of empirical results. The superordinate meta theory (theory of sciences) instructs scientific action (methodology). Because they can not be completely reduced to observational terms, theoretical terms (eg open or hypothetical constructs) are only partially interpretable.

Method Consult offers years of experience in dealing with object or meta theories, hypotheses etc. from a lot of scientific fields. We help you from your data to a theory, and/or to get from your theory to your data: For example, we can help to compile theories, analyse them (eg on logic, systematics and consistency), and derive chains of axioms or statements and hypotheses, and to collect the corresponding observational data. We also help you in the opposite direction by weaving a red thread from your observational data back to your starting point theory ...

## Interpretation

Research is a language game. Its meaning is fed by the theories and assumptions of these methods, models and designs, that are applied to the research object. Only certain interpretations are permitted within this framework. One has to learn this language before to be able to understand and mediate research. An insufficient methodical reflection leads only to an inadequate interpretation of, e.g. designs, results or hypotheses. Method Consult helps you to interpret science and research in all aspects (e.g. hypothesis formulation, statistics or design), for example, your data (e.g. distribution phenomenons), certain statistical conclusions (logics of research) or the outputs of several statistics packages. Examples:

**e.g. hypothesis formulation****e.g. methods****e.g. data analysis and statistics****e.g. results and interpretation****e.g. literature analysis**

We check whether your research hypotheses are precise, clear, and practically testable. For example, the sentence "There are researchers who never err" is not a practically falsifiable hypothesis. We also prevent you from so-called "collective hypotheses"

We check whether the chosen methods (experimental design, sampling, survey, analysis, etc.) are adequate for the research hypotheses, for example: Under no circumstances one must speak of measurement of change, when the design contains only one time point of measurement. If a test for a difference between two groups may not result in a statistical significance, this finding must not be misunderstood as if both groups were equal indeed.

We check what statistical methods suitable for your data (i.e. by testing the requirements). Especially with multivariate analysis you have to consider many special requirements, bias, and effects: e.g. multicollinearity, mediation effects, regression and interaction effects, or the bias of different clustering methods.

We check whether the results were obtained with appropriate methods, and also assist you with the statistical and substantive interpretation of your results, e.g. by explaining the outputs of various statistical programs, technical parameters and values, as well as the peculiarities of the statistical conclusion (logic of research). Statistical significance e.g. is closely related to the sample size, does not mean automatically clinical significance, and also, that a research hypothesis is ultimately proven.

We check the research literature that you provide on adherence to scientific standards. Trust, but verify. This way you are on the safe side.

## Methods for Business and Academics (a selection):

These methods can be used in every field: Research, evaluation, diagnosis. This includes the application areas such as e.g. quality control and improvement, market research, opinion and product research, city marketing, tourism research, client / consumer surveys, epidemiology, biostatistics / bioanalytics etc. The groups of approaches are listed alphabetically and represent only a small part of our services.

## Quantitative approaches

- e.g. Research methods
- e.g. Six Sigma
- e.g. Advanced Analytics
- e.g. Data Mining
- e.g. Business Intelligence
- e.g. Competitive Intelligence

Quantitative methods can be further classified into (non)parametric subvariants according to the measurement level of the variables used and the theoretically expected data distributions. Quantitative and qualitative methods can be combined. For example, qualitative variables can be analysed statistically, if they are coded on nominal or ordinal scales.

## Qualitative approaches

- Observation (e.g. by video, photo and part-taking observation)
- ex post-analyses (e.g. biographical and content analyses; CATI, CAPI, CAWI; atlas.ti, TEXTPAK, WINMAX)
- Conversation (e.g. single interview, group discussion)

Qualitative methods aim to describe social phenomena from the subjective perspective of individuals concerned. In spite of this closeness to social reality, qualitative methods are not necessarily superior methods. A subtype, the monological approach, only appears to capture the subjective world view of the persons concerned - the person in focus is usually not asked for a verifying/validating feedback, e.g. in content analyses. A validating combination of qualitative and quantitative methods may suggest itself, depending on the question of research.

## General Methods

- Analysis of many known psycho- and sociometric inventories (e.g. BDI, BFS, BIP, BSRI, CBCL, IOE, FKB-20, GBB, GDS, HAKEMP, HAS, KKG, MLDL, MOS, OQ, PTSD, RMBPCL, SAM, SCL, STAI, STAXI)
- tests of equivalence or difference
- reports
- data Entry (dual/double entry)
- data checking, cleaning and management
- designs
- documentation
- effect sizes
- experiments
- case-control Studies
- sample calculations (sample size)
- construction of questionnaires (question formulation, indices of difficulty, item analyses, selectivity, Cronbach's alpha, reliability, validity, etc.)
- visual presentations
- Interpretation
- coding of data
- plausibility analyses
- power computations
- sample drawing
- trouble-shooting (optimisation of suboptimal designs, critical evaluation of studies and publications, debugging of programs, checking of questionnaires, etc.)
- trial planning (randomisation/parallelisation, for details see section about Clinical Trials or the series by Dr. Schendera published in klinBiol, 2000-2001)

## Customized Research Methods

Special methods can be also developed and applied for the specific standards of your project. These project-specific methods would be customized to exactly fit your research requirement and your project (products, processes), and would meet quality standards of test and measurement theory. You specify what is to be measured and how it is to be done. You would not only learn what you want to know. You would also know exactly why you know it. If you want. Ask for further information.