What is a testable hypothesis

Make hypotheses

Crash course hypotheses

Sentence structure

A hypothesis is a clear and precise statement. It must be formulated in such a way that its content-related statement can either be clearly confirmed (“verified”) or rejected (“falsified”) at the end of the research work. Avoid words and phrases in the hypothesis that make this unambiguous verification or falsification difficult (e.g. "rather good", "relatively bad", "maybe", "probably", etc.).


You should carefully consider the selection and definition of the characteristics in the hypothesis. This determines the scope of your research work and also influences the results. In the example hypothesis above, the focus is on the characteristic “proportion of male apprentices”. It is important how this proportion is determined (e.g. using data from Statistics Austria, using our own survey, etc.). Different survey methods for the characteristic may also lead to different research results. With (d) research, it is therefore important to determine the characteristics as precisely as possible in advance and also to think about their measurement.

Types of hypotheses

Depending on the thematic context in which hypotheses are drawn up, there are distinctions between types of hypotheses.

At Connection hypotheses Direct connections between mostly two characteristics are examined (e.g. "The proportion of female apprentices in MINT occupations is lower than in other occupations.").

The Difference hypothesis compares the same characteristics of different groups (e.g. "The proportion of male apprentices is higher in Eastern Austria than in Western Austria.").

In the Change hypothesis a presumption is formulated that a characteristic has changed over time (e.g. "The proportion of male apprentices in Austria decreased between 2005 and 2015.").

By reviewing the Individual hypothesis No generalizations are made, but only assumptions about a specific case are examined (e.g. "Company XYZ has more than 50 percent women among its apprentices.").

Null hypothesis and alternative hypothesis

For a scientific work, two hypotheses are always set up for every suspected connection in a situation.

The Null hypothesis is formulated in such a way that a connection between the characteristics is excluded.

The Alternative hypothesis however, creates a clear connection between the features.

In general, examinations and research are used to check whether the null hypothesis can be clearly statistically rejected ("falsified"). In this case, the alternative hypothesis is then assumed to be correct. Only if the null hypothesis is not clearly falsified can the alternative hypothesis also be examined for correctness.