Hello Abdelhak Hammoudi,

Generally speaking,  a hypothesis is a tentative statement addressing a  theoretical, hypothetical explanation of of a given observation  or measurement.In other words, it tends to provide a prediction about a specific outcome or a suggested explanation about a targeted research topic. As such, researchers' failure in producing  interesting and testable predictions may result in the  reformulation  of the hypothesis or reconsideration of  the defined subject. Practically, hypotheses are answers to research questions which are formulated as yes/no or wh-questions. Therefore, the use of MAY or WILL would not be appropriate because the questions are in the form of does, do, is, are , etc. or in the form of how does, what is, and so on. The modals such as MAY and WILL are hedging signals that alleviate the force of the prediction posed by the hypothesis.

Best of Luck,

R. Biria

Hello Abdelhak Hammoudi,

Generally speaking,  a hypothesis is a tentative statement addressing a  theoretical, hypothetical explanation of of a given observation  or measurement.In other words, it tends to provide a prediction about a specific outcome or a suggested explanation about a targeted research topic. As such, researchers' failure in producing  interesting and testable predictions may result in the  reformulation  of the hypothesis or reconsideration of  the defined subject. Practically, hypotheses are answers to research questions which are formulated as yes/no or wh-questions. Therefore, the use of MAY or WILL would not be appropriate because the questions are in the form of does, do, is, are , etc. or in the form of how does, what is, and so on. The modals such as MAY and WILL are hedging signals that alleviate the force of the prediction posed by the hypothesis.

Best of Luck,

R. Biria

I absolutely agree with Mr. Biria. Such weak dubious modal verbs  alleviate the predictability of hypotheses. Hypotheses have to be direct, straightforward, and unilateral. 

Whether or not your observed differences in students' scores in writing are significant is a matter of interpretation. A significance test on the hypothesis that this difference is zero will give you a probability value for your data (or data that is more "extreme" w.r.t. this hypothesis), the "p value" or "significance" (as a gradual measure!). If this probability is low (what "low" means depends on the context!) enough you may find that the difference is "statistically significant(ly different to zero)" and thus reject the tested hypothesis. That's your (expert) opinion.

You test the hypothesis (H0) that the difference IS zero.

You MAY reject H0 in light of your data.

You WILL reject H0 if (in case or in future) the data is very unlikely under H0.

hypotheses are about true but unknown values of parameters of interest. a hypothesis states that such a true value IS 0 or IS at least 5 or IS not more than 17. it doesn't really make a lot of sense to say that a true value could or might be 0, because such a statement would "automatically" be true. it's like saying that the true but yet unknown weather of tomorrow might be good... - and therefore might also be bad. no need to test that because we already know it's true without any test.

moreover, you can't use the word "significant" when formulating a hypothesis.  as explained, hypotheses are about true parameter values, but significance refers to the estimations(!) of these unknown parameters! if you use a mean to estimate a true effect, then that mean value will hopefully be sufficiently far from your postulated true value 0 so that you can be pretty sure that this difference is not pure coincidence, maybe due to an unlucky choice of your random sample. then you can say that your mean is significantly different from 0. in other words, only "observed values" (or functions of observations like mean values or other test statistics) can be "significant" but not the parameters in (null-)hypotheses themselves!

@mr mahboudi: your use of words like "weak", "dubious" and "unpredictable" makes it sound like a hypothesis reflects the manly or less manly character of the researcher. you're right, one cannot use "may" when stating a hypothesis, but the reason why one cannot has nothing to do with the word being "weak". furthermore, i don't really know what you mean by "predictability of hypotheses"; you can predict outcomes from a regression model, but how do you want to predict hypotheses? and your statement that hypotheses have to be unilateral is just plain wrong, if you forgive me for saying that. "unilateral" and "bilateral" have a very specific meaning in this context...

It is necessary to take care in writing hypotheses and conclusions. The word 'may' implies permission, while 'might' implies probability. Neither should be in a hypothesis without qualification and explanation.

Hypothesis testing has become an unfortunate diversion in science. We should be doing science with clear statements based a on a structure of logic. Many investigations involve probability and could be expressed with terms of some, more, most, usually, in relation to, etc. Such terms require explanation and must be replaced by a definite expression. (As you suggested will for may.) The definite expression must be qualified and quantified. The quantification is what constitutes success or failure.

Thanks a lot! interesting answers.

Thank you all for the answers,

I do have a question, regarding writing the findings of a hypothesis. Is it correct to say that, It may be inferred that

Kindly advice me.

Thanks