When conducting a hypothesis test, you need to decide whether to accept or reject the null hypothesis at least partly on the basis of your observed data. This may be difficult when only a few variables are involved, such as when reading this article.

Many researchers encourage their students to run a hypothesis test before *teaching basic statistical inference*, so this article is for you! In this article, we will discuss how to run a hypothesis test and what questions to ask.

When doing a hypothesis test, there are some key questions that you should ask. These questions can be divided into two parts: acceptance and rejection. The first part of these questions can be answered in the following sections, but first we must discuss the second part.

Rejection of the *null hypothesis means changing something* about the study or adding another variable to the list. This **may require introducing new evidence** or **confirming existing evidence** with a new variable.

## Do not reject the null hypothesis

When developing hypotheses, it is important to be careful to define the null hypothesis and the expected hypothesis.

The null hypothesis can only be rejected if the *expected hypothesis holds true*. For example, if we were testing the theory that coffee is good for you, then the null hypothesis would be that it isn’t.

In this case, the expected hypothesis would be that it is not! As mentioned earlier, there are certain hypotheses that come with a “no-nos” list, but in this article we are going to focus on two: rejecting the null when there is no difference between the two groups and rejecting the **strongly suggested test** when there is a clear difference between groups.

Rejecting the null when there is no difference between groups is called passing on information and should be done with care. Passing information can lead to **either choosing one** of only two possible solutions or being told that something isn’t true.

## You can determine whether your sample is different from what was expected

One of the most important things you can do as a statistics professional is to determine whether your sample is different from what was expected. This means looking at your population and seeing if there are any examples of what people do not follow the rule set.

The *term population refers* to the total group of people that you are looking at. As an example, people who *wear glasses* have a population that includes those who wear glasses for only part of the day.

The term different from what was expected refers to a hypothesis that was presented and determined to be wrong. For example, when *wearing sunglasses*, **people typically look** through them and see the world through them. By testing whether they can do this, you can determine whether your hypothesis was wrong.

When determining whether a sample is different from what was expected, it is important to look at both quantity and quality of the data.

## You can determine the probability of getting a result as extreme or more so than what you observed

When you have a hypothesis test, you can *decide whether* to accept or reject the result based on what you observe. For instance, if you observe a high probability of getting a T, then you can conclude that the ** test resulted** in a positive result.

On the other hand, if you observe a low probability of getting a T, then you can conclude that the test resulted in a negative result. At least in this case, your hypothesis was not falsified!

This is important to keep in mind when performing an hypothesis test. When there is less than half of certainty that the test result is going to be positive or negative, then there is still time to make an *informed decision whether* to follow your intuition and accept or reject the result on the basis of what they observed.

When doing an hypothesis test, make sure to account for all possible outcomes.

## Understand the meaning of a p-value

When doing a hypothesis test, you should understand the p-value. A p-value is the probability that your **treatment group would** have beaten your control group’s treatment group by just one point on a 0 to 1 scale.

A *small difference* between the groups can have a large effect on the outcome, so using a **small difference makes sense**. That is why it is common to use 0.05 or lower for these tests!

Performing a hypothesis test can be done in several ways. You can do it by hand, using step-by-step instructions, or even use the Internet! Either way, you should learn how to do it!

The first way to perform a hypothesis test is by simply performing the test yourself! By simply taking what others say and trying to apply them on YOUR work environment and tests.

## Know how to calculate a p-value

When you perform a hypothesis test, you need to know how to calculate the p-value. The p-value is the standard measure of statistical significance.

The p-value was created so that we could determine if a hypothesis was True or Not using a very small difference. If the difference is large, it means that the population is not totally homogeneous and there is a chance that the **hypothesized thing exists**.

So, how do you calculate the p-value? Well, there are *two ways* to do it. You can use a software program or you can write it down. Both of these methods work!

Useful tip: If you want to learn more about *performing hypothesis tests*, read Through Hypothesis Testing: Building and Challenging Hypotheses by Gregory Westland.

## Know how to interpret a p-value

When doing a hypothesis test, you need to know the p-value. The p-value indicates how likely it is that the * observed result* was caused by theelmanageur vs. themanageur condition.

A low value of the p-value indicates that it is more likely that the observed result was **caused bytheelmanageur vs**. themanageur condition. A medium value of the p-value indicates that it is more likely that the **observed resultwas caused byeithertheelmanageurexpectations** or boththeemanagementandtheresults.

When doing a hypothesis test, you want to have a medium value of the p-value so that your confidence in your hypothesis can be measured.

## Understand the role of an alpha value

When you’re performing a hypothesis test, your goal is to find the smallest amount of contrast that produces a correct match.

This is done by knowing the alpha value of your competitor’s product. This value represents the smallest change in product performance that produces a correct match. For example, if the competitor’s product has an **orange square bottle** with a price of $5, then producing a tank top with this *size would produce* a match!

In order to meet this goal, you must know how much contrast your *tank top needs* to achieve its effect. This is called the Contrast Threshold Value (CTV). This value is what separates an experimental result from a chance result.

When doing hypothesis tests, it is important to know how much contrast your sample needs in order for you to reach the level of confidence required for the hypothesis.

## Know how to choose an alpha value

When deciding how *much money* to spend on advertising, you need to consider what alpha value your product or service represents.

An alpha value is the highest value a person considers when making decisions. It is the level of attractiveness, desirable quality, or success that someone looks for in a product or service.

When shopping for a car, the salesperson who offers the best deal for now and in the long run is the one that people choose over those that *offer less money* now but may not be as successful in the future.

When shopping for a business, you need to find the best bet that will **yield good results** in the long run. You do this by looking at what Alpha Values represent and how they matter when buying a business.