does data prove a hypothesis true

• Comprehensive Learning Paths
• 150+ Hours of Videos
• Complete Access to Jupyter notebooks, Datasets, References.

Hypothesis Testing – A Deep Dive into Hypothesis Testing, The Backbone of Statistical Inference

• September 21, 2023

Explore the intricacies of hypothesis testing, a cornerstone of statistical analysis. Dive into methods, interpretations, and applications for making data-driven decisions.

In this Blog post we will learn:

• What is Hypothesis Testing?
• Steps in Hypothesis Testing 2.1. Set up Hypotheses: Null and Alternative 2.2. Choose a Significance Level (α) 2.3. Calculate a test statistic and P-Value 2.4. Make a Decision
• Example : Testing a new drug.
• Example in python

1. What is Hypothesis Testing?

In simple terms, hypothesis testing is a method used to make decisions or inferences about population parameters based on sample data. Imagine being handed a dice and asked if it’s biased. By rolling it a few times and analyzing the outcomes, you’d be engaging in the essence of hypothesis testing.

Think of hypothesis testing as the scientific method of the statistics world. Suppose you hear claims like “This new drug works wonders!” or “Our new website design boosts sales.” How do you know if these statements hold water? Enter hypothesis testing.

2. Steps in Hypothesis Testing

• Set up Hypotheses : Begin with a null hypothesis (H0) and an alternative hypothesis (Ha).
• Choose a Significance Level (α) : Typically 0.05, this is the probability of rejecting the null hypothesis when it’s actually true. Think of it as the chance of accusing an innocent person.
• Calculate Test statistic and P-Value : Gather evidence (data) and calculate a test statistic.
• p-value : This is the probability of observing the data, given that the null hypothesis is true. A small p-value (typically ≤ 0.05) suggests the data is inconsistent with the null hypothesis.
• Decision Rule : If the p-value is less than or equal to α, you reject the null hypothesis in favor of the alternative.

2.1. Set up Hypotheses: Null and Alternative

Before diving into testing, we must formulate hypotheses. The null hypothesis (H0) represents the default assumption, while the alternative hypothesis (H1) challenges it.

For instance, in drug testing, H0 : “The new drug is no better than the existing one,” H1 : “The new drug is superior .”

2.2. Choose a Significance Level (α)

When You collect and analyze data to test H0 and H1 hypotheses. Based on your analysis, you decide whether to reject the null hypothesis in favor of the alternative, or fail to reject / Accept the null hypothesis.

The significance level, often denoted by $α$, represents the probability of rejecting the null hypothesis when it is actually true.

In other words, it’s the risk you’re willing to take of making a Type I error (false positive).

Type I Error (False Positive) :

• Symbolized by the Greek letter alpha (α).
• Occurs when you incorrectly reject a true null hypothesis . In other words, you conclude that there is an effect or difference when, in reality, there isn’t.
• The probability of making a Type I error is denoted by the significance level of a test. Commonly, tests are conducted at the 0.05 significance level , which means there’s a 5% chance of making a Type I error .
• Commonly used significance levels are 0.01, 0.05, and 0.10, but the choice depends on the context of the study and the level of risk one is willing to accept.

Example : If a drug is not effective (truth), but a clinical trial incorrectly concludes that it is effective (based on the sample data), then a Type I error has occurred.

Type II Error (False Negative) :

• Symbolized by the Greek letter beta (β).
• Occurs when you accept a false null hypothesis . This means you conclude there is no effect or difference when, in reality, there is.
• The probability of making a Type II error is denoted by β. The power of a test (1 – β) represents the probability of correctly rejecting a false null hypothesis.

Example : If a drug is effective (truth), but a clinical trial incorrectly concludes that it is not effective (based on the sample data), then a Type II error has occurred.

Balancing the Errors :

In practice, there’s a trade-off between Type I and Type II errors. Reducing the risk of one typically increases the risk of the other. For example, if you want to decrease the probability of a Type I error (by setting a lower significance level), you might increase the probability of a Type II error unless you compensate by collecting more data or making other adjustments.

It’s essential to understand the consequences of both types of errors in any given context. In some situations, a Type I error might be more severe, while in others, a Type II error might be of greater concern. This understanding guides researchers in designing their experiments and choosing appropriate significance levels.

2.3. Calculate a test statistic and P-Value

Test statistic : A test statistic is a single number that helps us understand how far our sample data is from what we’d expect under a null hypothesis (a basic assumption we’re trying to test against). Generally, the larger the test statistic, the more evidence we have against our null hypothesis. It helps us decide whether the differences we observe in our data are due to random chance or if there’s an actual effect.

P-value : The P-value tells us how likely we would get our observed results (or something more extreme) if the null hypothesis were true. It’s a value between 0 and 1. – A smaller P-value (typically below 0.05) means that the observation is rare under the null hypothesis, so we might reject the null hypothesis. – A larger P-value suggests that what we observed could easily happen by random chance, so we might not reject the null hypothesis.

2.4. Make a Decision

Relationship between $α$ and P-Value

When conducting a hypothesis test:

We then calculate the p-value from our sample data and the test statistic.

Finally, we compare the p-value to our chosen $α$:

• If $p−value≤α$: We reject the null hypothesis in favor of the alternative hypothesis. The result is said to be statistically significant.
• If $p−value>α$: We fail to reject the null hypothesis. There isn’t enough statistical evidence to support the alternative hypothesis.

3. Example : Testing a new drug.

Imagine we are investigating whether a new drug is effective at treating headaches faster than drug B.

Setting Up the Experiment : You gather 100 people who suffer from headaches. Half of them (50 people) are given the new drug (let’s call this the ‘Drug Group’), and the other half are given a sugar pill, which doesn’t contain any medication.

• Set up Hypotheses : Before starting, you make a prediction:
• Null Hypothesis (H0): The new drug has no effect. Any difference in healing time between the two groups is just due to random chance.
• Alternative Hypothesis (H1): The new drug does have an effect. The difference in healing time between the two groups is significant and not just by chance.

Calculate Test statistic and P-Value : After the experiment, you analyze the data. The “test statistic” is a number that helps you understand the difference between the two groups in terms of standard units.

For instance, let’s say:

• The average healing time in the Drug Group is 2 hours.
• The average healing time in the Placebo Group is 3 hours.

The test statistic helps you understand how significant this 1-hour difference is. If the groups are large and the spread of healing times in each group is small, then this difference might be significant. But if there’s a huge variation in healing times, the 1-hour difference might not be so special.

Imagine the P-value as answering this question: “If the new drug had NO real effect, what’s the probability that I’d see a difference as extreme (or more extreme) as the one I found, just by random chance?”

For instance:

• P-value of 0.01 means there’s a 1% chance that the observed difference (or a more extreme difference) would occur if the drug had no effect. That’s pretty rare, so we might consider the drug effective.
• P-value of 0.5 means there’s a 50% chance you’d see this difference just by chance. That’s pretty high, so we might not be convinced the drug is doing much.
• If the P-value is less than ($α$) 0.05: the results are “statistically significant,” and they might reject the null hypothesis , believing the new drug has an effect.
• If the P-value is greater than ($α$) 0.05: the results are not statistically significant, and they don’t reject the null hypothesis , remaining unsure if the drug has a genuine effect.

4. Example in python

For simplicity, let’s say we’re using a t-test (common for comparing means). Let’s dive into Python:

Making a Decision : “The results are statistically significant! p-value < 0.05 , The drug seems to have an effect!” If not, we’d say, “Looks like the drug isn’t as miraculous as we thought.”

5. Conclusion

Hypothesis testing is an indispensable tool in data science, allowing us to make data-driven decisions with confidence. By understanding its principles, conducting tests properly, and considering real-world applications, you can harness the power of hypothesis testing to unlock valuable insights from your data.

More Articles

Correlation – connecting the dots, the role of correlation in data analysis, sampling and sampling distributions – a comprehensive guide on sampling and sampling distributions, law of large numbers – a deep dive into the world of statistics, central limit theorem – a deep dive into central limit theorem and its significance in statistics, skewness and kurtosis – peaks and tails, understanding data through skewness and kurtosis”, similar articles, complete introduction to linear regression in r, how to implement common statistical significance tests and find the p value, logistic regression – a complete tutorial with examples in r.

Subscribe to Machine Learning Plus for high value data science content

© Machinelearningplus. All rights reserved.

Machine Learning A-Z™: Hands-On Python & R In Data Science

Free sample videos:.

Hypothesis Testing: Understanding the Basics, Types, and Importance

Hypothesis testing is a statistical method used to determine whether a hypothesis about a population parameter is true or not. This technique helps researchers and decision-makers make informed decisions based on evidence rather than guesses. Hypothesis testing is an essential tool in scientific research, social sciences, and business analysis. In this article, we will delve deeper into the basics of hypothesis testing, types of hypotheses, significance level, p-values, and the importance of hypothesis testing.

• Introduction

What is a hypothesis?

What is hypothesis testing, types of hypotheses, null hypothesis, alternative hypothesis, one-tailed and two-tailed tests, significance level and p-values, avoiding type i and type ii errors, making informed decisions, testing business strategies, a/b testing, formulating the null and alternative hypotheses, selecting the appropriate test, setting the level of significance, calculating the p-value, making a decision, common misconceptions about hypothesis testing, understanding hypothesis testing.

A hypothesis is an assumption or a proposition made about a population parameter. It is a statement that can be tested and either supported or refuted. For example, a hypothesis could be that a new medication reduces the severity of symptoms in patients with a particular disease.

Hypothesis testing is a statistical method that helps to determine whether a hypothesis is true or not. It is a procedure that involves collecting and analyzing data to evaluate the probability of the null hypothesis being true. The null hypothesis is the hypothesis that there is no significant difference between a sample and the population.

In hypothesis testing, there are two types of hypotheses: null and alternative.

The null hypothesis, denoted by H0, is a statement of no effect, no relationship, or no difference between the sample and the population. It is assumed to be true until there is sufficient evidence to reject it. For example, the null hypothesis could be that there is no significant difference in the blood pressure of patients who received the medication and those who received a placebo.

The alternative hypothesis, denoted by H1, is a statement of an effect, relationship, or difference between the sample and the population. It is the opposite of the null hypothesis. For example, the alternative hypothesis could be that the medication reduces the blood pressure of patients compared to those who received a placebo.

There are two types of alternative hypotheses: one-tailed and two-tailed. A one-tailed test is used when there is a directional hypothesis. For example, the hypothesis could be that the medication reduces blood pressure. A two-tailed test is used when there is a non-directional hypothesis. For example, the hypothesis could be that there is a significant difference in blood pressure between patients who received the medication and those who received a placebo.

The significance level, denoted by α, is the probability of rejecting the null hypothesis when it is true. It is set at the beginning of the test, usually at 5% or 1%. The p-value is the probability of obtaining a test statistic as extreme as

or more extreme than the observed one, assuming that the null hypothesis is true. If the p-value is less than the significance level, we reject the null hypothesis.

Importance of Hypothesis Testing

Hypothesis testing helps to avoid Type I and Type II errors. Type I error occurs when we reject the null hypothesis when it is actually true. Type II error occurs when we fail to reject the null hypothesis when it is actually false. By setting a significance level and calculating the p-value, we can control the probability of making these errors.

Hypothesis testing helps researchers and decision-makers make informed decisions based on evidence. For example, a medical researcher can use hypothesis testing to determine the effectiveness of a new drug. A business analyst can use hypothesis testing to evaluate the performance of a marketing campaign. By testing hypotheses, decision-makers can avoid making decisions based on guesses or assumptions.

Hypothesis testing is widely used in business analysis to test strategies and make data-driven decisions. For example, a business owner can use hypothesis testing to determine whether a new product will be profitable. By conducting A/B testing, businesses can compare the performance of two versions of a product and make data-driven decisions.

Examples of Hypothesis Testing

• A/B testing is a popular technique used in online marketing and web design. It involves comparing two versions of a webpage or an advertisement to determine which one performs better. By conducting A/B testing, businesses can optimize their websites and advertisements to increase conversions and sales.

A t-test is used to compare the means of two samples. It is commonly used in medical research, social sciences, and business analysis. For example, a researcher can use a t-test to determine whether there is a significant difference in the cholesterol levels of patients who received a new drug and those who received a placebo.

Analysis of Variance (ANOVA) is a statistical technique used to compare the means of more than two samples. It is commonly used in medical research, social sciences, and business analysis. For example, a business owner can use ANOVA to determine whether there is a significant difference in the sales performance of three different stores.

Steps in Hypothesis Testing

The first step in hypothesis testing is to formulate the null and alternative hypotheses. The null hypothesis is the hypothesis that there is no significant difference between the sample and the population, while the alternative hypothesis is the opposite.

The second step is to select the appropriate test based on the type of data and the research question. There are different types of tests for different types of data, such as t-test for continuous data and chi-square test for categorical data.

The third step is to set the level of significance, which is usually 5% or 1%. The significance level represents the probability of rejecting the null hypothesis when it is actually true.

The fourth step is to calculate the p-value, which represents the probability of obtaining a test statistic as extreme as or more extreme than the observed one, assuming that the null hypothesis is true.

The final step is to make a decision based on the p-value and the significance level. If the p-value is less than the significance level, we reject the null hypothesis. Otherwise, we fail to reject the null hypothesis.

There are several common misconceptions about hypothesis testing. One of the most common misconceptions is that rejecting the null hypothesis means that the alternative hypothesis is true. However

this is not necessarily the case. Rejecting the null hypothesis only means that there is evidence against it, but it does not prove that the alternative hypothesis is true. Another common misconception is that hypothesis testing can prove causality. However, hypothesis testing can only provide evidence for or against a hypothesis, and causality can only be inferred from a well-designed experiment.

Hypothesis testing is an important statistical technique used to test hypotheses and make informed decisions based on evidence. It helps to avoid Type I and Type II errors, and it is widely used in medical research, social sciences, and business analysis. By following the steps in hypothesis testing and avoiding common misconceptions, researchers and decision-makers can make data-driven decisions and avoid making decisions based on guesses or assumptions.

• What is the difference between Type I and Type II errors in hypothesis testing?
• Type I error occurs when we reject the null hypothesis when it is actually true, while Type II error occurs when we fail to reject the null hypothesis when it is actually false.
• How do you select the appropriate test in hypothesis testing?
• The appropriate test is selected based on the type of data and the research question. There are different types of tests for different types of data, such as t-test for continuous data and chi-square test for categorical data.
• Can hypothesis testing prove causality?
• No, hypothesis testing can only provide evidence for or against a hypothesis, and causality can only be inferred from a well-designed experiment.
• Why is hypothesis testing important in business analysis?
• Hypothesis testing is important in business analysis because it helps businesses make data-driven decisions and avoid making decisions based on guesses or assumptions. By testing hypotheses, businesses can evaluate the effectiveness of their strategies and optimize their performance.
• What is A/B testing?

If you want to learn more about statistical analysis, including central tendency measures, check out our  comprehensive statistical course . Our course provides a hands-on learning experience that covers all the essential statistical concepts and tools, empowering you to analyze complex data with confidence. With practical examples and interactive exercises, you’ll gain the skills you need to succeed in your statistical analysis endeavors. Enroll now and take your statistical knowledge to the next level!

If you’re looking to jumpstart your career as a data analyst, consider enrolling in our comprehensive  Data Analyst Bootcamp with Internship program . Our program provides you with the skills and experience necessary to succeed in today’s data-driven world. You’ll learn the fundamentals of statistical analysis, as well as how to use tools such as SQL, Python, Excel, and PowerBI to analyze and visualize data. But that’s not all – our program also includes a 3-month internship with us where you can showcase your Capstone Project.

2 Responses

This is a great and comprehensive article on hypothesis testing, covering everything from the basics to practical examples. I particularly appreciate the section on common misconceptions, as it’s important to understand what hypothesis testing can and cannot do. Overall, a valuable resource for anyone looking to understand this statistical technique.

Thanks, Ana Carol for your Kind words, Yes these topics are very important to know in Artificial intelligence.

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Genome Biol

The data-hypothesis relationship

Teppo felin.

1 Saïd Business School, University of Oxford, Oxford, UK

Jan Koenderink

2 Department of Physics, Delft University of Technology, Delft, The Netherlands

3 Department of Experimental Psychology, University of Leuven, Leuven, Belgium

Joachim I. Krueger

4 Department of Cognitive, Linguistic and Psychological Sciences, Brown University, Providence, USA

Denis Noble

5 Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, UK

George F.R. Ellis

6 Department of Mathematics, University of Cape Town, Cape Town, South Africa

Every conscious cognitive process will show itself to be steeped in theories; full of hypotheses. Rupert Riedl [ 1 ]

In a provocative editorial, Yanai and Lercher (henceforth Y&L) claim that “a hypothesis is a liability” [ 2 ]. They contend that having a hypothesis is costly because it causes scientists to miss hidden data and interesting phenomena. Y&L advocate “hypothesis-free” data exploration, which they argue can yield significant scientific discoveries.

We disagree. While we concur that a bad hypothesis is a liability, there is no such thing as hypothesis-free data exploration. Observation and data are always hypothesis- or theory-laden. Data is meaningless without some form of hypothesis or theory. Any exploration of data, however informal, is necessarily guided by some form of expectations. Even informal hunches or conjectures are types of proto-hypothesis. Furthermore, seemingly hypothesis-free statistical tools and computational techniques also contain latent hypotheses and theories about what is important—what might be interesting, worth measuring or paying attention to. Thus, while Y&L argue that a “hypothesis is a liability,” we argue that hypothesis-free observation is not possible (nor desirable) and that hypotheses in fact are the primary engine of scientific creativity and discovery.

The hidden gorilla

To illustrate their point about how a hypothesis is a liability, Y&L present their own version of the famous gorilla experiment [ 3 ]. In their experiment, subjects receive some made-up data featuring three variables: the BMI of individuals, the number of steps taken on a particular day, and their gender. One experimental group received three hypotheses to consider, while the other was “hypothesis-free.” Subjects in this latter group were simply asked to address the question “what do you conclude from the dataset?”

The “catch” of Y&L’s experiment was that a visual plot of the data showed a waving gorilla. And the key finding was that subjects in the hypothesis-free group were five times more likely to see the gorilla, compared with subjects in the hypothesis-focused group. Y&L concluded from this that hypotheses blind us to hidden patterns and insights in the data. Perhaps ironically, Y&L come to this conclusion based on their own hypothesis about the dangers of hypotheses.

But how exactly does missing the gorilla generalize to Y&L’s point about a hypothesis being a liability in scientific discovery? They argue that missing the gorilla is a problem, even though it is hard to see how finding an irrelevant gorilla mimics making a scientific insight. Now, we understand the gorilla is used as a metaphor for missing surprising or hidden things in science. But a meteorologist missing a cloud that looks like a gorilla is roughly equivalent to what Y&L are doing. A gorilla-shaped cloud has no scientific interest to the meteorologist, just as the gorilla-shaped data is irrelevant to Y&L’s context (the health data with three variables: BMI, steps taken and gender). Furthermore, the gorilla example does not generalize to scientific discovery because a gorilla is something that is universally recognized, while scientific discovery is essentially about finding new data, establishing new facts and relationships. New insights and scientific discoveries do not somehow “pop out” like the gorilla does once one plots the raw data. Hypotheses are needed. Thus, there is a mismatch between the experiment and what Y&L are claiming, on a number of levels.

Y&L import some of these problems from the original gorilla study [ 4 ]. The most serious concern is that various versions of the gorilla study can be seen as a form of attentional misdirection, similar to what is practiced by magicians. Experimental tasks are artificially constructed and designed to prove a specific hypothesis: that people are blind and miss large objects in their visual scenes. Experimenters first hide something in the visual scene, then distract their subjects with other tasks (whether counting basketball passes or asking them to analyze specific hypotheses), and then, voilà, reveal to them what they have missed. The problem is that—whether in science or in everyday life—an indefinite number of things remain undetected when we interact with data or visual scenes. It is not obvious what an apple falling means, without the right question, hypothesis, or theory. Visual scenes and data teem with possibilities, uses and meanings. Of course, the excitement generated by these studies comes from the fact that something so large and surprising—like a gorilla—goes undetected, even though it should be plainly obvious.

But there are deeper issues here. Reductionist forms of science assume that cues and data (somehow) jump out and tell us why they are relevant and important, based on the characteristics of the data itself (the physical properties of the world). In vision science, this assumption is based on research in psychophysics (and inverse optics and ideal observer theory) that focuses on salience as a function of cue or stimulus characteristics. From this perspective, cues and stimuli become data, information, and evidence due to their inherent nature [ 5 ].

To illustrate the problem with this, consider two stimulus or cue characteristics that are important to various versions of the gorilla study—and central to psychophysics and the cognitive sciences more generally—namely “size” and “surprisingness” [ 6 ]. The idea in psychophysics is that these characteristics should make cues salient. For example, researchers embedded an image of a gorilla in the CT scan images of patients’ lungs. They then asked expert radiologists to look for nodules as part of lung-cancer screening. Eighty-three percent of the radiologists missed the gorilla embedded in the image, despite the fact that the gorilla was 48 times the size of the nodules they were looking for [ 7 ].

But if radiologists or experimental subjects were asked to, say, “look for something unusual” or to “see if you can find the animal,” they would presumably find the gorilla. Thus, visual awareness or recognition has little to do with size or surprisingness. It has more to do with the question posed by the experimenter or the expectations of experimental subjects. In fact, experimental subjects themselves might suspect that the study actually is not about counting basketball passes or about analyzing health data or finding cancerous nodules in lungs. If subjects think that they are being tricked by experimenters—as is often the case—they might ignore the distracting tasks and priming questions and look for and find the gorilla. Note, again, that the a priori hypothesis of experimenters themselves is that people are blind, and so the experiments themselves are designed to prove this point. Alert subjects might suspect that they are being purposefully distracted and thus try to guess what they are meant to look for and find it.

The key point here is that the “transformation” of raw cues or data to information and evidence is not a straightforward process. It requires some form of hypothesis. Cues and data do not automatically tell us what they mean, whether or why they are relevant, or for which hypothesis they might provide evidence. Size is relevant in some situations, but not in others. Cues and data only become information and evidence in response to the questions and queries that we are asking.

Fishing expeditions require a net

One alternative to having a hypothesis, Y&L argue, is hypothesis- free exploration of data or what they call fishing expeditions. Of course, the idea of engaging in a fishing expedition—as Y&L recognize—has highly negative connotations, suggesting haphazard, unscientific, and perhaps even unethical practices. But they make a valid point: more exploratory and imaginative practices are important in science.

But fishing expeditions are hardly hypothesis-free. That is, fishing expeditions—to extend Y&L’s metaphor—require a net or some type of device for catching fish. Data and insights (just like fish) do not jump out and declare their relevance, meaning, or importance. As put by physical chemist Michael Polanyi, “things are not labelled ‘evidence’ in nature” [ 8 ]. The relevant data needs to be identified and lured in some fashion. Even the most exploratory process in science features choices and assumptions about what will count as data and evidence and what should be measured (and how). Any look at data—however preliminary it might be— necessarily represents some form of proto-hypothesis: a latent expectation, question, or even guess about what might be lurking, about what might potentially be interesting or relevant and how it might be caught.

In short, there’s no systematic way to extract and identify anything hidden without at least some rough idea of what one is looking for. The tools and devices scientists use are the net, sieve, or filter for capturing relevance and meaning. These nets come in vastly different materials and textures, sizes, types of weights, and anchors. Choices also need to be made about where to cast these nets. There are various ways to use and deploy them (trolling, longline, and so forth). Each choice implies a hypothesis. The choice of fishing net implies a hypothesis about what one is looking for and about what one might expect to catch and see as relevant [ 9 ].

Now, it might seem like we are stretching the definition of a hypothesis by including expectations, conjectures, and even the statistical and computational tools that are used to generate insights. But we think it is important to recognize that any tool—whether cognitive, computational, or statistical—functions like a net, as it already embodies implicit hypotheses about what matters and what does not. Perhaps these are not full-fledged, formal hypotheses in the sense that Y&L discuss. But they certainly are proto-hypotheses that direct awareness and attention toward what should be measured and what counts as data and evidence. A hypothesis is some form of expectation or question about what one is looking for and about what one expects to find. The identification and collection of data necessarily is of the same form, as one cannot collect all data about what is going on in the world at a specific time: flu patterns in China, weather patterns in the Pacific, sunspot cycles, the state of the New York stock exchange, earthquakes in Tahiti, and so on. Science is about making decisions about what subset of all this “stuff” should be focused on and included in the analysis.

Y&L specifically emphasize correlations and the generation of various statistical patterns as a way to make hypothesis-free discoveries in data. Correlations are one form of “net” for looking at data. But correlations are ubiquitous and their strength tells us little [ 10 ]. One needs a hypothesis to arbitrate between which correlation might be worth investigating and which not. The genome-wide associational studies have pointed this out. With the exception of the usual outliers (rare genetic diseases), the association levels are relatively small. More data may offer more stable statistical estimates, but it will not achieve the identification of causality required for a physiological explanation. On the contrary, the extremely low association data can be hiding substantial causality or perhaps more complex or interconnected, omnigenic factors are at play in the genome [ 11 ]. A causal hypothesis, tested rigorously with quantitative modeling, can reveal the potential pathways for understanding genetic variation, epigenetic factors, and disease or traits [ 12 ].

Science: bottom-up versus top-down

Y&L argue that scientific discoveries are “undiscoverable without data.” While this is correct in principle, Y&L mis-specify the data-hypothesis relationship by privileging the role the data to the detriment of hypothesis and theory. They ignore the temporal primacy of theory and hypothesis. A hypothesis tells us what data to look for. Data emerges and becomes evidence in response to a hypothesis. In physics, for example, the existence of gravitational waves had long been hypothesized. The hypothesis guided scientists to look for this data. This specifically led to the invention and construction of exquisitely sensitive devices to detect and measure gravitational radiation (e.g., LIGO and VIRGO observations). Eventually, in 2015, gravitational waves were discovered. The data emerged because of the conceptualization, design, and construction of relevant devices for measurement. The data was manifest due to the hypothesis rather than the other way around. And the data analysis itself is theory-based [ 13 ]: it depends on templates of waves expected from the gravitational coalescence of black holes or neutron stars.

Einstein aptly captured the relationship between hypotheses and data when noting that “whether you can observe a thing or not depends on the theory which you use. It is the theory which decides what can be observed.” Einstein’s point might be illustrated by the so-called DIKW hierarchy (Fig.  1 ) [ 14 ]. Currently popular data-first approaches assume that scientific understanding is built from the bottom-up. But to the contrary, many of the greatest insights have come “top-down,” where scientists start with theories and hypotheses that guide them to identify the right data and evidence. One of the most profound ways this happens is when scientists query fundamental assumptions that are taken for granted, such as that species are fixed for all time, or that simultaneity is independent of the state of motion. This questioning of axiomatic assumptions drives the creation of transformational theories (the theory of evolution, special relativity) and the subsequent collection of associated data that tests such profound reshaping of the foundations.

The DIKW “hierarchy” is often seen as “bottom-up.” But, as we argue, top-down mechanisms play a critical role in discovering data, relevance, and meaning

There certainly are significant reciprocal influences between these “levels” of the hierarchy. But Y&L’s central argument that a “hypothesis is a liability” simply does not recognize the profound, top-down influence played by hypotheses and theories in science, and how these enable the identification and generation of data.

Our concern is that starting at the bottom—as suggested by Y&L’s notion of hypothesis-free exploration of data—will inadvertently lead to an overly descriptive science: what Ernest Rutherford called “stamp collecting.” Charles Darwin anticipated this problem when he wrote to a friend:

It made me laugh to read of [Edwin Lankester’s] advice or rather regret that I had not published facts alone . How profoundly ignorant he must be of the very soul of observation. About 30 years ago there was much talk that Geologists ought only to observe and not theorise ; and I well remember someone saying, that at this rate a man might as well go into a gravel-pit and count the pebbles and describe their colours. How odd it is that everyone should not see that all observation must be for or against some view , if it is to be of any service [ 15 ].

Acknowledgements

TF, DN and GFRE  gratefully acknowledge University of Oxford's Foundations of Value and Values-initiative for providing a forum to discuss these types of interdisciplinary issues.

Authors’ contributions

TF wrote the initial draft of the manuscript. JK, JIK, DN and GFRE added many ideas, examples and further edits to subsequent iterations of the article. The authors read and approved the final manuscript.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Forgot password? New user? Sign up

Existing user? Log in

Hypothesis Testing

Already have an account? Log in here.

A hypothesis test is a statistical inference method used to test the significance of a proposed (hypothesized) relation between population statistics (parameters) and their corresponding sample estimators . In other words, hypothesis tests are used to determine if there is enough evidence in a sample to prove a hypothesis true for the entire population.

The test considers two hypotheses: the null hypothesis , which is a statement meant to be tested, usually something like "there is no effect" with the intention of proving this false, and the alternate hypothesis , which is the statement meant to stand after the test is performed. The two hypotheses must be mutually exclusive ; moreover, in most applications, the two are complementary (one being the negation of the other). The test works by comparing the \(p\)-value to the level of significance (a chosen target). If the \(p\)-value is less than or equal to the level of significance, then the null hypothesis is rejected.

When analyzing data, only samples of a certain size might be manageable as efficient computations. In some situations the error terms follow a continuous or infinite distribution, hence the use of samples to suggest accuracy of the chosen test statistics. The method of hypothesis testing gives an advantage over guessing what distribution or which parameters the data follows.

Definitions and Methodology

Hypothesis test and confidence intervals.

In statistical inference, properties (parameters) of a population are analyzed by sampling data sets. Given assumptions on the distribution, i.e. a statistical model of the data, certain hypotheses can be deduced from the known behavior of the model. These hypotheses must be tested against sampled data from the population.

The null hypothesis \((\)denoted \(H_0)\) is a statement that is assumed to be true. If the null hypothesis is rejected, then there is enough evidence (statistical significance) to accept the alternate hypothesis \((\)denoted \(H_1).\) Before doing any test for significance, both hypotheses must be clearly stated and non-conflictive, i.e. mutually exclusive, statements. Rejecting the null hypothesis, given that it is true, is called a type I error and it is denoted \(\alpha\), which is also its probability of occurrence. Failing to reject the null hypothesis, given that it is false, is called a type II error and it is denoted \(\beta\), which is also its probability of occurrence. Also, \(\alpha\) is known as the significance level , and \(1-\beta\) is known as the power of the test. \(H_0\) \(\textbf{is true}\)\(\hspace{15mm}\) \(H_0\) \(\textbf{is false}\) \(\textbf{Reject}\) \(H_0\)\(\hspace{10mm}\) Type I error Correct Decision \(\textbf{Reject}\) \(H_1\) Correct Decision Type II error The test statistic is the standardized value following the sampled data under the assumption that the null hypothesis is true, and a chosen particular test. These tests depend on the statistic to be studied and the assumed distribution it follows, e.g. the population mean following a normal distribution. The \(p\)-value is the probability of observing an extreme test statistic in the direction of the alternate hypothesis, given that the null hypothesis is true. The critical value is the value of the assumed distribution of the test statistic such that the probability of making a type I error is small.

Methodologies: Given an estimator \(\hat \theta\) of a population statistic \(\theta\), following a probability distribution \(P(T)\), computed from a sample \(\mathcal{S},\) and given a significance level \(\alpha\) and test statistic \(t^*,\) define \(H_0\) and \(H_1;\) compute the test statistic \(t^*.\) \(p\)-value Approach (most prevalent): Find the \(p\)-value using \(t^*\) (right-tailed). If the \(p\)-value is at most \(\alpha,\) reject \(H_0\). Otherwise, reject \(H_1\). Critical Value Approach: Find the critical value solving the equation \(P(T\geq t_\alpha)=\alpha\) (right-tailed). If \(t^*>t_\alpha\), reject \(H_0\). Otherwise, reject \(H_1\). Note: Failing to reject \(H_0\) only means inability to accept \(H_1\), and it does not mean to accept \(H_0\).

Assume a normally distributed population has recorded cholesterol levels with various statistics computed. From a sample of 100 subjects in the population, the sample mean was 214.12 mg/dL (milligrams per deciliter), with a sample standard deviation of 45.71 mg/dL. Perform a hypothesis test, with significance level 0.05, to test if there is enough evidence to conclude that the population mean is larger than 200 mg/dL. Hypothesis Test We will perform a hypothesis test using the \(p\)-value approach with significance level \(\alpha=0.05:\) Define \(H_0\): \(\mu=200\). Define \(H_1\): \(\mu>200\). Since our values are normally distributed, the test statistic is \(z^*=\frac{\bar X - \mu_0}{\frac{s}{\sqrt{n}}}=\frac{214.12 - 200}{\frac{45.71}{\sqrt{100}}}\approx 3.09\). Using a standard normal distribution, we find that our \(p\)-value is approximately \(0.001\). Since the \(p\)-value is at most \(\alpha=0.05,\) we reject \(H_0\). Therefore, we can conclude that the test shows sufficient evidence to support the claim that \(\mu\) is larger than \(200\) mg/dL.

If the sample size was smaller, the normal and \(t\)-distributions behave differently. Also, the question itself must be managed by a double-tail test instead.

Assume a population's cholesterol levels are recorded and various statistics are computed. From a sample of 25 subjects, the sample mean was 214.12 mg/dL (milligrams per deciliter), with a sample standard deviation of 45.71 mg/dL. Perform a hypothesis test, with significance level 0.05, to test if there is enough evidence to conclude that the population mean is not equal to 200 mg/dL. Hypothesis Test We will perform a hypothesis test using the \(p\)-value approach with significance level \(\alpha=0.05\) and the \(t\)-distribution with 24 degrees of freedom: Define \(H_0\): \(\mu=200\). Define \(H_1\): \(\mu\neq 200\). Using the \(t\)-distribution, the test statistic is \(t^*=\frac{\bar X - \mu_0}{\frac{s}{\sqrt{n}}}=\frac{214.12 - 200}{\frac{45.71}{\sqrt{25}}}\approx 1.54\). Using a \(t\)-distribution with 24 degrees of freedom, we find that our \(p\)-value is approximately \(2(0.068)=0.136\). We have multiplied by two since this is a two-tailed argument, i.e. the mean can be smaller than or larger than. Since the \(p\)-value is larger than \(\alpha=0.05,\) we fail to reject \(H_0\). Therefore, the test does not show sufficient evidence to support the claim that \(\mu\) is not equal to \(200\) mg/dL.

The complement of the rejection on a two-tailed hypothesis test (with significance level \(\alpha\)) for a population parameter \(\theta\) is equivalent to finding a confidence interval \((\)with confidence level \(1-\alpha)\) for the population parameter \(\theta\). If the assumption on the parameter \(\theta\) falls inside the confidence interval, then the test has failed to reject the null hypothesis \((\)with \(p\)-value greater than \(\alpha).\) Otherwise, if \(\theta\) does not fall in the confidence interval, then the null hypothesis is rejected in favor of the alternate \((\)with \(p\)-value at most \(\alpha).\)

• Statistics (Estimation)
• Normal Distribution
• Correlation
• Confidence Intervals

Problem Loading...

Note Loading...

Set Loading...

What is Hypothesis Testing in Statistics? Types and Examples

Varun Saharawat is a seasoned professional in the fields of SEO and content writing. With a profound knowledge of the intricate aspects of these disciplines, Varun has established himself as a valuable asset in the world of digital marketing and online content creation.

Hypothesis testing in statistics involves testing an assumption about a population parameter using sample data. Learners can download Hypothesis Testing PDF to get instant access to all information!

What exactly is hypothesis testing, and how does it work in statistics? Can I find practical examples and understand the different types from this blog?

Hypothesis Testing : Ever wonder how researchers determine if a new medicine actually works or if a new marketing campaign effectively drives sales? They use hypothesis testing! It is at the core of how scientific studies, business experiments and surveys determine if their results are statistically significant or just due to chance.

Hypothesis testing allows us to make evidence-based decisions by quantifying uncertainty and providing a structured process to make data-driven conclusions rather than guessing. In this post, we will discuss hypothesis testing types, examples, and processes!

Table of Contents

Hypothesis Testing

Hypothesis testing is a statistical method used to evaluate the validity of a hypothesis using sample data. It involves assessing whether observed data provide enough evidence to reject a specific hypothesis about a population parameter. 

Hypothesis Testing in Data Science

Hypothesis testing in data science is a statistical method used to evaluate two mutually exclusive population statements based on sample data. The primary goal is to determine which statement is more supported by the observed data.

Hypothesis testing assists in supporting the certainty of findings in research and data science projects. This statistical inference aids in making decisions about population parameters using sample data. For those who are looking to deepen their knowledge in data science and expand their skillset, we highly recommend checking out Master Generative AI: Data Science Course by Physics Wallah .

Also Read: What is Encapsulation Explain in Details

What is the Hypothesis Testing Procedure in Data Science?

The hypothesis testing procedure in data science involves a structured approach to evaluating hypotheses using statistical methods. Here’s a step-by-step breakdown of the typical procedure:

1) State the Hypotheses:

• Null Hypothesis (H0): This is the default assumption or a statement of no effect or difference. It represents what you aim to test against.
• Alternative Hypothesis (Ha): This is the opposite of the null hypothesis and represents what you want to prove.

2) Choose a Significance Level (α):

• Decide on a threshold (commonly 0.05) beyond which you will reject the null hypothesis. This is your significance level.

3) Select the Appropriate Test:

• Depending on your data type (e.g., continuous, categorical) and the nature of your research question, choose the appropriate statistical test (e.g., t-test, chi-square test, ANOVA, etc.).

4) Collect Data:

• Gather data from your sample or population, ensuring that it’s representative and sufficiently large (or as per your experimental design).

5)Compute the Test Statistic:

• Using your data and the chosen statistical test, compute the test statistic that summarizes the evidence against the null hypothesis.

6) Determine the Critical Value or P-value:

• Based on your significance level and the test statistic’s distribution, determine the critical value from a statistical table or compute the p-value.

7) Make a Decision:

• If the p-value is less than α: Reject the null hypothesis.
• If the p-value is greater than or equal to α: Fail to reject the null hypothesis.

8) Draw Conclusions:

• Based on your decision, draw conclusions about your research question or hypothesis. Remember, failing to reject the null hypothesis doesn’t prove it true; it merely suggests that you don’t have sufficient evidence to reject it.

9) Report Findings:

• Document your findings, including the test statistic, p-value, conclusion, and any other relevant details. Ensure clarity so that others can understand and potentially replicate your analysis.

Also Read: Binary Search Algorithm

How Hypothesis Testing Works?

Hypothesis testing is a fundamental concept in statistics that aids analysts in making informed decisions based on sample data about a larger population. The process involves setting up two contrasting hypotheses, the null hypothesis and the alternative hypothesis, and then using statistical methods to determine which hypothesis provides a more plausible explanation for the observed data.

The Core Principles:

• The Null Hypothesis (H0): This serves as the default assumption or status quo. Typically, it posits that there is no effect or no difference, often represented by an equality statement regarding population parameters. For instance, it might state that a new drug’s effect is no different from a placebo.
• The Alternative Hypothesis (H1 or Ha): This is the counter assumption or what researchers aim to prove. It’s the opposite of the null hypothesis, indicating that there is an effect, a change, or a difference in the population parameters. Using the drug example, the alternative hypothesis would suggest that the new drug has a different effect than the placebo.

Testing the Hypotheses:

Once these hypotheses are established, analysts gather data from a sample and conduct statistical tests. The objective is to determine whether the observed results are statistically significant enough to reject the null hypothesis in favor of the alternative.

Examples to Clarify the Concept:

• Null Hypothesis (H0): The sanitizer’s average efficacy is 95%.
• By conducting tests, if evidence suggests that the sanitizer’s efficacy is significantly less than 95%, we reject the null hypothesis.
• Null Hypothesis (H0): The coin is fair, meaning the probability of heads and tails is equal.
• Through experimental trials, if results consistently show a skewed outcome, indicating a significantly different probability for heads and tails, the null hypothesis might be rejected.

What are the 3 types of Hypothesis Test?

Hypothesis testing is a cornerstone in statistical analysis, providing a framework to evaluate the validity of assumptions or claims made about a population based on sample data. Within this framework, several specific tests are utilized based on the nature of the data and the question at hand. Here’s a closer look at the three fundamental types of hypothesis tests:

The z-test is a statistical method primarily employed when comparing means from two datasets, particularly when the population standard deviation is known. Its main objective is to ascertain if the means are statistically equivalent. 

A crucial prerequisite for the z-test is that the sample size should be relatively large, typically 30 data points or more. This test aids researchers and analysts in determining the significance of a relationship or discovery, especially in scenarios where the data’s characteristics align with the assumptions of the z-test.

The t-test is a versatile statistical tool used extensively in research and various fields to compare means between two groups. It’s particularly valuable when the population standard deviation is unknown or when dealing with smaller sample sizes. 

By evaluating the means of two groups, the t-test helps ascertain if a particular treatment, intervention, or variable significantly impacts the population under study. Its flexibility and robustness make it a go-to method in scenarios ranging from medical research to business analytics.

3. Chi-Square Test:

The Chi-Square test stands distinct from the previous tests, primarily focusing on categorical data rather than means. This statistical test is instrumental when analyzing categorical variables to determine if observed data aligns with expected outcomes as posited by the null hypothesis. 

By assessing the differences between observed and expected frequencies within categorical data, the Chi-Square test offers insights into whether discrepancies are statistically significant. Whether used in social sciences to evaluate survey responses or in quality control to assess product defects, the Chi-Square test remains pivotal for hypothesis testing in diverse scenarios.

Also Read: Python vs Java: Which is Best for Machine learning algorithm

Hypothesis Testing in Statistics

Hypothesis testing is a fundamental concept in statistics used to make decisions or inferences about a population based on a sample of data. The process involves setting up two competing hypotheses, the null hypothesis H 0​ and the alternative hypothesis H 1​. 

Through various statistical tests, such as the t-test, z-test, or Chi-square test, analysts evaluate sample data to determine whether there’s enough evidence to reject the null hypothesis in favor of the alternative. The aim is to draw conclusions about population parameters or to test theories, claims, or hypotheses.

Hypothesis Testing in Research

In research, hypothesis testing serves as a structured approach to validate or refute theories or claims. Researchers formulate a clear hypothesis based on existing literature or preliminary observations. They then collect data through experiments, surveys, or observational studies. 

Using statistical methods, researchers analyze this data to determine if there’s sufficient evidence to reject the null hypothesis. By doing so, they can draw meaningful conclusions, make predictions, or recommend actions based on empirical evidence rather than mere speculation.

Hypothesis Testing in R

R, a powerful programming language and environment for statistical computing and graphics, offers a wide array of functions and packages specifically designed for hypothesis testing. Here’s how hypothesis testing is conducted in R:

• Data Collection : Before conducting any test, you need to gather your data and ensure it’s appropriately structured in R.
• Choose the Right Test : Depending on your research question and data type, select the appropriate hypothesis test. For instance, use the t.test() function for a t-test or chisq.test() for a Chi-square test.
• Set Hypotheses : Define your null and alternative hypotheses. Using R’s syntax, you can specify these hypotheses and run the corresponding test.
• Execute the Test : Utilize built-in functions in R to perform the hypothesis test on your data. For instance, if you want to compare two means, you can use the t.test() function, providing the necessary arguments like the data vectors and type of t-test (one-sample, two-sample, paired, etc.).
• Interpret Results : Once the test is executed, R will provide output, including test statistics, p-values, and confidence intervals. Based on these results and a predetermined significance level (often 0.05), you can decide whether to reject the null hypothesis.
• Visualization : R’s graphical capabilities allow users to visualize data distributions, confidence intervals, or test statistics, aiding in the interpretation and presentation of results.

Hypothesis testing is an integral part of statistics and research, offering a systematic approach to validate hypotheses. Leveraging R’s capabilities, researchers and analysts can efficiently conduct and interpret various hypothesis tests, ensuring robust and reliable conclusions from their data.

Do Data Scientists do Hypothesis Testing?

Yes, data scientists frequently engage in hypothesis testing as part of their analytical toolkit. Hypothesis testing is a foundational statistical technique used to make data-driven decisions, validate assumptions, and draw conclusions from data. Here’s how data scientists utilize hypothesis testing:

• Validating Assumptions : Before diving into complex analyses or building predictive models, data scientists often need to verify certain assumptions about the data. Hypothesis testing provides a structured approach to test these assumptions, ensuring that subsequent analyses or models are valid.
• Feature Selection : In machine learning and predictive modeling, data scientists use hypothesis tests to determine which features (or variables) are most relevant or significant in predicting a particular outcome. By testing hypotheses related to feature importance or correlation, they can streamline the modeling process and enhance prediction accuracy.
• A/B Testing : A/B testing is a common technique in marketing, product development, and user experience design. Data scientists employ hypothesis testing to compare two versions (A and B) of a product, feature, or marketing strategy to determine which performs better in terms of a specified metric (e.g., conversion rate, user engagement).
• Research and Exploration : In exploratory data analysis (EDA) or when investigating specific research questions, data scientists formulate hypotheses to test certain relationships or patterns within the data. By conducting hypothesis tests, they can validate these relationships, uncover insights, and drive data-driven decision-making.
• Model Evaluation : After building machine learning or statistical models, data scientists use hypothesis testing to evaluate the model’s performance, assess its predictive power, or compare different models. For instance, hypothesis tests like the t-test or F-test can help determine if a new model significantly outperforms an existing one based on certain metrics.
• Business Decision-making : Beyond technical analyses, data scientists employ hypothesis testing to support business decisions. Whether it’s evaluating the effectiveness of a marketing campaign, assessing customer preferences, or optimizing operational processes, hypothesis testing provides a rigorous framework to validate assumptions and guide strategic initiatives.

Hypothesis Testing Examples and Solutions

Let’s delve into some common examples of hypothesis testing and provide solutions or interpretations for each scenario.

Example: Testing the Mean

Scenario : A coffee shop owner believes that the average waiting time for customers during peak hours is 5 minutes. To test this, the owner takes a random sample of 30 customer waiting times and wants to determine if the average waiting time is indeed 5 minutes.

Hypotheses :

• H 0​ (Null Hypothesis): 5 μ =5 minutes (The average waiting time is 5 minutes)
• H 1​ (Alternative Hypothesis): 5 μ =5 minutes (The average waiting time is not 5 minutes)

Solution : Using a t-test (assuming population variance is unknown), calculate the t-statistic based on the sample mean, sample standard deviation, and sample size. Then, determine the p-value and compare it with a significance level (e.g., 0.05) to decide whether to reject the null hypothesis.

Example: A/B Testing in Marketing

Scenario : An e-commerce company wants to determine if changing the color of a “Buy Now” button from blue to green increases the conversion rate.

• H 0​: Changing the button color does not affect the conversion rate.
• H 1​: Changing the button color affects the conversion rate.

Solution : Split website visitors into two groups: one sees the blue button (control group), and the other sees the green button (test group). Track the conversion rates for both groups over a specified period. Then, use a chi-square test or z-test (for large sample sizes) to determine if there’s a statistically significant difference in conversion rates between the two groups.

Hypothesis Testing Formula

The formula for hypothesis testing typically depends on the type of test (e.g., z-test, t-test, chi-square test) and the nature of the data (e.g., mean, proportion, variance). Below are the basic formulas for some common hypothesis tests:

Z-Test for Population Mean :

Z=(σ/n​)(xˉ−μ0​)​

• ˉ x ˉ = Sample mean
• 0 μ 0​ = Population mean under the null hypothesis
• σ = Population standard deviation
• n = Sample size

T-Test for Population Mean :

t= (s/ n ​ ) ( x ˉ −μ 0 ​ ) ​ 

s = Sample standard deviation 

Chi-Square Test for Goodness of Fit :

χ2=∑Ei​(Oi​−Ei​)2​

• Oi ​ = Observed frequency
• Ei ​ = Expected frequency

Also Read: Full Form of OOPS

Hypothesis Testing Calculator

While you can perform hypothesis testing manually using the above formulas and statistical tables, many online tools and software packages simplify this process. Here’s how you might use a calculator or software:

• Z-Test and T-Test Calculators : These tools typically require you to input sample statistics (like sample mean, population mean, standard deviation, and sample size). Once you input these values, the calculator will provide you with the test statistic (Z or t) and a p-value.
• Chi-Square Calculator : For chi-square tests, you’d input observed and expected frequencies for different categories or groups. The calculator then computes the chi-square statistic and provides a p-value.
• Software Packages (e.g., R, Python with libraries like scipy, or statistical software like SPSS) : These platforms offer more comprehensive tools for hypothesis testing. You can run various tests, get detailed outputs, and even perform advanced analyses, including regression models, ANOVA, and more.

When using any calculator or software, always ensure you understand the underlying assumptions of the test, interpret the results correctly, and consider the broader context of your research or analysis.

Hypothesis Testing FAQs

What are the key components of a hypothesis test.

The key components include: Null Hypothesis (H0): A statement of no effect or no difference. Alternative Hypothesis (H1 or Ha): A statement that contradicts the null hypothesis. Test Statistic: A value computed from the sample data to test the null hypothesis. Significance Level (α): The threshold for rejecting the null hypothesis. P-value: The probability of observing the given data, assuming the null hypothesis is true.

What is the significance level in hypothesis testing?

The significance level (often denoted as α) is the probability threshold used to determine whether to reject the null hypothesis. Commonly used values for α include 0.05, 0.01, and 0.10, representing a 5%, 1%, or 10% chance of rejecting the null hypothesis when it's actually true.

How do I choose between a one-tailed and two-tailed test?

The choice between one-tailed and two-tailed tests depends on your research question and hypothesis. Use a one-tailed test when you're specifically interested in one direction of an effect (e.g., greater than or less than). Use a two-tailed test when you want to determine if there's a significant difference in either direction.

What is a p-value, and how is it interpreted?

The p-value is a probability value that helps determine the strength of evidence against the null hypothesis. A low p-value (typically ≤ 0.05) suggests that the observed data is inconsistent with the null hypothesis, leading to its rejection. Conversely, a high p-value suggests that the data is consistent with the null hypothesis, leading to no rejection.

Can hypothesis testing prove a hypothesis true?

No, hypothesis testing cannot prove a hypothesis true. Instead, it helps assess the likelihood of observing a given set of data under the assumption that the null hypothesis is true. Based on this assessment, you either reject or fail to reject the null hypothesis.

• Top 10 Tech Skills to Master in 2024

I have compiled a list of in-demand top 10 Tech Skills to master in 2024 to help you navigate the…

• Top 30 Excel Formulas And Functions You Should Know

Microsoft Excel is the most common instrument for working with data and their structures. A handful of people probably haven’t…

• Best Web Designing: Top 10 Website Designs to Inspire You in 2024

Great web design is essential when it comes to any online business presence. Find best web designing options, ideas, and…

Related Articles

• PW Skills Vishwas Diwas: List of Affordable Courses in Offer
• PW Skills Vishwas Diwas: List of Premium Courses Available on Discount Offers 2024
• Top 50 Manual Testing Interview Questions and Answers
• What is AWS DevOps? | Architecture, Tools, and Benefits in 2024
• SQL Interview Questions CHEAT SHEET (2024)
• Top 30+ Angular Interview Questions and Answers for 2024
• Top 10 Online Computer Programming Courses To Enroll In 2024

• Teesside University Student & Library Services
• Learning Hub Group

Quantitative data collection and analysis

• Testing hypotheses
• Quantitative data collection
• Averages and percentiles
• Measures of Spread or Dispersion
• Samples and population
• Statistical tests - parametric
• Statistical tests - non-parametric
• Probability
• Reliability and Validity
• Analysing relationships
• Useful Books

Testing Hypotheses

• What is a hypothesis?
• Significance testing
• One-tailed or two-tailed?
• Degrees of freedom

A hypothesis is a statement that we are trying to prove or disprove. It is used to express the relationship between variables  and whether this relationship is significant. It is specific and offers a prediction on the results of your research question.

Your research question  will lead you to developing a hypothesis, this is why your research question needs to be specific and clear.

The hypothesis will then guide you to the most appropriate techniques you should use to answer the question. They reflect the literature and theories on which you basing them. They need to be testable (i.e. measurable and practical).

Null hypothesis  (H 0 ) is the proposition that there will not be a relationship between the variables you are looking at. i.e. any differences are due to chance). They always refer to the population. (Usually we don't believe this to be true.)

e.g. There is  no difference in instances of illegal drug use by teenagers who are members of a gang and those who are not..

Alternative hypothesis  (H A ) or ( H 1 ):  this is sometimes called the research hypothesis or experimental hypothesis. It is the proposition that there will be a relationship. It is a statement of inequality between the variables you are interested in. They always refer to the sample. It is usually a declaration rather than a question and is clear, to the point and specific.

e.g. The instances of illegal drug use of teenagers who are members of a gang  is different than the instances of illegal drug use of teenagers who are not gang members.

A non-directional research hypothesis - reflects an expected difference between groups but does not specify the direction of this difference (see two-tailed test).

A directional research hypothesis - reflects an expected difference between groups but does specify the direction of this difference. (see one-tailed test)

e.g. The instances of illegal drug use by teenagers who are members of a gang will be higher t han the instances of illegal drug use of teenagers who are not gang members.

Then the process of testing is to ascertain which hypothesis to believe. 

It is usually easier to prove something as untrue rather than true, so looking at the null hypothesis is the usual starting point.

The process of examining the null hypothesis in light of evidence from the sample is called significance testing . It is a way of establishing a range of values in which we can establish whether the null hypothesis is true or false.

The debate over hypothesis testing

There has been discussion over whether the scientific method employed in traditional hypothesis testing is appropriate.  

See below for some articles that discuss this:

• Gill, J. (1999) 'The insignificance of null hypothesis testing',  Politics Research Quarterly , 52(3), pp. 647-674 .
• Wainer, H. and Robinson, D.H. (2003) 'Shaping up the practice of null hypothesis significance testing',  Educational Researcher, 32(7), pp.22-30 .
• Ferguson, C.J. and Heener, M. (2012) ' A vast graveyard of undead theories: publication bias and psychological science's aversion to the null' ,  Perspectives on Psychological Science, 7(6), pp.555-561 .

Taken from: Salkind, N.J. (2017)  Statistics for people who (think they) hate statistics. 6th edn. London: SAGE pp. 144-145.

• Null hypothesis - a simple introduction (SPSS)

A significance level defines the level when your sample evidence contradicts your null hypothesis so that your can then reject it. It is the probability of rejecting the null hypothesis when it is really true.

e.g. a significance level of 0.05 indicates that there is a 5% (or 1 in 20) risk of deciding that there is an effect when in fact there is none.

The lower the significance level that you set,  then the evidence from the sample has to be stronger to be able to reject the null hypothesis.

N.B.  - it is important that you set the significance level before you carry out your study and analysis.

Using Confidence Intervals

I t is possible to test the significance of your null hypothesis using Confidence Interval (see under samples and populations tab).

- if the range lies outside our predicted null hypothesis value we can reject it and accept the alternative hypothesis  

The test statistic

This is another commonly used statistic

• Write down your null and alternative hypothesis
• Find the sample statistic (e.g.the mean of your sample)
• Calculate the test statistic Z score (see under Measures of spread or dispersion and Statistical tests - parametric). In this case the sample mean is compared to the population mean (assumed from the null hypothesis) and the standard error (see under Samples and population) is used rather than the standard deviation.
• Compare the test statistic with the critical values (e.g. plus or minus 1.96 for 5% significance)
• Draw a conclusion about the hypotheses - does the calculated z value lies in this critical range i.e. above 1.96 or below -1.96? If it does we can reject the null hypothesis. This would indicate that the results are significant (or an effect has been detected) - which means that if there were no difference in the population then getting a result that you have observed would be highly unlikely therefore you can reject the null hypothesis.

Type I error  - this is the chance of wrongly rejecting the null hypothesis even though it is actually true, e.g. by using a 5% p  level you would expect the null hypothesis to be rejected about 5% of the time when the null hypothesis is true. You could set a more stringent p  level such as 1% (or 1 in 100) to be more certain of not seeing a Type I error. This, however, makes more likely another type of error (Type II) occurring.

Type II error  - this is where there is an effect, but the  p  value you obtain is non-significant hence you don’t detect this effect.

• Statistical significance - what does it really mean?
• Statistical tables

One-tailed tests - where we know in which direction (e.g. larger or smaller) the difference between sample and population will be. It is a directional hypothesis.

Two-tailed tests - where we are looking at whether there is a difference between sample and population. This difference could be larger or smaller. This is a non-directional hypothesis.

If the difference is in the direction you have predicted (i.e. a one-tailed test) it is easier to get a significant result. Though there are arguments against using a one-tailed test (Wright and London, 2009, p. 98-99)*

*Wright, D. B. & London, K. (2009)  First (and second) steps in statistics . 2nd edn. London: SAGE.

N.B. - think of the ‘tails’ as the regions at the far-end of a normal distribution. For a two-tailed test with significance level of 0.05% then 0.025% of the values would be at one end of the distribution and the other 0.025% would be at the other end of the distribution. It is the values in these ‘critical’ extreme regions where we can think about rejecting the null hypothesis and claim that there has been an effect.

Degrees of freedom ( df)  is a rather difficult mathematical concept, but is needed to calculate the signifcance of certain statistical tests, such as the t-test, ANOVA and Chi-squared test.

It is broadly defined as the number of "observations" (pieces of information) in the data that are free to vary when estimating statistical parameters. (Taken from Minitab Blog ).

The higher the degrees of freedom are the more powerful and precise your estimates of the parameter (population) will be.

Typically, for a 1-sample t-test it is considered as the number of values in your sample minus 1.

For chi-squared tests with a table of rows and columns the rule is:

(Number of rows minus 1) times (number of columns minus 1)

Any accessible example to illustrate the principle of degrees of freedom using chocolates.

• You have seven chocolates in a box, each being a different type, e.g. truffle, coffee cream, caramel cluster, fudge, strawberry dream, hazelnut whirl, toffee. 
• You are being good and intend to eat only one chocolate each day of the week.
• On the first day, you can choose to eat any one of the 7 chocolate types  - you have a choice from all 7.
• On the second day, you can choose from the 6 remaining chocolates, on day 3 you can choose from 5 chocolates, and so on.
• On the sixth day you have a choice of the remaining 2 chocolates you haven't ate that week.
• However on the seventh day - you haven't really got any choice of chocolate - it has got to be the one you have left in your box.
• You had 7-1 = 6 days of “chocolate” freedom—in which the chocolate you ate could vary!
• << Previous: Samples and population
• Next: Statistical tests - parametric >>
• Last Updated: Jan 9, 2024 11:01 AM
• URL: https://libguides.tees.ac.uk/quantitative

User Preferences

Content preview.

Arcu felis bibendum ut tristique et egestas quis:

• Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris
• Duis aute irure dolor in reprehenderit in voluptate
• Excepteur sint occaecat cupidatat non proident

Keyboard Shortcuts

6a.1 - introduction to hypothesis testing, basic terms section  .

The first step in hypothesis testing is to set up two competing hypotheses. The hypotheses are the most important aspect. If the hypotheses are incorrect, your conclusion will also be incorrect.

The two hypotheses are named the null hypothesis and the alternative hypothesis.

The goal of hypothesis testing is to see if there is enough evidence against the null hypothesis. In other words, to see if there is enough evidence to reject the null hypothesis. If there is not enough evidence, then we fail to reject the null hypothesis.

Consider the following example where we set up these hypotheses.

Example 6-1 Section  

A man, Mr. Orangejuice, goes to trial and is tried for the murder of his ex-wife. He is either guilty or innocent. Set up the null and alternative hypotheses for this example.

Putting this in a hypothesis testing framework, the hypotheses being tested are:

• The man is guilty
• The man is innocent

Let's set up the null and alternative hypotheses.

\(H_0\colon \) Mr. Orangejuice is innocent

\(H_a\colon \) Mr. Orangejuice is guilty

Remember that we assume the null hypothesis is true and try to see if we have evidence against the null. Therefore, it makes sense in this example to assume the man is innocent and test to see if there is evidence that he is guilty.

The Logic of Hypothesis Testing Section  

We want to know the answer to a research question. We determine our null and alternative hypotheses. Now it is time to make a decision.

The decision is either going to be...

• reject the null hypothesis or...
• fail to reject the null hypothesis.

Consider the following table. The table shows the decision/conclusion of the hypothesis test and the unknown "reality", or truth. We do not know if the null is true or if it is false. If the null is false and we reject it, then we made the correct decision. If the null hypothesis is true and we fail to reject it, then we made the correct decision.

So what happens when we do not make the correct decision?

When doing hypothesis testing, two types of mistakes may be made and we call them Type I error and Type II error. If we reject the null hypothesis when it is true, then we made a type I error. If the null hypothesis is false and we failed to reject it, we made another error called a Type II error.

Types of errors

The “reality”, or truth, about the null hypothesis is unknown and therefore we do not know if we have made the correct decision or if we committed an error. We can, however, define the likelihood of these events.

\(\alpha\) and \(\beta\) are probabilities of committing an error so we want these values to be low. However, we cannot decrease both. As \(\alpha\) decreases, \(\beta\) increases.

Example 6-1 Cont'd... Section  

A man, Mr. Orangejuice, goes to trial and is tried for the murder of his ex-wife. He is either guilty or not guilty. We found before that...

• \( H_0\colon \) Mr. Orangejuice is innocent
• \( H_a\colon \) Mr. Orangejuice is guilty

Interpret Type I error, \(\alpha \), Type II error, \(\beta \).

As you can see here, the Type I error (putting an innocent man in jail) is the more serious error. Ethically, it is more serious to put an innocent man in jail than to let a guilty man go free. So to minimize the probability of a type I error we would choose a smaller significance level.

Try it! Section  

An inspector has to choose between certifying a building as safe or saying that the building is not safe. There are two hypotheses:

• Building is safe
• Building is not safe

Set up the null and alternative hypotheses. Interpret Type I and Type II error.

\( H_0\colon\) Building is not safe vs \(H_a\colon \) Building is safe

Power and \(\beta \) are complements of each other. Therefore, they have an inverse relationship, i.e. as one increases, the other decreases.

• school Campus Bookshelves
• menu_book Bookshelves
• perm_media Learning Objects
• login Login
• how_to_reg Request Instructor Account
• hub Instructor Commons
• Download Page (PDF)
• Download Full Book (PDF)
• Periodic Table
• Physics Constants
• Scientific Calculator
• Reference & Cite
• Tools expand_more
• Readability

selected template will load here

This action is not available.

9.1: Introduction to Hypothesis Testing

• Last updated
• Save as PDF
• Page ID 10211

• Kyle Siegrist
• University of Alabama in Huntsville via Random Services

Basic Theory

Preliminaries.

As usual, our starting point is a random experiment with an underlying sample space and a probability measure \(\P\). In the basic statistical model, we have an observable random variable \(\bs{X}\) taking values in a set \(S\). In general, \(\bs{X}\) can have quite a complicated structure. For example, if the experiment is to sample \(n\) objects from a population and record various measurements of interest, then \[ \bs{X} = (X_1, X_2, \ldots, X_n) \] where \(X_i\) is the vector of measurements for the \(i\)th object. The most important special case occurs when \((X_1, X_2, \ldots, X_n)\) are independent and identically distributed. In this case, we have a random sample of size \(n\) from the common distribution.

The purpose of this section is to define and discuss the basic concepts of statistical hypothesis testing . Collectively, these concepts are sometimes referred to as the Neyman-Pearson framework, in honor of Jerzy Neyman and Egon Pearson, who first formalized them.

A statistical hypothesis is a statement about the distribution of \(\bs{X}\). Equivalently, a statistical hypothesis specifies a set of possible distributions of \(\bs{X}\): the set of distributions for which the statement is true. A hypothesis that specifies a single distribution for \(\bs{X}\) is called simple ; a hypothesis that specifies more than one distribution for \(\bs{X}\) is called composite .

In hypothesis testing , the goal is to see if there is sufficient statistical evidence to reject a presumed null hypothesis in favor of a conjectured alternative hypothesis . The null hypothesis is usually denoted \(H_0\) while the alternative hypothesis is usually denoted \(H_1\).

An hypothesis test is a statistical decision ; the conclusion will either be to reject the null hypothesis in favor of the alternative, or to fail to reject the null hypothesis. The decision that we make must, of course, be based on the observed value \(\bs{x}\) of the data vector \(\bs{X}\). Thus, we will find an appropriate subset \(R\) of the sample space \(S\) and reject \(H_0\) if and only if \(\bs{x} \in R\). The set \(R\) is known as the rejection region or the critical region . Note the asymmetry between the null and alternative hypotheses. This asymmetry is due to the fact that we assume the null hypothesis, in a sense, and then see if there is sufficient evidence in \(\bs{x}\) to overturn this assumption in favor of the alternative.

An hypothesis test is a statistical analogy to proof by contradiction, in a sense. Suppose for a moment that \(H_1\) is a statement in a mathematical theory and that \(H_0\) is its negation. One way that we can prove \(H_1\) is to assume \(H_0\) and work our way logically to a contradiction. In an hypothesis test, we don't prove anything of course, but there are similarities. We assume \(H_0\) and then see if the data \(\bs{x}\) are sufficiently at odds with that assumption that we feel justified in rejecting \(H_0\) in favor of \(H_1\).

Often, the critical region is defined in terms of a statistic \(w(\bs{X})\), known as a test statistic , where \(w\) is a function from \(S\) into another set \(T\). We find an appropriate rejection region \(R_T \subseteq T\) and reject \(H_0\) when the observed value \(w(\bs{x}) \in R_T\). Thus, the rejection region in \(S\) is then \(R = w^{-1}(R_T) = \left\{\bs{x} \in S: w(\bs{x}) \in R_T\right\}\). As usual, the use of a statistic often allows significant data reduction when the dimension of the test statistic is much smaller than the dimension of the data vector.

The ultimate decision may be correct or may be in error. There are two types of errors, depending on which of the hypotheses is actually true.

Types of errors:

• A type 1 error is rejecting the null hypothesis \(H_0\) when \(H_0\) is true.
• A type 2 error is failing to reject the null hypothesis \(H_0\) when the alternative hypothesis \(H_1\) is true.

Similarly, there are two ways to make a correct decision: we could reject \(H_0\) when \(H_1\) is true or we could fail to reject \(H_0\) when \(H_0\) is true. The possibilities are summarized in the following table:

Of course, when we observe \(\bs{X} = \bs{x}\) and make our decision, either we will have made the correct decision or we will have committed an error, and usually we will never know which of these events has occurred. Prior to gathering the data, however, we can consider the probabilities of the various errors.

If \(H_0\) is true (that is, the distribution of \(\bs{X}\) is specified by \(H_0\)), then \(\P(\bs{X} \in R)\) is the probability of a type 1 error for this distribution. If \(H_0\) is composite, then \(H_0\) specifies a variety of different distributions for \(\bs{X}\) and thus there is a set of type 1 error probabilities.

The maximum probability of a type 1 error, over the set of distributions specified by \( H_0 \), is the significance level of the test or the size of the critical region.

The significance level is often denoted by \(\alpha\). Usually, the rejection region is constructed so that the significance level is a prescribed, small value (typically 0.1, 0.05, 0.01).

If \(H_1\) is true (that is, the distribution of \(\bs{X}\) is specified by \(H_1\)), then \(\P(\bs{X} \notin R)\) is the probability of a type 2 error for this distribution. Again, if \(H_1\) is composite then \(H_1\) specifies a variety of different distributions for \(\bs{X}\), and thus there will be a set of type 2 error probabilities. Generally, there is a tradeoff between the type 1 and type 2 error probabilities. If we reduce the probability of a type 1 error, by making the rejection region \(R\) smaller, we necessarily increase the probability of a type 2 error because the complementary region \(S \setminus R\) is larger.

The extreme cases can give us some insight. First consider the decision rule in which we never reject \(H_0\), regardless of the evidence \(\bs{x}\). This corresponds to the rejection region \(R = \emptyset\). A type 1 error is impossible, so the significance level is 0. On the other hand, the probability of a type 2 error is 1 for any distribution defined by \(H_1\). At the other extreme, consider the decision rule in which we always rejects \(H_0\) regardless of the evidence \(\bs{x}\). This corresponds to the rejection region \(R = S\). A type 2 error is impossible, but now the probability of a type 1 error is 1 for any distribution defined by \(H_0\). In between these two worthless tests are meaningful tests that take the evidence \(\bs{x}\) into account.

If \(H_1\) is true, so that the distribution of \(\bs{X}\) is specified by \(H_1\), then \(\P(\bs{X} \in R)\), the probability of rejecting \(H_0\) is the power of the test for that distribution.

Thus the power of the test for a distribution specified by \( H_1 \) is the probability of making the correct decision.

Suppose that we have two tests, corresponding to rejection regions \(R_1\) and \(R_2\), respectively, each having significance level \(\alpha\). The test with region \(R_1\) is uniformly more powerful than the test with region \(R_2\) if \[ \P(\bs{X} \in R_1) \ge \P(\bs{X} \in R_2) \text{ for every distribution of } \bs{X} \text{ specified by } H_1 \]

Naturally, in this case, we would prefer the first test. Often, however, two tests will not be uniformly ordered; one test will be more powerful for some distributions specified by \(H_1\) while the other test will be more powerful for other distributions specified by \(H_1\).

If a test has significance level \(\alpha\) and is uniformly more powerful than any other test with significance level \(\alpha\), then the test is said to be a uniformly most powerful test at level \(\alpha\).

Clearly a uniformly most powerful test is the best we can do.

\(P\)-value

In most cases, we have a general procedure that allows us to construct a test (that is, a rejection region \(R_\alpha\)) for any given significance level \(\alpha \in (0, 1)\). Typically, \(R_\alpha\) decreases (in the subset sense) as \(\alpha\) decreases.

The \(P\)-value of the observed value \(\bs{x}\) of \(\bs{X}\), denoted \(P(\bs{x})\), is defined to be the smallest \(\alpha\) for which \(\bs{x} \in R_\alpha\); that is, the smallest significance level for which \(H_0\) is rejected, given \(\bs{X} = \bs{x}\).

Knowing \(P(\bs{x})\) allows us to test \(H_0\) at any significance level for the given data \(\bs{x}\): If \(P(\bs{x}) \le \alpha\) then we would reject \(H_0\) at significance level \(\alpha\); if \(P(\bs{x}) \gt \alpha\) then we fail to reject \(H_0\) at significance level \(\alpha\). Note that \(P(\bs{X})\) is a statistic . Informally, \(P(\bs{x})\) can often be thought of as the probability of an outcome as or more extreme than the observed value \(\bs{x}\), where extreme is interpreted relative to the null hypothesis \(H_0\).

Analogy with Justice Systems

There is a helpful analogy between statistical hypothesis testing and the criminal justice system in the US and various other countries. Consider a person charged with a crime. The presumed null hypothesis is that the person is innocent of the crime; the conjectured alternative hypothesis is that the person is guilty of the crime. The test of the hypotheses is a trial with evidence presented by both sides playing the role of the data. After considering the evidence, the jury delivers the decision as either not guilty or guilty . Note that innocent is not a possible verdict of the jury, because it is not the point of the trial to prove the person innocent. Rather, the point of the trial is to see whether there is sufficient evidence to overturn the null hypothesis that the person is innocent in favor of the alternative hypothesis of that the person is guilty. A type 1 error is convicting a person who is innocent; a type 2 error is acquitting a person who is guilty. Generally, a type 1 error is considered the more serious of the two possible errors, so in an attempt to hold the chance of a type 1 error to a very low level, the standard for conviction in serious criminal cases is beyond a reasonable doubt .

Tests of an Unknown Parameter

Hypothesis testing is a very general concept, but an important special class occurs when the distribution of the data variable \(\bs{X}\) depends on a parameter \(\theta\) taking values in a parameter space \(\Theta\). The parameter may be vector-valued, so that \(\bs{\theta} = (\theta_1, \theta_2, \ldots, \theta_n)\) and \(\Theta \subseteq \R^k\) for some \(k \in \N_+\). The hypotheses generally take the form \[ H_0: \theta \in \Theta_0 \text{ versus } H_1: \theta \notin \Theta_0 \] where \(\Theta_0\) is a prescribed subset of the parameter space \(\Theta\). In this setting, the probabilities of making an error or a correct decision depend on the true value of \(\theta\). If \(R\) is the rejection region, then the power function \( Q \) is given by \[ Q(\theta) = \P_\theta(\bs{X} \in R), \quad \theta \in \Theta \] The power function gives a lot of information about the test.

The power function satisfies the following properties:

• \(Q(\theta)\) is the probability of a type 1 error when \(\theta \in \Theta_0\).
• \(\max\left\{Q(\theta): \theta \in \Theta_0\right\}\) is the significance level of the test.
• \(1 - Q(\theta)\) is the probability of a type 2 error when \(\theta \notin \Theta_0\).
• \(Q(\theta)\) is the power of the test when \(\theta \notin \Theta_0\).

If we have two tests, we can compare them by means of their power functions.

Suppose that we have two tests, corresponding to rejection regions \(R_1\) and \(R_2\), respectively, each having significance level \(\alpha\). The test with rejection region \(R_1\) is uniformly more powerful than the test with rejection region \(R_2\) if \( Q_1(\theta) \ge Q_2(\theta)\) for all \( \theta \notin \Theta_0 \).

Most hypothesis tests of an unknown real parameter \(\theta\) fall into three special cases:

Suppose that \( \theta \) is a real parameter and \( \theta_0 \in \Theta \) a specified value. The tests below are respectively the two-sided test , the left-tailed test , and the right-tailed test .

• \(H_0: \theta = \theta_0\) versus \(H_1: \theta \ne \theta_0\)
• \(H_0: \theta \ge \theta_0\) versus \(H_1: \theta \lt \theta_0\)
• \(H_0: \theta \le \theta_0\) versus \(H_1: \theta \gt \theta_0\)

Thus the tests are named after the conjectured alternative. Of course, there may be other unknown parameters besides \(\theta\) (known as nuisance parameters ).

Equivalence Between Hypothesis Test and Confidence Sets

There is an equivalence between hypothesis tests and confidence sets for a parameter \(\theta\).

Suppose that \(C(\bs{x})\) is a \(1 - \alpha\) level confidence set for \(\theta\). The following test has significance level \(\alpha\) for the hypothesis \( H_0: \theta = \theta_0 \) versus \( H_1: \theta \ne \theta_0 \): Reject \(H_0\) if and only if \(\theta_0 \notin C(\bs{x})\)

By definition, \(\P[\theta \in C(\bs{X})] = 1 - \alpha\). Hence if \(H_0\) is true so that \(\theta = \theta_0\), then the probability of a type 1 error is \(P[\theta \notin C(\bs{X})] = \alpha\).

Equivalently, we fail to reject \(H_0\) at significance level \(\alpha\) if and only if \(\theta_0\) is in the corresponding \(1 - \alpha\) level confidence set. In particular, this equivalence applies to interval estimates of a real parameter \(\theta\) and the common tests for \(\theta\) given above .

In each case below, the confidence interval has confidence level \(1 - \alpha\) and the test has significance level \(\alpha\).

• Suppose that \(\left[L(\bs{X}, U(\bs{X})\right]\) is a two-sided confidence interval for \(\theta\). Reject \(H_0: \theta = \theta_0\) versus \(H_1: \theta \ne \theta_0\) if and only if \(\theta_0 \lt L(\bs{X})\) or \(\theta_0 \gt U(\bs{X})\).
• Suppose that \(L(\bs{X})\) is a confidence lower bound for \(\theta\). Reject \(H_0: \theta \le \theta_0\) versus \(H_1: \theta \gt \theta_0\) if and only if \(\theta_0 \lt L(\bs{X})\).
• Suppose that \(U(\bs{X})\) is a confidence upper bound for \(\theta\). Reject \(H_0: \theta \ge \theta_0\) versus \(H_1: \theta \lt \theta_0\) if and only if \(\theta_0 \gt U(\bs{X})\).

Pivot Variables and Test Statistics

Recall that confidence sets of an unknown parameter \(\theta\) are often constructed through a pivot variable , that is, a random variable \(W(\bs{X}, \theta)\) that depends on the data vector \(\bs{X}\) and the parameter \(\theta\), but whose distribution does not depend on \(\theta\) and is known. In this case, a natural test statistic for the basic tests given above is \(W(\bs{X}, \theta_0)\).

Fundamentals of Statistics

2. what is hypothesis, 2.2. what is hypothesis testing.

Module 8: Inference for One Proportion

Hypothesis testing (2 of 5), learning outcomes.

• Recognize the logic behind a hypothesis test and how it relates to the P-value.

In this section, our focus is hypothesis testing, which is part of inference. On the previous page, we practiced stating null and alternative hypotheses from a research question. Forming the hypotheses is the first step in a hypothesis test. Here are the general steps in the process of hypothesis testing. We will see that hypothesis testing is related to the thinking we did in Linking Probability to Statistical Inference .

Step 1: Determine the hypotheses.

The hypotheses come from the research question.

Step 2: Collect the data.

Ideally, we select a random sample from the population. The data comes from this sample. We calculate a statistic (a mean or a proportion) to summarize the data.

Step 3: Assess the evidence.

Assume that the null hypothesis is true. Could the data come from the population described by the null hypothesis? Use simulation or a mathematical model to examine the results from random samples selected from the population described by the null hypothesis. Figure out if results similar to the data are likely or unlikely. Note that the wording “likely or unlikely” implies that this step requires some kind of probability calculation.

Step 4: State a conclusion.

We use what we find in the previous step to make a decision. This step requires us to think in the following way. Remember that we assume that the null hypothesis is true. Then one of two outcomes can occur:

• One possibility is that results similar to the actual sample are extremely unlikely. This means that the data do not fit in with results from random samples selected from the population described by the null hypothesis. In this case, it is unlikely that the data came from this population, so we view this as strong evidence against the null hypothesis. We reject the null hypothesis in favor of the alternative hypothesis.
• The other possibility is that results similar to the actual sample are fairly likely (not unusual). This means that the data fit in with typical results from random samples selected from the population described by the null hypothesis. In this case, we do not have evidence against the null hypothesis, so we cannot reject it in favor of the alternative hypothesis.

Data Use on Smart Phones

According to an article by Andrew Berg (“Report: Teens Texting More, Using More Data,” Wireless Week , October 15, 2010), Nielsen Company analyzed cell phone usage for different age groups using cell phone bills and surveys. Nielsen found significant growth in data usage, particularly among teens, stating that “94 percent of teen subscribers self-identify as advanced data users, turning to their cellphones for messaging, Internet, multimedia, gaming, and other activities like downloads.” The study found that the mean cell phone data usage was 62 MB among teens ages 13 to 17. A researcher is curious whether cell phone data usage has increased for this age group since the original study was conducted. She plans to conduct a hypothesis test.

The null hypothesis is often a statement of “no change,” so the null hypothesis will state that there is no change in the mean cell phone data usage for this age group since the original study. In this case, the alternative hypothesis is that the mean has increased from 62 MB.

• H 0 : The mean data usage for teens with smart phones is still 62 MB.
• H a : The mean data usage for teens with smart phones is greater than 62 MB.

The next step is to obtain a sample and collect data that will allow the researcher to test the hypotheses. The sample must be representative of the population and, ideally, should be a random sample. In this case, the researcher must randomly sample teens who use smart phones.

For the purposes of this example, imagine that the researcher randomly samples 50 teens who use smart phones. She finds that the mean data usage for these teens was 75 MB with a standard deviation of 45 MB. Since it is greater than 62 MB, this sample mean provides some evidence in favor of the alternative hypothesis. But the researcher anticipates that samples will vary when the null hypothesis is true. So how much of a difference will make her doubt the null hypothesis? Does she have evidence strong enough to reject the null hypothesis?

To assess the evidence, the researcher needs to know how much variability to expect in random samples when the null hypothesis is true. She begins with the assumption that H 0 is true – in this case, that the mean data usage for teens is still 62 MB. She then determines how unusual the results of the sample are: If the mean for all teens with smart phones actually is 62 MB, what is the chance that a random sample of 50 teens will have a sample mean of 75 MB or higher? Obviously, this probability depends on how much variability there is in random samples of this size from this population.

The probability of observing a sample mean at least this high if the population mean is 62 MB is approximately 0.023 (later topics explain how to calculate this probability). The probability is quite small. It tells the researcher that if the population mean is actually 62 MB, a sample mean of 75 MB or higher will occur only about 2.3% of the time. This probability is called the P-value .

Note: The P-value is a conditional probability, discussed in the module Relationships in Categorical Data with Intro to Probability . The condition is the assumption that the null hypothesis is true.

Step 4: Conclusion.

The small P-value indicates that it is unlikely for a sample mean to be 75 MB or higher if the population has a mean of 62 MB. It is therefore unlikely that the data from these 50 teens came from a population with a mean of 62 MB. The evidence is strong enough to make the researcher doubt the null hypothesis, so she rejects the null hypothesis in favor of the alternative hypothesis. The researcher concludes that the mean data usage for teens with smart phones has increased since the original study. It is now greater than 62 MB. ( P = 0.023)

Notice that the P-value is included in the preceding conclusion, which is a common practice. It allows the reader to see the strength of the evidence used to draw the conclusion.

How Small Does the P-Value Have to Be to Reject the Null Hypothesis?

A small P-value indicates that it is unlikely that the actual sample data came from the population described by the null hypothesis. More specifically, a small P-value says that there is only a small chance that we will randomly select a sample with results at least as extreme as the data if H 0 is true. The smaller the P-value, the stronger the evidence against H 0 .

But how small does the P-value have to be in order to reject H 0 ?

In practice, we often compare the P-value to 0.05. We reject the null hypothesis in favor of the alternative if the P-value is less than (or equal to) 0.05.

Note: This means that sampling variability will produce results at least as extreme as the data 5% of the time. In other words, in the long run, 1 in 20 random samples will have results that suggest we should reject H 0 even when H 0 is true. This variability is just due to chance, but it is unusual enough that we are willing to say that results this rare suggest that H 0 is not true.

Statistical Significance: Another Way to Describe Unlikely Results

When the P-value is less than (or equal to) 0.05, we also say that the difference between the actual sample statistic and the assumed parameter value is statistically significant . In the previous example, the P-value is less than 0.05, so we say the difference between the sample mean (75 MB) and the assumed mean from the null hypothesis (62 MB) is statistically significant. You will also see this described as a significant difference . A significant difference is an observed difference that is too large to attribute to chance. In other words, it is a difference that is unlikely when we consider sampling variability alone. If the difference is statistically significant, we reject H 0 .

Other Observations about Stating Conclusions in a Hypothesis Test

In the example, the sample mean was greater than 62 MB. This fact alone does not suggest that the data supports the alternative hypothesis. We have to determine that the data is not only larger than 62 MB but larger than we would expect to see in a random sampling if the population mean is 62 MB. We therefore need to determine the P-value. If the sample mean was less than or equal to 62 MB, it would not support the alternative hypothesis. We don’t need to find a P-value in this case. The conclusion is clear without it.

We have to be very careful in how we state the conclusion. There are only two possibilities.

• We have enough evidence to reject the null hypothesis and support the alternative hypothesis.
• We do not have enough evidence to reject the null hypothesis, so there is not enough evidence to support the alternative hypothesis.

If the P-value in the previous example was greater than 0.05, then we would not have enough evidence to reject H 0 and accept H a . In this case our conclusion would be that “there is not enough evidence to show that the mean amount of data used by teens with smart phones has increased.” Notice that this conclusion answers the original research question. It focuses on the alternative hypothesis. It does not say “the null hypothesis is true.” We never accept the null hypothesis or state that it is true. When there is not enough evidence to reject H 0 , the conclusion will say, in essence, that “there is not enough evidence to support H a .” But of course we will state the conclusion in the specific context of the situation we are investigating.

We compared the P-value to 0.05 in the previous example. The number 0.05 is called the significance level for the test, because a P-value less than or equal to 0.05 is statistically significant (unlikely to have occurred solely by chance). The symbol we use for the significance level is α (the lowercase Greek letter alpha). We sometimes refer to the significance level as the α-level. We call this value the significance level because if the P-value is less than the significance level, we say the results of the test showed a significant difference.

If the P-value ≤ α, we reject the null hypothesis in favor of the alternative hypothesis.

If the P-value > α, we fail to reject the null hypothesis.

In practice, it is common to see 0.05 for the significance level. Occasionally, researchers use other significance levels. In particular, if rejecting H 0 will be controversial or expensive, we may require stronger evidence. In this case, a smaller significance level, such as 0.01, is used. As with the hypotheses, we should choose the significance level before collecting data. It is treated as an agreed-upon benchmark prior to conducting the hypothesis test. In this way, we can avoid arguments about the strength of the data. We will look more at how to choose the significance level later. On this page, we continue to use a significance level of 0.05.

First, work through the interactive exercise below to practice the four steps of hypothesis testing and related concepts and terms.

Next, let’s look at some exercises that focus on the P-value and its meaning. Then we’ll try some that cover the conclusion.

For many years, working full-time has meant working 40 hours per week. Nowadays, it seems that corporate employers expect their employees to work more than this amount. A researcher decides to investigate this hypothesis.

• H 0 : The average time full-time corporate employees work per week is 40 hours.
• H a : The average time full-time corporate employees work per week is more than 40 hours.

To substantiate his claim, the researcher randomly selects 250 corporate employees and finds that they work an average of 47 hours per week with a standard deviation of 3.2 hours.

According to the Centers for Disease Control (CDC), roughly 21.5% of all high school seniors in the United States have used marijuana. (The data were collected in 2002. The figure represents those who smoked during the month prior to the survey, so the actual figure might be higher.) A sociologist suspects that the rate among African American high school seniors is lower. In this case, then,

• H 0 : The rate of African American high-school seniors who have used marijuana is 21.5% (same as the overall rate of seniors).
• H a : The rate of African American high-school seniors who have used marijuana is lower than 21.5%.

To check his claim, the sociologist chooses a random sample of 375 African American high school seniors and finds that 16.5% of them have used marijuana.

Contribute!

Improve this page Learn More

• Interactive: Concepts in Statistics - Hypothesis Testing (2 of 5). Authored by : Deborah Devlin and Lumen Learning. Located at : https://lumenlearning.h5p.com/content/1291194018762009888 . License : CC BY: Attribution
• Concepts in Statistics. Provided by : Open Learning Initiative. Located at : http://oli.cmu.edu . License : CC BY: Attribution

Hypotheses and Proofs

In this post

What is a hypothesis?

A hypothesis is basically a theory that somebody states that needs to be tested in order to see if it is true. Most of the time a hypothesis is a statement which someone claims is true and then a series of tests are made to see if the person is correct.

Hypothesis – a proposed true statement that acts as a starting point for further investigation.

Devising theories is how all scientists progress, not just mathematicians, and the evidence that is found must be collected and interpreted to see if it gives any light on the truth in the statement. Statistics can either prove or disprove a theory, which is why we need the evidence that we gather to be as close to the truth as possible: so that we can give an answer to the question with a high level of confidence.

Hypotheses are just the plural of a single hypothesis. A hypothesis is the first thing that someone must come up with when doing a test, as we must initially know what it is we wish to find out rather than blindly going into carrying out certain surveys and tests.

Some examples of hypotheses are shown below:

• Britain is colder than Spain
• A dog is faster than a cat
• Blondes have more fun
• The square of the hypotenuse of a triangle is equal to the sum of the squares of the other two sides

Obviously, some of these hypotheses are correct and others are not. Even though some may look wrong or right we still need to test the hypothesis either way to find out if it is true or false.

Some hypotheses may be easier to test than others, for example it is easy to test the last hypothesis above as this is very mathematical. However, when it comes to measuring something like ‘fun’ which is shown in the hypothesis ‘Blondes have more fun’ we will begin to struggle! How do you measure something like fun and in what units? This is why it is much easier to test certain hypotheses when compared with others.

Another way to come up with a hypothesis is by doing some ‘trial and error’ type testing. When finding data you may realise that there is in fact a pattern and then state this as a hypothesis of your findings. This pattern should then be tested using mathematical skills to test its authenticity. There is still a big difference between finding a pattern in something and finding that something will always happen no matter what. The pattern that is found at any point may just be a coincidence as it is much harder to prove something using mathematics rather than simply noticing a pattern. However, once something is proved with mathematics it is a very strong indication that the hypothesis is not only a guess but is scientific fact.

A hypothesis must always:

• Be a statement that needs to be proven or disproven, never a question
• Be applied to a certain population
• Be testable, otherwise the hypothesis is rather pointless as we can never know any information about it!

There are also two different types of hypothesis which are explained here:

An Experimental Hypothesis –  This is a statement which should state a difference between two things that should be tested. For example, ‘Cheetahs are faster than lions’.

A Null Hypothesis –  This kind of hypothesis does not say something is more than another, instead it states that they are the same. For example, ‘There is no difference between the number of late buses on Tuesday and on Wednesday’.

Subjects and samples

We have already talked in an earlier lesson of different types of samples and how these are formed, so we will not dwell for too long on this. The main thing to make sure of when choosing subjects for a test is to link them to the hypothesis that we are looking into. This will then give a much better data set that will be a lot more relevant to the questions we are asking. There is no point in us gathering data from people that live in Ireland if our original hypothesis states something about Scottish people, so we need to also make sure that the sample taken is as relevant to the hypothesis as possible. As with all samples that are taken, there should never be any bias towards one subject or another (unless we are using something like quota sampling as outlined in an earlier lesson). This will then mean that a random collection of subjects is taken into account and will mean that the information that is acquired will be more useful to the hypothesis that we wish to look at.

The experimental method

By treating the hypothesis and the data collection as an experiment, we should use as many scientific methods as possible to ensure that the data we are collecting is very accurate.

The most important and best way of doing this is the  control of variables . A variable is basically anything that can change in a situation, which means there are a lot in the vast majority as lots of different things can be altered. By keeping all variables the same and only changing the ones which we wish to test, we will get data that is as reliable as possible. However, if variables are changed that can affect an outcome we may end up getting false data.

For example, when testing ‘A cheetah is faster than a lion’ we could simply make the two animals run against each other and see which is quickest. However, if we allowed the cheetah to run on flat ground and made the lion run up hill, then the times would not be accurate to the truth as it is much harder to run up a slope than on flat ground. It is for this reason that any variables should be the same for all subjects.

The only variable that is mentioned in the hypothesis ‘A cheetah runs faster than a lion’ is the animal that runs. Therefore, this is called the  independent variable  and is the only thing that we wish to change between experiments as it is the thing we wish to  prove has an effect on other results.

A  dependent variable  is something that we wish to measure in experiments to see if there is an effect. This is the speed at which something runs in our example, as we are changing the animal and measuring the speed.

Independent variable – something that stands alone and is not changed by other variables in the experiment. This variable is changed by the person carrying out the investigation to see if it influences the dependent variables. This can also be seen as an input when an experiment is created.

Dependent variable – this variable is measured in an experiment to see if it changes when the independent variable is changed. These represent an output after the experiment is carried out.

Standardised instructions

Another thing that is essential to carrying out experiments is to give both of the participants the same instructions in what you wish them to do. Although this may seem a little picky, there will be a definite difference in how a subject performs if they are given clear and concise instructions as opposed to given misleading and rushed ones.

Turning data into information

Experiments are carried out to produce a set of data but this is not the end of the problem! We will then need to interpret and change this information into something that will tell us what we need to know. This means we need to turn data in the form of numbers into actual information that can be useful to our investigation. Figures that are found through experiments are first shown as ‘raw data’ before we can use different tables and charts to show the patterns that have been found in the surveys and experiments that have been carried out. Once all the data is collected and in tables we can move on to using these to find patterns.

Once a hypothesis has been stated, we can look to prove or disprove it. In mathematics, a proof is a little different to what people usually think. A mathematical proof must show that something is the case without any doubt. We do this by working through step-by-step to build a proof that shows the hypothesis as being either right or wrong. Each small step in the proof must be correct so that the entire thing cannot be argued.

Setting out a proof

Being able to write a proof does not mean that you must work any differently to how you would usually answer a question. It simply means that you must show that something is the case. Questions on proofs may ask you to ‘prove’, ‘verify’ or ‘check’ a statement.

When doing this you will need to first understand the hypothesis that has been stated. Look at the example below to see how we would go about writing a simple proof.

Prove that 81 is not a prime number.

Here we have a hypothesis that 81 is not prime. So, to prove this, we can try to find a factor of 81 that is not 1 as we know the definition of a prime number is that it is only divisible by itself and 1. Therefore, we could simply show that:

The fact that 81 divided by 9 gives us 9 proves the hypothesis that 81 is not prime.

A proof for a hypothesis does not have to be very complex – it simply has to show that a statement is either true or false. Doing this will use your problem-solving skills though, as you may need to think outside the box and ensure that all of the information that you have is fully understood.

Harder examples

Being able to prove something can be very challenging. It is true that some mathematical equations are still yet to be proved and many mathematicians work on solving extremely complex proofs every day.

When looking at harder examples of proofs you will need to find like terms in equations and then think about how you can work through the proof to get the desired result.

Here we need to use the left-hand side to get to the right-hand side in order to prove that they are equal. We can do this by expanding the brackets on the left and collecting the like terms:

We have now expanded the brackets and collected the like terms. It is now that we will need to look at our hypothesis again and try to make the above equation into the right-hand side by moving terms around. We can see from the right-hand side of our hypothesis that we have a double bracket and then 2 added to this so we can begin by bringing 2 out of the above:

So we have now worked through an entire proof from start to finish. Here it is again using only mathematics and no writing:

In the above we have shown that the hypothesis is true by working through step-by-step and rearranging the equation on the left to get the one on the right.

The step-by-step approach to proofs

To prove something is correct we have used a step-by-step approach so far. This method is a very good way to get from the left-hand side of an equation to the right-hand side through different steps. To do this we can use specific rules:

1) Try to multiply out brackets early on where possible.  This will help you to cancel out certain terms in order to simplify the equation.

3) Take small steps each time.  A proof is about working through a problem slowly so that it is easy to spot what has been done in each step. Do not take big leaps in your work such as multiplying out brackets and collecting like terms all at once. Remember that the person marking your paper needs to see your working, so it is good to work in small stages.

4) Go back and check your work.  Once you have finished your proof you can go back and check each individual stage. One of the good things about carrying out a proof is that you will know if a mistake has been made in your arithmetic because you will not be able to get to the final solution. If this happens, go back and check your working throughout.

Harder proofs

When working through a proof that is more difficult it can be quite tricky. Sometimes we may have to carry out a lot of different steps or even prove something using another piece of knowledge. For example, it might be that we are asked to prove that an expression will always be even or that it will always be positive.

In the above equation we have worked through to get an answer that is completely multiplied by 4. This must therefore be even as any number (whether even or odd) will be even when multiplied by 4.

In this example we have had to use our knowledge that anything multiplied by 4 must be even. This information was not included in the question but is something that we know from previous lessons. Some examples of information that you may need to know in order to solve more difficult proofs are:

Any number that is multiplied by an even number must be even

A number multiplied by an even number and then added to an odd number will be odd

Any number multiplied by a number will give an answer that is divisible by the same number (e.g. 3 n  must be divisible by 3)

Any number that is squared must be positive

Above we have come to an answer that is multiplied by 3. This means that the answer has to be divisible by 3 also.

Interested in a Maths GCSE?

We offer the Edexcel IGCSE in Mathematics through our online campus.

Learn more about our maths GCSE courses

Read another one of our posts

Understanding dementia: types, symptoms, and care needs.

How GCSE Business Prepares You for Real-World Entrepreneurship

Preparing for a Career in Adult Social Care: What You Need to Know

Parent’s Guide to Supporting A-Level Students

The Importance of Compassion in Healthcare

The Role of Palliative Care in End of Life Care

Community Health Initiatives – Promoting Wellness Locally

Caring for Older People – Strategies for Providing Quality Senior Care

Save your cart?