TYPE I and Type II Errors, All about P- Value

-

Continuing our last blog about Hypothesis testing, which allows us to formulate hypothesis, and we use this hypothesis to perform a Statistic-Test. We also reject or approve the hypothesis as per the significance level selected by the researcher.

What is Type I error? I’ll explain with the help of an example, suppose a patient experiencing a headache visited a doctor, and after doing the preliminary examination, doctor formulates a hypothesis

Null Hypothesis:- The patient has Migraine

Alternative Hypothesis: - The patient doesn’t have Migraine.

Suppose a doctor ends up making the decision that the patient doesn’t have a migraine and sends him home which means he rejected the Null hypothesis, whereas originally, the patient was suffering from migraine. This could prove fatal, as the patient’s condition may worsen. This is the Type I error.

In the Type II error, if originally the patient doesn’t have a migraine, however doctor ended up concluding that he is suffering from migraine, he might end up prescribing some medicine, which may not make the situation grave. This is the example of Type II error.

When someone rejects the null hypothesis, however it turns out that null hypothesis was true, then this error is Termed as TYPE I error.

All about p- value



P-value is the probability of the Null hypothesis being true. Null hypothesis treats everything equal, which means, if we use Medicine A and Medicine B on few people, the effect of both the medicines will be same or equal, or the number of boys is equal to number of girls in a school.

When there is a null hypothesis, an alternate hypothesis is formulated, so Medicine A and Medicine B has a different effect on the people, or number of boys is different from number of girls in a school beocmes  an Alternate Hypothesis.

 If we perform a test on the above hypothesis, p value is calculated and if p value is 0.01, which means only 1% cases will be there where Medicine A and Medicine B has the same effect on the patient, and rest of the patients will have a different effect. In this case we will reject null hypothesis. To make it more simpler to understand, if we perform a statistical test, to check if the number of boys and the number of girls are same in a school, and gets a p value of 0.05, it will mean out of 100 cases only 5 cases will have null hypothesis true. There will be only 5 cases where the number of boys will be equal to number of girls.

In general if p value < 0.05, then we reject null hypothesis, otherwise we accept null hypothesis.

A bit about Statistical tests, in hypothesis testing, we formulate a hypothesis, and perform some statistical test to obtain a p value. These statistical tests are

     1.)  t- Test

     2.)  Chi- Squared Test

     3.)  Anova Test

     4.)  z- test

     5.)  f-test

We will study in detail about each one of these test and when to perform these tests in the upcoming blogs. Stay tuned!


Comments

Post a Comment

Popular posts from this blog

Quick Mind Map for Statistics - Part 1

-
Image
Statistics is the one of the pillar or a prerequisite of a Data Science and is used in Data analysis and Visualization in the lifecycle of a Data Science Project. In this series, presenting you a quick mind map of Statistics for Data Science. Statistics needs to be studied in the following order.   Mind- Map for Statistics Statistics is subdivided into 1.) Descriptive Statistics 2.) Inferential Statistics.   This diagram represents the mind-map of Descriptive Statistics. Descriptive statistics describes the characteristics of the data set. It is further subdivided into three categories 1.) Measures of Central Tendency 2.) Measures of Variability 3.) Measures of a Symmetry.   Measures of Central Tendency tells about the center of the dataset and are 1.) Mean 2.) Median 3.) Mode   Measures of Variability tells about the dispersion of data points around the mean and are 1.) Variance 2.) Standard Deviation 3.) Co Variance 4.) Coefficient Correlation
Read more

Quick Mind Map for Statistics - Part 2

-
Image
  Statistics is the one of the pillar or a prerequisite of a Data Science and is used in Data analysis and Visualization in the lifecycle of a Data Science Project. In this series, presenting you a quick mind map of Statistics for Data Science. Statistics needs to be studied in the following order. Statistics is subdivided into 1.) Descriptive Statistics 2.) Inferential Statistics. This diagram represents the mind-map of Inferential Statistics. Inferential statistics is used to make the inferences about the population parameters, using sample statistics. For example we will use Sample Mean to calculate the Population mean. Like descriptive statistics, inferential statistics does not have any subsequent divisions, but concepts that help us to make inferences or predictions about Population parameters.   So the first thing you need to know when you are learning the inferential statistics is the difference between 1.) Population 2.) Sample You can find the difference  betwee
Read more

Statistics - Measure Of Central Tendency

-
Image
Statistics  is the discipline that concerns the collection, organization, analysis, interpretation, and presentation of DATA. Statistics is applied to a Scientific, Industrial or Social problems.  Statistic has been extensively used in the field Of Genetics or Bio Statistics to be precise. Statistics helps us to gain insights from the Data of any kind.  Image by  GraphicMama-team  from  Pixabay   DESCRIPTIVE STATISTICS In this blog we will hover over DESCRIPTIVE STATISTICS, which describes or summarizes characteristics of the data, it consists of two basic categories Measure of Central Tendency and Measures of Variability. This section will deal with Measure of Central Tendency. POPULATION VS SAMPLE Population in general means the number of people in a country or a specific place however, In Statistics, Population means a collection of people, collection of objects, collection of numbers, collection of observations , collection of instances or a collection of anything (data) that n
Read more