Posted: September 4th, 2022

Statistical tests

Statistical tests

Statistical tests involve a process of accepting or rejecting a hypothesis by comparing evidence generated from data against the hypothesis.

ANOVA

Analysis of Variance (ANOVA) is a statistical method that seeks to find out the difference between groups. For example,  a group of patients trying three types of pain reliever paracetamol, ibuprofen and aspirin. You want to find out which one works best

There are different types of ANOVA models such as

1. MANOVA
2. SPANOVA
3. ANCOVA
4. T-tests
5. One way ANOVA

Chi-square

It is a statistical method that seeks to find out whether the distribution of data observed in a sample fits well with the expected data. If the variables are constant. There are several unique characteristics of the Chi-square test. These include;

1. It can only analyze categorical data.
2. It examines the null hypothesis when the variables are independent.
3. Data must be in numbers but not in percentages

T-tests

It is a statistical test that seeks to find out the significant difference between two data sets that are related. It has an alternative and null hypothesis. The null hypothesis states that there is no difference between the two sets of data. The alternative hypothesis states that there is a difference between the two data sets.

The T-test looks for a critical value which helps identify whether there is a significant difference between the datasets. It compares against 0.05 p-value. If the critical value after the end of the test is less than 0.05 this indicates that there is only a 5% chance that the data is random. It proves that the null hypothesis is correct.

For example

Finding out whether there is a relationship between the grades of public schools and private hypothesis.

Types of T-tests

1. Un equal variance T-test
2. Equal variance T-test
3. Correlated T-test

Correlation test

It is a statistical test that seeks to find out the relationship between two or more variables. For example, whether there is a correlation between the height of a mother and her daughters. There are different types of correlation analysis

1. Pearson correlation – it measures linear dependence when there is a normal distribution between x and y.
2. Kendall tau: it is used to measure the relationship between rank based data.
3. Spearman rho: it is also used to measure the relationship between rank based data that does not have a normal distribution.
4. Compute correlation in R: It is the use of R functions to check the association between two variables. The functions can be computed using cor () or cor.test (or)

Regression test

It is a software test that seeks to establish whether a code or program change has affected the existing features. It involves re-running functional and non-functional tests to ensure that the existing functionalities are working well. There are different techniques for conducting a regression test

1. Retest all – It is re-running all the tests in the existing test suite.
2. Regression test selection – It is running of a test in parts of the test suite.
3. Prioritization of test cases – involves running of tests depending on the impact and use of parts.

Z-tests

It is a statistical test used in large samples to determine if there is a noteworthy difference between two means of groups when the variance are known and some features are similar. There are several steps involved in Z-test. These include:

1. Define the null and alternative hypothesis
2. State alpha
3. State decision rule
4. Calculate test static
5. State results
6. State conclusion

Parametric tests

It is a statistical test that examines a hypothesis assuming that the sample has a normal distribution. It is the complete opposite of non-parametric tests which assumes that the sample does not have a normal distribution.

Statistical software

These are programs used in collection, organization, analysis, interpretation, and presentation of statistical data.

SPSS

Statistical Package for the Social Sciences (SPSS) is a software package that analyzes statistical data. It analyzes data from different sources such as survey results, organization customer databases, and scientific research results among others. SPSS analyzes many kinds of data with different formats such as spreadsheets, plain text files, and relational databases.

STATA

It is a statistical software package that offers data science needs. Such as analysis, manipulation, interpretation, and presentations.

STATA capabilities

1. Management of data
2. Graphical representations
3. Statistical analysis
4. Simulations
5. Customizing programs.

R software

It is a software that has a collection of analysis tools, it is also a programming language for statistical computing.

Uses

1. Data handling and storage
2. Doing calculations on spreadsheets and arrays
3. Displaying graphics on the screen.

E VIEWS

It is a statistical software for windows that is used for data analysis, time series estimation and forecasting.

Uses

1. Statistical and panel data analysis.
2. Econometric analysis
3. Time series estimation

SAS

It is a software package that does the following:

1. Managing data
2. Data analysis
3. Business intelligence

ARC GIS

It is a software that works with maps and geographical information. It used to collect, organize, interpret, map, analyze and share geographical information to access real-world problems.

Conclusion

In conclusion, statistical tests play a vital part in proving a hypothesis right or wrong, while statistical software is essential in data analysis. They both play an important role in developing beneficial information that is crucial for decision making.

Order | Check Discount