This site attempts to protect users against Cross-Site Request Forgeries attacks. In order to do so, you must have JavaScript enabled in your web browser otherwise this site will fail to work correctly for you. See details of your web browser for how to enable JavaScript. Skip to Main Content Library - University of Liverpool
Toggle mobile navigation

Library News and Blogs

How to choose the right statistical test for your data

by University of Liverpool Library on 2021-04-28T12:40:22+01:00 | 0 Comments

Laptop on a sofa showing charts and statistics

Knowing which statistical test is most appropriate for your data can feel overwhelming. We’ve got some tips to help you get started.

  1. Take your time

Don’t feel rushed into choosing the right test. Slow down and think it through logically. When conducting research or in your future career you will always have time to consider your options and talk to others. 

  1. What do you want to do with your data?

Start by considering the hypothesis of your study and what you want to test.

  • Are you looking for a difference?
  • Are you looking for a relationship?
  • Are you trying to predict an outcome?
  1. How many variables do you have?

Most tests usually involve just one outcome (dependent variable), and there are several tests such as t-tests which can only be used with one independent variable too. Other tests such as ANOVAs and regressions can be used when there are one or more independent variables.

  • Variables are data you have recorded or measured e.g. height, weight, gender
  • A dependent variable is your outcome variable, in other words the result you are interested in
  • An independent variable is something you have recorded or measured, that might affect your result/outcome (dependent variable)
  1. What type of data do you have?

Next, consider what type of data you have collected, nominal, ordinal or continuous? What type of data are your dependent and independent variables?

  • Nominal- categorical data which fall into groups with no order to them .e.g gender, eye colour. Can only be used as total counts
  • Ordinal- categorical data with an order e.g. passenger class, Body Mass index, grades etc.
  • Continuous- data which can be used to create an average e.g. weight, height, distance

To ensure you are familiar with different types of data. Complete the KnowHow online tutorial Getting started with statistics- Types of data.

  1. Are your data paired?

Paired data usually involve measuring something at least twice on the same individual e.g. weight before a new diet and weight after a new diet. Data which are not paired are called independent data e.g. measuring the weight loss of people who had diet 1, versus the weight loss of people who were given diet 2.

  1. Are your data normally distributed?

After working through this list, you should have a good idea which test you’d like to use. Several tests will have assumptions and it’s important to check that your data meet these assumptions before using a test. One of the most common and important assumptions is that your data follow a normal distribution.

Further support

If you’d like to know more about choosing the right statistical test for your data, you can work through the KnowHow online tutorial Choosing Statistical Tests

Look out for KnowHow sessions on statistics and if you have a question about statistics, you can book a one-to-one appointment with one of the Stats@Liverpool tutors.


 Add a Comment

0 Comments.

  Subscribe



Enter your e-mail address to receive notifications of new posts by e-mail.


  Archive



  Return to Library News
This post is closed for further discussion.

title
Loading...