Statistics is now an essential component of biology. As most science students have to carry out a final year research project, it is good to know some basic information about statistics.
Here we learn the following terms: One sample t-test
Music from www.bensounds.com
Script of text
Welcome to Lesson 5. To all who have subscribed, thanks a million. Your support is very much appreciated.
Earlier, I have explained how statistical analysis can be divided into descriptive statistics that applies only to the sample studied, and inferential statistics where we test hypothesis and make conclusions about the population.
In Lessons 3 and 4, we have covered the different descriptive statistics for central tendencies and dispersion of mean. From this lesson 5 onwards, we will be looking at inferential statistics.
In inferential statistics, we will cover comparative statistics to test the means for similarity, association statistics to look for relationships, patterns and trends, and also frequency statistics where we focus mostly on using Chi-square tests for association and Goodness of Fit tests.
NEXT SLIDE:
In comparative statistics, we can compare the means of one group, two groups and more than two groups.
For one group, usually we will compare the mean against a criterion or limit using one sample t-test. This is for parametric testing.
For non-parametric testing when your data does not fulfill normal distribution, you may use the Wilconox Rank Sum or Signed Rank test to test the median against a criterion. However if data transformation normalizes your data, then I will suggest that you transform your data and use parametric tests. This is a more robust approach to statistical testing.
For all the following statistical tests, I will be using Excel and PAST. Excel is probably a more familiar option for most of you. However you will need to have the Analysis Tool Pak installed.
You may learn how to do this in this video:
[ Ссылка ]
However if you are going to be working with a fair amount of data e.g., for your final year project, may I suggest you try out PAST? This is a free statistical software that also handles advanced level statistics including multivariate analysis.
Of course, Excel continues to be helpful in handling, managing and arranging your data.
You may start learning PAST by this link:
PAST Lesson 1: How to start using PAST
[ Ссылка ]
Or if you wish, I have a playlist of lessons in PAST that may be helpful.
In the Analysis Tool Pak add-in that is available in Excel, there is actually no option for a one-sample t-test whereas PAST offers a useful one sample testing. You can see how it is done in PAST on the link here.
PAST Lesson 5: One sample t-test
[ Ссылка ]
However, you can still use Excel by doing the following:
EXCEL
Thanks for watching, and please remember to subscribe.
