Hello,
I want to know when can we use Student's test, Pearson's correlation coefficient test and Spearman rank correlation, what can i put us matrix (numeric variables as means..sd (standard deviations..) ?
Thank you
Have you tried searching online? I'm sorry, but these tests (especially the t-test and Pearson correlation coefficient) are some of the most basic/foundational tests in [frequentist] statistics, so you could find the answers to these questions very easily.
Briefly, a t-test is for when your dependent/outcome variable is continuous and you want to compare the mean of this variable in one group to the mean in another group (or to a fixed number). A Pearson correlation coefficient is for when not only is your outcome variable continuous, but your independent/predictor variable is also continuous and you want to see if the two variables have a systematic relationship. The Spearman correlation is similar, except it also tests relationships that are not necessarily a straight line (you can find lots of examples with illustrations online).
Pearson correlations can be used when the residuals have a normal distribution. If the residuals are not normal then use spearman correlations.
Thank you for the responses. Can you give a reference of an accessible biostatistic book please ? I need more explanations about all the biostatistics tests
Dear Meriem
Independent samples T- test, used for normally distributed data ( also, there are dependent two samples T- test , and one sample T- test ).
Pearson correlation used for two scale independent variables which normally distributed, while spearman correlation used in case of ordinal variables and not normal distributed data.
For more comments, you can use the link:
https://www.researchgate.net/post/Why_should_data_be_normally_distributed_and_continuous_in_order_to_apply_Pearson_correlation
Good Luck
Hi, This is what I have made from net information .May be it is useful
Hi,
I understand what you write but cannot decide which test to use. One of my research questions for my dissertation is "What differences/similarities are there between primary school teachers' beliefs about using L1 and their actual use of L1?" so should I use t-test to find correlation between teachers' beliefs and their actual use of L1?
Hi,
As outlined above, Pearson coefficient correlation is used when we want to estimate the linear relationship between two quantitative measures (normally distributed....). Yet, here is the trick : we need to use the t-test to assess the strength of the evidence against the null hypothesis (rho=0). What does it mean ?
To test the hypothesis that the Pearson correlation coefficient is different than 0 in the universe, we can not use the value of the calculated R coefficient (because the distribution of the statistic R under the null hypothesis is unkown). Therefore, we need to calculate a t-value from the R value in order to obtain the value of a random variable ~ t distribution with df = N-2 (because the distribution of the t statistic under the null hypothesis is known). The formula of this t-test (which is different than the formula used to compare two means ) is explained in many websites. Then you just have to use a t table two-tailed to find the critical t value for your number of df and to check if your calculated t value > critical value for df = N-2.
Usually, we don't report this value when reporting our results, we just quot (R=..., p=....), but we can not obtain the p-value if the t-value is not first calculated from the Pearson R-value. This t-value is automatically calculated by softwares, therefore, most of the time, they provide 1) the R value, 2) the related t-value and 3) the p-value.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis