Not always but preferable, specially if sample size is higher than 50.

It depends on the type of research and the statistical methods you intend to use. A normality test is not always a mandatory step, but it is crucial in certain cases. Here’s why normality testing is important:

1. Importance of Normality Testing

Many parametric statistical tests (e.g., t-tests, ANOVA, regression, Pearson correlation) assume that the data is normally distributed.

If the normality assumption is violated, the results of these tests may be misleading or inaccurate unless adjustments (e.g., transformations or non-parametric alternatives) are applied.

2. Is Normality Testing a Thumb Rule?

Not always. It depends on the research method and the type of data:

If you're using parametric tests, normality should be checked.

If you're using non-parametric tests (e.g., Mann-Whitney U test, Kruskal-Wallis test, Spearman correlation), normality is not required.

In large samples (n > 30), the Central Limit Theorem (CLT) states that sample means tend to be normally distributed, reducing the need for a normality test.

In machine learning or qualitative research, normality testing is usually not necessary.

3. What to Do If Data is Not Normal?

Transform the data (e.g., log transformation, square root transformation).

Use non-parametric tests, which do not assume normality.

Use bootstrapping or other resampling methods.

if you need more help let me know

Normality tests are carried out after data entry to see what kind of tests can actually be applied. While most parametric tests such as Pearson's Correlation and ANOVA require the data to be normally distributed whereas non-parametric tests do not require the data to be normally distributed.

However, the kind of tests one should adopt depends on several factors. I have made a Youtube Video on that- https://youtu.be/F4AKZ9-u_ic?si=-CeYV7OmuQoN_RYe

You can find a lot more about normality and data analysis on my channel.

https://www.youtube.com/@ramagokulakrishnan

Yes, it's important to perform a Normality test before certain statistical analyses.

Why? Many statistical tests assume the data follows a normal distribution (bell-shaped curve). Violating this assumption can lead to inaccurate results.

When?

Parametric tests: Essential (e.g., t-tests, ANOVA)

Non-parametric tests: Not required (designed for non-normal data)

It is a thumb up rule in some cases such as parametric testing. This is not important in non-parametric testing where normal distribution is not required. Normality test is carried after data is collected and analysed.