A Simple Guide on How the T-test and Analysis of Variance Works for Nurses

A Simple Guide on How the T-test and Analysis of Variance Works

For people doing their research, whether it’s a thesis or a dissertation, the use of statistical tools such as the T-test and Analysis of Variance is an essential component to test the hypothesis mentioned in the study. These statistical methods are termed as parametric statistical techniques. When you say parametrical statistical technique, it means that there should be a normal distribution in order to make standardized comparisons across different populations or treatments.

Testing of differences among means between two groups require the use of T-test. It can be used when testing means among topics such as self-esteem or gender wherein the variable (dependent) becomes an interval which is continuous for the former and a 2 level categorical variable similar to the latter.

The p-value will be determined in order to get how likely the hypothesis is. A result which is greater than or equal to 5% p-value indicates that there is a significant difference between the groups and so, the null hypothesis is rejected.

The two sample T-test determines if the 2 independent variables have different mean values on some degrees. One example of a two-sample t-test is testing the variables between a group of patients who were given a certain treatment and a control group given blank or placebo.

Another type is the paired sample T-test which computes for the differences in tests scores on the same sample of patients using a pre-test and another post-test paradigm. This is applied in cases such as determining before and after treatment of level of cholesterol.

The following information should be included in reporting the result of T-test:

• Checking for validity of parametric assumptions
• dependent variable scores
• independent variable levels
• statistical data used for computational purposes

Meanwhile, if the study needs to compare means between three or more groups, ANOVA or Analysis of Variance is used. Here is a step by step procedure in performing ANOVA (in this example, the one way ANOVA is demonstrated):

• Step 1: Use the Ryan-Joiner Test to test for normality of distribution for each interval scale response variable.
If P < 5%, the variable is not normally distributed (proceed to step 2)
If P > 5%, the variable is normally distributed (proceed to step 4)
• Step 2: Transform the variables that are not normally distributed. Using the calculator, the log transformation, square-root transformation and arc-sin transformation will be used for these variables. (proceed to step 3)

• Step 3: Test for the normality of distribution for transformed variables using the Ryan-Joiner Test
If P < 5%, the variables are not normally distributed (proceed to step 5)
If P > 5%, the variables are normally distributed (proceed to step 4)

• Step 4: Perform ONE-WAY ANOVA for variables which are normally distributed. Using Turkey’s Test, compare the variables.
If P <5%, the variables are significantly different.
If P >5%, the variables are not significantly different.

• Step 5: Use the Kruskal-Wallis for variables that are not Normally Distributed and the transformed variables are also not Normally Distributed.

If P <5%, the variables are significantly different.
If P >5%, the variables are not significantly different.

And so, there you have it! Performing statistical analyses is easy as one two three. You only need to research and perform it carefully in order to avoid miscalculations and other unnecessary errors.