Repeated Measures Design

Paired samples t-test

The paired sample t-test, also referred to as the dependent samples t-test examines the statistical significance of the changes between two paired observations. The test involves measuring same subjects at two time points or observations using two different methods. This test's application requires use of paired variables in situations where paired variables need to be continuous and have a normal distribution. The mean and the standard deviation of the paired differences and the sample size enable the calculation of the significance level.

Repeated Measures ANOVA

ANOVA is a testing mode to determine equality means in different treatment groups. Designs containing the within-subjects factors are referred to as repeated measures ANOVA. Within-subjects factors are the factors are factors where measurements acquisitions are acquired on the same subject over time or under similar conditions. Within-subjects designs are advantageous over other the between-subject ANOVA. One of the advantages is that in clinical setting, only few subjects are needed compared to the between-subject ANOVA. The repeated measures design efficiently use subjects in the aspect of having fewer subjects than between-subject while in the statistical aspect having minimal error variance.

Correlational Analysis-Pearson's Correlations

Correlation is a statistical term denoting the close relationship between two variables. Correlation is also the closeness of the points relative to the straight line. In a positive correlation one variable goes up as the other variable also goes up. In the negative correlation, one variable goes up as the other variable goes down. Pearson correlation, also known as the Pearson product moment correlation is the commonly used correlation. The person correlation coefficient determines the strengths of the linear relationship existing between two variables. In a sample, r represents Pearson correlation coefficient. The scale measuring the coefficient has no units, and take a value ranging from -1 to +1. Pearson's calculations can only be undertaken if having numerical variables with at least one having a normal distribution.

Spearman's Rank-order correlations

The Spearman's Rank-order correlations, also referred to as the Spearman's p, compares the relationship between ordinal, or rank-ordered variables. This correlation is a statistical process formulated to determine the relationship between two variables positioned on an ordinal measurement scale when the sample size is n≥4. The spearman rank correlation coefficient could apply if the assumptions for Pearson's Correlational coefficient are insufficient, where it is impossible to assume the linear association without any of the variable having a normal distribution, or if having at least a single discrete variable, such as when measured on an ordinal scale. Ρs represents the population parameter for Spearman's rank correlation coefficient.

Partial Correlations

Partial correlations refer to the correlations between two variables following the elimination of the startling impact of the third variable. The first two variables could be X and Y, with the third variable being Z. Variable Z is the controlling variable. Partial correlation is founded on Spearman rank correlation (proportion correlation). Partial correlation eradicates the indirect impact emanating from the remaining taxa. A statistically instrumental partial correlation denotes the primary direct interaction between two taxa. The partial correlational coefficient was developed to deal with the challenge emanating from the inability to differentiate direct associations from those associated with confounding factors. The bone of contention is deducting from with the information in controlled variables through linear regression.