Hypothesis Testing and T Tests

Understanding Hypothesis Testing using the T Test

  • Concept of Hypothesis Testing:

    • Involves making a statement (hypothesis) about the population, which can be tested through sample data.
    • The focus is on determining whether differences in sample means indicate true differences in population means.

  • Data Gathering:

    • Collect sample data from two groups (conditions) to generate sample means.
    • Example: If Group 1 has a mean of M1 and Group 2 has a mean of M2, the difference between these sample means is M1 - M2.

  • Infinite Data Collection:

    • Imagine gathering sample data numerous times to create a distribution of differences between sample means.
    • This forms a hypothetical distribution of differences from which T values can be generated.

  • Formula for the T Statistic:

    • The T statistic is calculated using:
    • T = (X̄1 - X̄2) / SE
    • Where SE (Standard Error) is based on the standard deviations and sample sizes of the two groups.

  • Interpreting Large vs. Small Differences in Means:

    • Larger Difference: Leads to a larger T value and greater likelihood of obtaining significant results.
    • Smaller Difference: Results in a smaller T value and is less likely to indicate statistical significance.

  • T Distribution:

    • The population of T values forms a distribution often resembling the bell-shaped normal curve.
    • Characteristics:
    • Symmetrical around the mean.
    • Extreme values (tails) become less probable as you approach them.

  • Understanding Alpha (α):

    • Alpha is the threshold for significance, commonly set at 0.05 (5%).
    • This implies 2.5% in each tail of the distribution (lower and upper extremes).
    • Alpha determines the probability of incorrectly rejecting a true null hypothesis (Type I error).

  • Decision Rule:

    • If the calculated T falls into the critical regions (extreme T values), the null hypothesis (H₀: μ1 = μ2) is rejected.
    • This leads to a conclusion that the population means are statistically different.

  • Calculating P Values:

    • Compare P value from sample data to alpha.
    • If P < α, reject the null hypothesis.
    • Example: If P = 0.017 < 0.025 (α/2), reject H₀.

  • Possible Errors:

    • Even with rejection, there is a 5% chance the null is true (Type I error).
    • Maintain rigor in testing to keep this error rate low.

  • Key Points on Hypothesis Testing:

    • Gather sample means from two conditions.
    • Calculate T using the difference of means and standard error.
    • Use the T distribution to evaluate significance against alpha.
    • Decide to reject or not reject H₀ based on comparisons of P and α.