Independent Samples T-test

Definition (#f7aeae)

Important (#edcae9)

Extra (#fffe9d)


T-test:

Introduction to t-test:

  • Involves comparison of 2 “groups” or “conditions” or “levels” (IV).

  • Whether differences between the groups/levels are significant or not.

  • DV must be a continuous variable.

  • 3 types:

    1. One sample t-test.

    2. Independent-sample t-test.

    3. Paired sample t-test.


General assumptions on t-test:

  • DV must be interval or ratio scale.

  • Random sampling from the population.

  • Data is normally distributed.

  • Variances of groups / conditions are equal.


T-test:

  • Compares 2 types of variability:

    • Explained variability: The difference between groups (caused by the IV).

    • Unexplained variability: Differences within each group (caused by other factors like random error or individual differences).

  • t: Explained variance (difference in means)/unexplained variance (standard error).


  • If there's a big difference between group means and small differences within groups, the t-value will be large. Suggesting the groups are truly different.

  • t-value close to zero means the group means are similar, supporting the null hypothesis (no real difference).

  • t-value far from zero means the group means are likely different, and the p-value will be small (p < α), suggesting a statistically significant result.


Understanding Degrees of Freedom (df):

  • df: Number of independent values that can vary in a statistical calculation.

  • Represents the amount of information available to estimate the parameters.

  • Related to sample size – more data, more freedom.

  • Used in many statistical tests (t-tests, ANOVA, chi-square).

  • Affects the shape of the sampling distribution.

  • The 'flexibility' in your data before constraints are applied.


Degrees of freedom in t-test:

  1. One-sample t-test: df = n - 1 [ex: n = 20 → df = 19]

  2. Independent samples t-test: df = n₁ + n₂ - 2 [ex: n₁ = 15, n₂ = 15 → df = 28]

  3. Paired samples t-test: df = n - 1 [ex: n = 12 → df = 11]


Independent Samples T-test:

Overview:

  • Compares the means of two independent groups.

  • Used to test whether there is a statistically significant difference between the two groups.

  • Groups must be unrelated (Group A ≠ Group B).

  • Each participant contributes only one score to the analysis.


When Do We Use Independent Samples t-test:

  • Used in between-subjects design.

  • Each group must consist of different participants.

  • Commonly used in experimental and quasi-experimental research.

  • Useful when comparing two conditions/groups (control vs. treatment).

  • Works best when the sample sizes of the two groups are similar.


Sample size:

  • Each group should have at least 20-30 participants to ensure statistical power and reliability.

  • Small samples (< 20 per group) can be used but may:

    • Violate assumptions (especially normality).

    • Reduce power (harder to detect a real effect).

    • Be more vulnerable to outliers and sampling bias.


Assumptions:

  • Independence of observations: No participant is in more than one group.

  • Normality: Scores in each group should be approximately normally distributed.

  • Homogeneity of variance: The variances of the two groups should be roughly equal.

  • Dependent variable should be measured at interval or ratio level.


Analysis:

  1. Analyze → Compare means → Independent sample T-test.

  2. Put DV in “test variable” → Put IV in “grouping variable”.

    Step 2
  3. Click “define groups” → Put in the values you used for each group → “Click continue”.

    Step 3
  4. Click “OK


Interpretation:

  1. Check Levene’s test:

    • If not significant (variances are equal), use values from the first row. If significant (variances not equal), use values from the second row.

    • Table: Independent Samples Test.

  2. See “sig”:

    • If sig (p) <.05, there is a significant difference between groups. If sig (p) >.05, there is no significant difference between groups.

    • Table: Independent Samples Test.

  3. See the means of each group to see which group scored higher only if there is a significant group difference.

    • Table: Group Statistics

Output


Homogeneity of variances:

  • Levene’s test not significant (p > .05): Variances of the 2 samples are not statistically different.

  • Levene’s test significant (p < .05) : Variances of the 2 samples are statistically different.

    • The assumption is not met.

    • Have to use corrected t-test instead.


Template:

  • What test you ran and what variables were plugged in.

  • Is there a significant difference between group.

  • Report the “equation” above with the values.

  • Interpretation:

    • If significant: Which group scored higher and lower.

    • If not significant: Groups had similar scores.