t TEST

What is t test
◻ The common-use definition or description of t tests
is simply comparing two means to see if they are
significantly different from each other.

◻ The two groups could be, for example, patients

who received drug A once and drug B once, and
you want to know if there is a difference in blood
pressure between these two groups.
t Distributions
◻ When the population standard deviation (σ) is
unknown or the sample size (n) is small (typically
n<30), the normal distribution-based z-scores are
no longer accurate.
◻ The t-distribution adjusts for this by accounting
for the sample size, making it more reliable for
smaller datasets or unknown population
parameters.
Characteristics of t-Distributions
◻ Shape: The t-distribution is bell-shaped like the normal
distribution but it has heavier tails, meaning it is more
spread out and allows for more extreme values..
◻ Dependence on Sample Size: The exact shape of the
t-distribution depends on degrees of freedom (df),
which is related to the sample size:
df=n−1 for a single sample.
The smaller the sample size (or df), result in wider and
flatter distribution.
As n becomes large or as df increases, the t-distribution
approaches the normal distribution. (n>120)
When to Use the t-Distribution
◻ Small Sample Sizes (n<120):
Use the t-distribution when your sample size is small, and
you are estimating probabilities or performing hypothesis
tests.
Example: Estimating the mean height of a population with
a sample of n=25.
◻ Population Standard Deviation Unknown:
If the population standard deviation (σ) is unknown, you
estimate it with the sample standard deviation (s).
This introduces variability, which the t-distribution
accounts for.
The Independent Samples t Test
◻ You use this test when you want to compare the
means of two independent samples on a given
variable.
For example, if you wanted to compare the average
height of 50 randomly selected men to that of 50
randomly selected women, you would conduct an
independent samples t test.
◻ there is no overlap between these two samples,
these groups are independent
◻ To conduct an independent samples t test, you need
one categorical or nominal independent variable
and one continuous or intervally scaled dependent
variable
◻ For example, we may want to know if the average
height of people (height is the dependent,
continuous variable) depends on whether the
person is a man or a woman (gender of the person
is the independent, categorical variable).
Dependent Samples t Test
◻ A dependent samples t test is also used to compare
two means on a single dependent variable.
◻ Unlike the independent samples test, however, a
dependent samples t test is used to compare the
means of a single sample or of two matched or
paired samples.
Examples
◻ If a group of students took a math test in March
and that same group of students took the same math
test two months later in May, we could compare
their average scores on the two test dates using a
dependent samples t test.
◻ Or, suppose that we wanted to compare a sample of
boys’ Scholastic Aptitude Test (SAT) scores with
their fathers’ SAT scores.
◻ In this example, each boy in our study would be
matched with his father.
In both of these examples, each score is matched, or
paired with, a second score.
Because of this pairing, we say that the scores are
dependent on each other, and a dependent samples t
test is warranted.
Conceptual Issues with the
Independent Samples t Test
◻ The most complicated conceptual issue in the
independent samples t test involves the standard
error for the test
◻ It is designed to answer the most simple question.
Do two independent samples differ from each other
significantly in their average scores on some
variable?
◻ a random sample of 50 men differs significantly
from a random sample of 50 women in their
average enjoyment of a new television show
◻ In male sample average rating is 7.5 and my sample
of women gave the show an average rating of 6.5
◻ The question is whether they differed significantly
in their average enjoyment of the show?
◻ In other words, is this difference of 1.0 between my
two samples large enough to represent a real
difference between the populations of men and
women?
◻ “Is the difference between the means of these two
samples statistically significant?”
◻ You must know how much difference you should
expect to see between two samples of this size
drawn from these two populations?
◻ The critical question here is this: What is the
average expected difference between the means of
two samples of this size (i.e., 50 each) selected
randomly from these two populations?
In other words, what is the standard error of the
difference between the means?
The Standard Error of the Difference
between Independent Sample Means
◻ The standard error of the difference measures:
How much difference we should expect, on
average, between two sample means selected
randomly from a population.
◻ The standard error of the difference between
independent sample means is a little bit more
complex than the standard error of the mean
◻ Unlike the standard error of a single sample, this
involves two samples, so we need to combine their
standard errors.
◻ If the two samples:
◻ Have unequal sizes or
◻ Different standard deviations,
the formula must be adjusted to account for these
difference
Formula
Determining the Significance of the t Value for
an Independent Samples t Test

◻ It involves checking if the difference between two

group means is likely due to chance or represents a
real difference. Here's how it works in simple
terms:
◻ Calculate the t-value: This is a number that tells us
how much the two groups differ, considering the
variation in the data.
◻ Find the p-value: The p-value tells us the
probability of seeing such a difference (or more
extreme) if the groups are truly the same.
◻ Compare the p-value to the significance level
(α):
The most common α is 0.05 (5%).
If the p-value is less than 0.05, it means the difference
is statistically significant (not likely due to chance).
If the p-value is greater than 0.05, the difference is
not statistically significant (it could just be random).
Paired or Dependent Samples t Tests in
Depth
◻ The goal of the dependent samples t-test is to
determine if an observed difference in a sample
reflects a true difference in the population.
◻ Example:
◻ A factory owner compares the productivity of 30
employees before and after a 2-week vacation.
Pre-vacation average: 250 widgets.
Post-vacation average: 300 widgets.
Observed difference: 50 widgets.
◻ Key Question:
◻ Does this 50-widget difference represent a real
improvement in the productivity of all employees,
or is it just due to chance?
◻ Comparison with Independent Samples t-Test:
Instead of comparing two different groups, a dependent
samples t-test compares the same group on two
occasions (e.g., pre-vacation vs. post-vacation).
◻ Another difference between dependent and
independent samples t tests can be found in the
calculation of the degrees of freedom
◻ Whereas we had to add the two samples together
and subtract 2 in the independent samples formula,
for dependent samples we find the number of pairs
of scores and subtract 1
T test procedure in SPSS
◻ Analyze, Compare Means, Independent Samples
T-Test
◻ Make the variable for which you want to compute a
mean the “Test Variable”
◻ Make the variable that defines the two groups the
“Grouping Variable”
◻ Click “Define Groups”, then enter the value that
each group has on the “Grouping Variable”
What to remember
◻ The t-test for the difference in means is an
hypothesis test that tests the null hypothesis that the
means for both groups are equal, versus the
alternative hypothesis that the means are not equal
(2-tail) or that the mean for one of the groups is
larger than the mean for the other group (1-tail).
◻ To interpret the t-test results, all you need to find
on the output is the p-value for the test.
◻ To do an hypothesis test at a specific alpha
(significance) level, just compare the p-value on the
output (labeled as a “Sig.” value on the SPSS
output) to the chosen alpha level.
Example SPSS Output for T-Test for
Difference in Means
Interpretation of output
◻ Note the mean for each of the two groups in the “Group
Statistics” section.
◻ This output shows that the average weight for European
cars is 2431 pounds, versus 2221 pounds for Japanese
cars.
◻ To determine whether this difference in means is
statistically significant, we examine the p-value for the
t-test.
◻ The p-value is labeled as “Sig. (2-tailed)” in the SPSS
output, which stands for the significance level of the
test.
How to Identify the Correct p-value:

◻ In the “Independent Samples Test” output, look

under the section labeled “t-test for Equality of
Means.”
◻ Focus on the column labeled “Sig. (2-tailed)” —
this gives the p-value for the t-test.
◻ Do not confuse this with the column labeled “Sig.”
in the Levene’s Test for Equality of Variances
section, which tests for equal variances but is not
the p-value for the t-test itself.
Levene’s Test for Equality of
Variances (choosing correct row)
◻ Levene’s Test for Equality of Variances is a test
that is used to check whether the variances (spread
of data) of two groups are equal or not.
◻ Equal Variances Assumed
◻ This assumes that the variability (spread) of data is
equal for both groups (weights is similar for
European and Japanese vehicles).
◻ If the Sig. (p-value) from Levene’s test is greater
than 0.05, it indicates no significant difference in
variances, so we assume the variances are equal. In
such cases, use the “Equal variances assumed”
row to interpret the t-test results.
Read from the top row.
◻ If the Sig. Value (p-value) is Less Than 0.05
◻ If the p-value is ≤ 0.05, it means there is a
significant difference in the variances, so the
assumption of equal variances is violated.
◻ In this case, you use the “Equal variances not
assumed” row, which adjusts for this difference
and gives more reliable results.
Interpretation of the Results
◻ The p-value (Sig. 2-tailed) in the second row is
0.002.
◻ This indicates that the difference in mean weights
between European and Japanese cars is statistically
significant at the 0.1, 0.05, and 0.01 levels.
◻ This means there is strong evidence that the
difference in average weights is not due to random
chance.