BASICS OF
STATISTICAL INFERENCE
1
SAMPLE VS. POPULATION
In the context of educational research, the sample is
the group of individuals who actually participate in
your study. For example, the pre-course learning
need questionnaire sample…
The population is the broader group of people to
whom you intend to generalize the results of your
study. A sample is a subset of a population.
Population has unknown parameters (mean,
variance, etc.) while sample provides estimates of
2
the parameters.
SAMPLING DISTRIBUTION
A sampling distribution or finite-sample distribution
is the probability distribution of a given statistic
based on a random sample.
Example — Assume we repeatedly take samples of a
given size from a normal population. The distribution
of these means is called the sampling distribution
of the sample mean.
3
STANDARD ERROR
Standard error = Standard deviation of the sampling
distribution of a statistic (e.g. mean)
Reflect the accuracy with which a sample represents
a population. The smaller the standard error, the
more representative the sample will be of the
population.
Inversely proportional to the sample size. The larger
the sample size, the smaller the standard error.
4
STANDARD DEVIATION VS.
STANDARD ERROR
Standard deviation — A representation of the spread
of each of the data points
Standard error — To determine the accuracy of the
sample or the accuracy of multiple samples by
analyzing deviation within the means
5
NORMAL DISTRIBUTION
Assumption of many statistical tests
Mean and variance are independent from each other
Distribution approach the “bell” curve with large
sample (
[Link]
The video is also posted under additional
resources
6
CENTRAL LIMIT THEOREM
As you take more samples, especially large ones,
the distribution of the sample means will approach
normal distribution.
Dice demo (
[Link]
The video is also posted under additional
resources
7
BINOMIAL DISTRIBUTION
Discrete probability distribution of the number of
successes in a sequence of n independent
experiments each asking a yes/no, success/failure,
true/false outcome
Parameters n and p
Each single trial is called a Bernoulli trial
8
HYPOTHESIS TESTING
Hypothesis testing helps decide whether the tested
ideas are probably true or false. The conclusions are
never made with 100% confidence.
What is a hypothesis?
9
NULL HYPOTHESIS
Denoted as H0
A statement on the value of a population parameter,
usually the mean value
Initially regarded to be true until the evidence
rejects the null hypothesis
1
0
ALTERNATIVE HYPOTHESIS
Denoted as Ha
A statement that will be accepted when the null
hypothesis is rejected
We do not directly test the alternative hypothesis
1
1
TYPE I ERROR
When a null hypothesis is rejected even though it is
true (should not be rejected) = False positive
False positive is serious! Usually we control it at 5%
including both tails
A type error I would cause the appearance that a
treatment for a disease has the effect of reducing
the severity of the disease when in fact it does
not.
1
2
TYPE II ERROR
When a null hypothesis is accepted when actually
the alternative hypothesis is true
Equal to 1 minus power of the test
Larger sample size -> Larger power of test ->
Smaller Type II error
1
3
ILLUSTRATION
1
4
DETERMINING SIGNIFICANCE
Significance represented by p, which stands for
probability and is the probability of Type I error
Usually 5%, or sometimes 1%
1
5
CONFIDENCE INTERVAL
1
6
SPSS ACTIVITY
One sample t-test — to determine whether a sample
of observations could have been generated by a
process with a specific mean
USU students’ Step 3 score equal 215?
1
7
STEPS OF HYPOTHESIS
TESTING PROCEDURE
What are the null hypothesis and the alternative hypothesis?
Which test statistic is appropriate, and what is the probability
distribution?
What is the required level of significance?
What is the decision rule?
Based on the sample data, what is the value of the population
statistic?
Do we reject or retain the null hypothesis?
Based on our decision, what is the conclusion?
1
8
STATISTICAL POWER
The probability that the test correctly rejects the null
hypothesis when the alternative hypothesis is true
Range from 0 to 1. As power increases, type II error
decreases.
Can be used to calculate the minimum sample size
required so that one can reasonably likely to detect
an effect of a given size. Often set power to 80%.
1
9
STATISTICAL POWER
ILLUSTRATION
2
0
EFFECT SIZE
A measure of the magnitude of a phenomenon
For most types of effect size, a larger absolute value
indicates a stronger effect (Small: <.30; Medium:
between .30 and .50; Large: >.50)
Reporting effect sizes is considered good practice
when presenting results of inferential statistics in
addition to p value
2
1
TWO-TAILED VS ONE-TAILED
TEST
Two tailed test: To show whether the mean of the sample is
significantly greater than and significantly less than the mean of a
population.
H0: u = 230
Ha: u =\ 230
One tailed test: To show the sample mean would be higher or
lower than the population mean.
H0: u < = 230
Ha: u > 230
2
2
TUTORIAL DATA WRITE-UP
Mean Step 3 score from USUHS (M = 212.91, SD =
14.68) was lower than the national Step 3 score of
230, a statistically significant mean difference of
2.09, t(653) = -3.649, p <.001.
2
4