Statistical & Data Analysis for Nursing

Research

Statistical Inferences
Hypothesis Testing
Statistical Inference

◼ Descriptive statistics—for describing samples

◼ Inferential statistics—for coming to
conclusions about what is probably true in a
population, based on sample values
◼ Inferential statistics uses the laws of
probability to provide guidance on what is
probably true
Hypothesis Testing
◼ What do you know about it ?
Hypotheses
◼ Hypotheses articulate the expected
relationships between variables.
◼ Stem directly from the research questions
◼ Grounded in theory or conceptual models
◼ Tested using data and inferential statistics
◼ Testable hypotheses define
◼ The groups that are being compared
◼ The variables on which the groups are being
compared
◼ The expected relationship
Hypotheses (cont.)
◼ Types of relationships
◼ Association
◼ No causal effect is presumed.
◼ Causal
◼ The independent variable is said to cause changes in
the dependent variable.
Types of Hypotheses

◼ Null hypothesis (H0) states that there will be

no relationship between the two variables.
◼ Alternative hypothesis (Ha) states that there
will be a relationship between the two
variables.
◼ Directional
◼ Nondirectional
Types of Hypotheses: Example
◼ Null hypothesis (H0)
– There will be no relationship between height and
weight in adolescent boys.
◼ Alternative hypothesis (Ha)
◼ Directional: Height will be positively related to
weight in adolescent boys (e.g., taller boys will
weigh more).
◼ Nondirectional: There will be a relationship
between height and weight in adolescent boys.
Hypothesis Testing
◼ Hypotheses are tested with inferential
statistics.
◼ The null hypothesis (H0) is always the
hypothesis that is being tested.
◼ Rejecting the null: This means that the
researchers believe the variables are statistically
associated with one another.
◼ Accepting the null hypothesis (failing to reject
the null): This means that the researchers do
not believe that the variables are statistically
associated with one another.
 Hypothesis Testing (cont.)

◼ The decision to reject or accept the null

hypothesis is based upon two things:
– The p-value (probability value): This is the
probability that the results of the statistical test
were obtained by chance alone. This is
computed from the data and is not known until
the statistical test is completed.
– The α-level (alpha-level): This is p-value that is
defined by the researcher as “statistically
significant.” This is defined before any statistical
tests are conducted. Common α-levels are
“0.10”, “0.05”, and “0.01”.
 Question
◼ Which type of hypothesis is the following
statement?
◼ People who are immunized against the flu
will be less likely to contract the flu than
those who are not immunized.

A. Null hypothesis
B. Directional hypothesis
C. Nondirectional hypothesis
D. An alpha-level hypothesis
Answer

◼ B. Directional hypothesis
◼ Rationale: This hypothesis states that a
relationship between immunization and
contracting the flu is expected and also
states the direction of that hypothesis.
Outcome of Testing Hypotheses

◼ When we test hypotheses with statistics,

one of the four things can happen:
Types of Errors
◼ Type I Error
◼ Reject the null when it is true
◼ In other words, we say that there is a
relationship between the variables when one
really does not exist.
◼ Type II Error
◼ Accept the null when it is false
◼ In other words, we say that there is no
relationship between the variables when one
really does exist.
◼ In any given study, we will never know if we
have committed either one of these errors.
 Errors

◼ Type I Error: reject the null when it is true

◼ The probability of making a type I error is
defined by the α-level of the study.
◼ Given an α-level of 0.10, we will make a type I error
10% of the time.
◼ Type II Error: not rejecting the null when
it is false.
This is referred to as “β” (beta).
Power of a Test

◼ The power of a test is its ability to detect

statistically significant differences.
◼ Mathematically, this is defined as 1-β.
◼ Power is a function of the α-level, the
sample size, and the population effect size.
◼ There are numerous statistical packages
that will compute the power of a study.
 Question

◼ If there is, in fact, a relationship between

immunization and contracting the flu and
the researcher concludes there is no
relationship, what type of error have they
committed?
A. Insufficient power in the study
B. Logical error
C. Type I
D. Type II
 Answer

◼ D. Type II Error
◼ Rationale: A type II error is when we say
that there is no relationship between the
variables, but one really does exist.
 Critical Regions

◼ Researchers calculate a test statistic using

their sample data

◼ They reject the null hypothesis if the test

statistic falls at or beyond a critical region
on the theoretical distribution for their test
statistic; they accept the null hypothesis
otherwise
Critical Region for Two-Tailed Test

◼ The sample mean of 5.5 fell in the upper

critical region for a two-tailed test
 Statistical Significance
◼ When the null hypothesis is rejected, the
results are statistically significant
◼ If the null hypothesis is retained (whenever
p > .05), the results are statistically
nonsignificant
◼ A nonsignificant result reflects an outcome that
could have been obtained as a result of chance
more than five out of 100 times
Meaning of Statistical Significance

◼ A statistically significant result is one that

has a high probability of being “real” in the
population, and probably does not merely
reflect a chance fluctuation
◼ Statistical significance does not mean the
result is important, relevant, or clinically
meaningful
 Six-Steps for Hypothesis Testing
1. State the hypothesis (in both the null and
the alternative form).
2. Specify the test statistic (what test? Have
assumptions been met?)
3. Decision rule (α, direction of hypothesis,
critical values & critical region)
4. Draw & label sampling distribution
Six-Steps for Hypothesis Testing
5. Calculate observed test statistic from data.
6. Interpret results and state a conclusion.
Example: Hypothesis Testing

◼ We are looking at the relationship between

APACHE scores (X) and hospital LOS. We
randomly select our sample of 7 cortical ill
subjects.
◼ Research Question: Is there a linear
relationship between APACHE scores and
hospital LOS for critically ill subjects?
Raw Data & Measures of Central Tendency
APACHE LOS
23 12
21 28
28 13
23 41
33 22
17 34
29 18

M (APACHE): 24.8571, sd: 5.0265

M (LOS): 24.0000, sd: 10.0712
Skewness (APACHE): .127
Kurtosis (APACHE): -.649
Skewness (LOS): .488
Kurtosis (LOS): -1.052
Example Step 1: State the Null and
Alternative Hypotheses.

Step 1: State the hypothesis

◼ H0: What should we write here?

◼ Ha : And here …..????

Step 2: Specify the Test Statistic

#DV LOM # IPV LOM Category What

OF IPV
OF OF Same/
Test?
DV IPV Different?
 Step 2: Have we met the
Assumptions of the Test Statistic
(Pearson r)?
◼ Hold on. We will discuss the assumptions of
each test in this course!!!
◼ Assume we me all of the assumptions
related to correlation.
Step 3: Decision Rule

◼ Which level of α to select?

◼ Directional or nondirectional
◼ Critical values and N
◼ Degree of Freedom
➢ df = N – 2 (degree freedom for r)
Determining Critical Value
Our example:
◼ df = 7 – 2 = 5
◼  = .05
◼ Hypothesis: non-directional
◼ rcv = + .7545
◼ Critical region: r > .7545 OR r < -.7545
Step 4: Draw and Label Sampling
Distribution

◼ Remember that in our example rcv = + .7545

SPSS Output (Step 5)
Step 6: Interpret the Results and
State a Conclusion
◼ In our example:

✓ Fail to reject the null hypothesis

✓ English form: There is no linear relationship
between APACHE score and number of
readmission days.