Statistical Treatment: Tests of Difference

Parametric Tests

1
Introduction

Purpose: Tests of difference compare two or more groups to

determine if their means or distributions differ significantly.

Relevance to Nursing: In nursing research, these tests are

essential for comparing treatment effects, patient outcomes, and
other clinical variables.

2
Parametric Tests
Definition: Tests that assume data follows a specific
distribution, typically a normal distribution.
• Assumptions:
• Normality: Data should follow a normal distribution (bell-shaped curve).
• Homogeneity of Variance (Homoscedasticity): Variability within each
group should be similar.
• Sample size and Randomization: The sample size is large (a minimum
of 30) and have been randomly selected from a population
• Interval or Ratio Data: Dependent variables should be continuous, not
categorical.
3
 • Examples of Parametric Tests:
o t-test:

▪ Independent t-test: Compares means of two independent groups

(e.g., comparison of blood pressure between male and female
patients).
• Assumptions:

Parametric Independent variable is categorical (i.e., nominal or

ordinal) and has exactly two categories in addition to the
Tests assumptions of other parametric tests.
▪ Paired t-test: Compares means within the same group at two
different time points (e.g., before and after treatment in the same
group of patients).
▪Assumptions:

▪The independent variable must have two groups that have

“matched pairs”; means that we are measuring the same persons
twice. This in addition to the assumptions of other parametric tests.
▪Homogeneity of Variance can be tested using Levene’s test. If the
test is not significant (p-value= >.05) meaning there is homogeneity of
variances, and we can proceed. If the test is significant (p-value=
≤.05) , an alternative test would be to conduct. 4
 Research Problem
Recruited twenty pupils. Ten took
part in the school counseling
program, while the remaining ten
did not.

The level of life satisfaction will be

measured to determine whether
there are variances between the
two groups.
MDM 5
Nov. ‘24
 Experienced Did not Experienced
counselling i counselling i
DATA

MDM 6
Nov. ‘24
 Step 1: Compute the Mean from Each Group
Experienced Did not Experience
counselling i counselling i

DATA

MDM 7
Nov. ‘24
 Step 2: Compute the Variance from Each Group

Experienced Did not Experience

counselling i counselling i

DATA

MDM 8
Nov. ‘24
 Step 2: Compute the Variance from Each Group

MDM 9
Nov. ‘24
 Step 3: Compute the Standard Error of Difference Between Means

MDM 10
Nov. ‘24
 Step 4: Compute the t- statistics

MDM 11
Nov. ‘24
 Kingdom of Saudi Arabia
Ministry of Education
Shaqra University
College of Nursing- Dawadmi

Step 5: Find the Critical Value

MDM 12
Nov. ‘24
MDM 13
Nov. ‘24
 Step 5: Compare the t- statistics (computed Value) to the critical value
(table value)

IF the value of T is greater than the critical value in a t-table,

reject the Ho

CONCLUSION: There is significant difference

between the two group
MDM 14
Nov. ‘24
 Obtaining an Independent t-Test Using SPSS
Research Question: Is there a difference in weight
lose between male and female after 3 months?
H0: There is no significant difference in weight
loss between male and female.
H1: There is a significant difference in weight loss
between male and female.

Checking of assumptions :
• The total sample size is greater than 30 (n 32).
• The grouping variable is “gender,” and it has
two levels, female=1 and male= 2.
• The dependent variable of interest is the
“wtlosslb” which is pounds of weight lost in
the past 3 months” (a ratio-level variable).
• Weight loss is normally distributed.
15
Obtaining an Independent t-Test Using SPSS

16
Obtaining an Independent t-Test Using SPSS

17
 Obtaining an Independent t-Test Using SPSS
The SPSS output has two parts. The first part of the output provides the
descriptive statistics of the two groups.
The women (group 1) had a mean weight loss of 12.13 lb (SD, 5.462),
and the men had a mean weight loss of 18.59 lb (SD, 6.021).

The second part of the output provides statistics that are critical to
determine the difference between two means.
Levene’s test for equality of variances test helps researchers choose the
proper independent t-test. when the equality of variances is assumed (e.g.,
homogeneity of variance) or when equal variances are not assumed (e.g.,
heterogeneity of variance).

The actual p-value is .782 (and greater than the -level of .05), it can be
determined that the variances of the two groups are not significantly
different and that there is homogeneity of variance (the assumption of
homogeneity of variance is met). Therefore, the row labeled “Equal
variances assumed” is examined to obtain the independent t-test value.
18
 Obtaining an Independent t-Test Using SPSS
Result reporting:

The results were [t(30) = -3.159, p = .004] . The

p-value is less than .05, indicating a significant
difference in weight lose between males and
females. Males (M=18.59, SD=6.021) lost
significantly more weight than women (M=12.13,
SD=5.462). The researchers are 95% sure that
the true wight loss difference between males
and females after 3 months ranged from [-
10.627 to -2.282].

19
 • Examples of Parametric Tests:
o ANOVA (Analysis of Variance):
▪ Compares means among three or more groups (e.g.,

Parametric ▪
comparing the efficacy of three different medications).
One-Way ANOVA: Compares means of three or more

Tests •
independent groups.
Assumptions:
Independent variable is categorical (i.e., nominal or
ordinal) and has exactly more than two categories in addition to
the assumptions of other parametric tests.

20
PROCEDURE FOR CALCULATION OF ONE WAY ANOVA
Procedure for calculation of one way ANOVA
1. Obtain the mean of each sample.
2. Take out the average mean of the sample means.
3. Calculate sum of squares for variance between the sample SSB(SB²)
4. Obtain variance for mean square (MS) between samples.
5. Calculate sum of squares for variance within samples SSW(SW²)
6. Obtain the variance or means square (MS) within samples.
7. Find sum of squares or deviation for total variance
8. Finally, find F- ratio
 SAMPLE TABLE FOR ONE WAY ANOVA CALCULATION

DATA MILD MODERATE SEVERE

FEVER
FEVER FEVER

FLUE
VACCINE:
RANGE OF
FEVER

MDM
Nov. ‘24
Step 1. Compute the Mean of the group

(37.3 to 38.0 C ) 38.1 to 39.0 C) (39.1 to 41 C)

MDM
Nov. ‘24
 Step 2. Complete the table below
X1 (37.3 to X2 (38.1 X3 (39.1
38.0 C ) to 39.0 to 41 C)
C)

MDM
Nov. ‘24
Step 2. Complete the table below
X1 (37.3 to X2 (38.1 X3 (39.1
38.0 C ) to 39.0 to 41 C)
C)
 Step 3. Complete the ANOVA Table

MDM
Nov. ‘24
 Step 3. Complete the ANOVA Table

MDM
Nov. ‘24
 Step 3. Complete the ANOVA Table

MDM
Nov. ‘24
 Step 3. Complete the ANOVA Table

MDM
Nov. ‘24
 Step 3. Complete the ANOVA Table

MDM
Nov. ‘24
 Step 4. Find the Tabular Value of F

MDM
Nov. ‘24
 Step 5. Compare the F - statistic (Computed Value)
with the Critical Value ( from the Table )

IF the value of F is greater than the value of table F,

reject the Ho

CONCLUSION: There is significant difference in the ranges of fever

between the three group
MDM
Nov. ‘24
 Obtaining One-Way ANOVA Using SPSS
Research question: is there a difference in the number of visits to the physician among people with different types of
insurance?

Variables:
The independent/grouping variable is “insure” which is the type of insurance,” and it has four categories, (1= no
insurance, 2= private, 3=Medicare, 4= TRICARE
The dependent variable is the “mdvisits” which is number of physician visits in the past year”.

Hypothesis
H0: There is no significant difference in the number of visits to the physician among people with different types of
insurance. OR there is no significant difference in the number of visits to the physician among no insurance,
Medicare, TRICAR, and private insurance groups
H1: There is a significant difference in the number of visits to the physician among people with different types of
insurance. OR there is a significant difference in the number of visits to the physician among no insurance, Medicare,
TRICAR, and private insurance groups.

After entering data into SPSS and checking the assumptions (MET), we run the one-way ANOVA test as follow: 33
34
 SPSS output (One-Way ANOVA)

“Test of Homogeneity of Variances.” Levene’s statistic is

nonsignificant because the p-value is .451 (above our -level of
.05), indicating that homogeneity of variance is present

The output of the ANOVA analysis shows a statistically

significant difference between group means (F(2,27) =
4.467, p = .021). We can see that the significance value is
0.021, which is below 0.05. and, therefore, there is a
statistically significant difference in the mean the number of
visits to the physician between the insurance groups.
To know which of the specific groups differed, look at the Post
Hoc or Multiple Comparisons table.

35
The table shows that group with no insurance made
significantly fewer physician visits (2.57 a year) than
those with private insurance (4.46 visits a year, p =
.001), Medicare (4.19 visits a year, p = .005), or
TRICARE (4.08 visits a year, p = .013), and no other
differences were found between the groups in number
of outpatient visits.

36