Week 8: Inferential Statistics: Test of Independence

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 4

WEEK 8: INFERENTIAL STATISTICS: indicator of how successfully diabetes is being controlled.

TEST OF INDEPENDENCE Having the test performed regularly is important, and is


therefore a valid indicator of healthcare quality in
diabetes. We can calculate that 64.6% (128/198) of
Asians received the check, compared with 74.7%
(430/576) of non- Asians. As such we know that a lower
proportion of Asian patients was checked, but is there a
significant association between ethnicity and receiving
the check? Our null hypothesis is that there is no
association between ethnicity and receiving a HbA1c
check.

Characteristics of Chi-Square Distribution

1. The computed chi-square is always positive.


2. The shape of the chi-square distribution does not
depend on the size of the sample.
3. The chi-square is positively skewed
4. It is applied for nominal and ordinal data.

Assumptions in Chi-Square
The frequencies for Asian/non- Asian patients with
1. Subjects are randomly selected diabetes are assembled in a 2 × 2 table and tabulated
2. Categories are mutually exclusive against the frequencies in each group of patients who
have/have not received the HbA1c test, as shown in
Procedures in Chi-Square Table 20.2.
1. Set up the hypothesis for Chi-square test. Procedures in Chi-Square
2. Compute for the degree of freedom
3. Select the critical region based on the critical Step 1. Set up the hypothesis for Chi-square test.
value for Chi-square test based on the set level
• H0: There is no significant association with
of significance.
ethnicity and receipt of HBA1c checking.
4. Compute for the expected frequencies
5. Compute for the value of Chi-square using the • HA: There is a significant association with
formula: ethnicity and receipt of HBA1c checking.
6. Conclude according to statistical decision for
hypothesis testing

SAMPLE PROBLEM

A study asks whether Asians with diabetes receive worse


treatment in primary care than non- Asians with diabetes.
This is important, since Asians are more likely to develop
diabetes than non- Asians. A number of variables are
studied, including whether patients with diabetes have
received a HbA1c test within the previous year (we
mentioned HbA1c in Chapter 7), as this is a valuable
Step 5. Calculate the t-value and determine the statistical
INFERENTIAL STATISTICS: TEST OF CORRELATION decision for hypothesis testing
Pearson’s Product-Moment Correlation Test
Step 6. State the conclusion
• In statistics, correlation assesses the strength of
association between variables (usually interval or
ratio).
• AKA Pearson’s r (as in the Greek letter rho “ρ”).
• It is a parametric statistical test used to measure
the degree of relationship (correlation coefficient)
between two sets of data.

SAMPLE PROBLEM

A rheumatologist measures and records the bone mineral


density (BMD) in a group of women. She has a
hypothesis that BMD decreases with age, and decides to
• The value of correlation coefficient scales from -1
use correlation and linear regression to explore this. The
(perfect negative correlation) to +1 (perfect
data collected by our consultant rheumatologist are
positive correlation), with 0 indicating that there is
shown in Table 18.1.
no correlation between two data set.

Pearson’s Product-Moment Correlation Test

The table below lists the correlation coefficients and the


corresponding degree or strength of correlation:

Procedures in Pearson’s r

Step 1. Set up the hypotheses

Step 2. Set the level of significance

Step 3. Calculate the degrees of freedom

Step 4. Calculate the value of Pearson’s r using the


formula
Procedures in Pearson’s r

Step 1. Set up the hypothesis for Pearson’s r

H0: ρ = 0

There is no significant relationship between the data sets.


Or it could be stated otherwise as.

The correlation is zero

HA: ρ > 0, ρ < 0, ρ ≠ 0

There is a significant relationship between the data sets

The correlation is not zero

Step 2. Level of significance: α = 0.05

Step 3. Calculate the degrees of freedom

(df = n – 2)

df = 10 – 2

df = 8

refer to the table for the t-critical.

t-critical = ±2.306

Step 4. Compute for the value of Pearson’s r using a


formula given:

You might also like