1. Consider the following partially incomplete Analysis of Variance.

Source SS df MS Fstats (0.01)

Treatments 12.5138 B C E F
Errors A 12 D
Total 15.8333 14

1.1. Complete the table.
1.2. Test for the difference among the treatment means at α = 0.01.
1.3. If it’s known that n_1=4; n_2=6; ( Y) ̅_1=8.875; Y ̅_2=6.7. Test for the difference between the two treatment means, by constructing a 99% confidence interval for the mean differences.

2. A study was done on 15 renal dialysis patients, to assess oral home care practices. The patients were divided into 3 groups were in (1) dialysis was for less than 1 year, (2) dialysis was for 1 – 3 years, and (3) dialysis was for more than 3 years. The table below shows plagues index scores for these subjects. A higher value indicates a greater amount of plague.

Group Observations Ti (T_i^2)/n
1 2.33 1.50 2.67 1.17 7.670 D
2 2.83 2.67 2.17 1.83 1.67 3.00 A E
3 2.17 2.33 2.50 2.83 2.67 B F
C G

∑_all▒Y_ij^2 =? ; CM= ? n= ?

2.1. Complete the table and test at 5% significance level if there’s a difference amongst the 3 groups.

1 answer

Let's break down the problem step-by-step.

1. Analysis of Variance Completion

1.1: Complete the ANOVA Table

Given Information:

  • SS Total = 15.8333
  • SS Treatments = 12.5138
  • SS Errors = A
  • df Total = 14
  • df Treatments = B
  • df Errors = D
  • MS Treatments = E
  • F-statistic = F

Step 1: Find Degrees of Freedom

For ANOVA:

  • \( df_{Total} = n - 1 \) (Where \( n \) is the total number of observations)
  • \( df_{Treatments} = k - 1 \) (Where \( k \) is the number of treatment groups)
  • \( df_{Errors} = df_{Total} - df_{Treatments} \)

Due to \( df_{Total} = 14 \) and assuming there are 3 treatments, we can set:

  • \( k = 3 \)
  • \( df_{Treatments} = k - 1 = 3 - 1 = 2 \)

Thus:

  • \( df_{Errors} = df_{Total} - df_{Treatments} = 14 - 2 = 12 \)

Step 2: Find Sum of Squares for Errors

Using the Total SS: \[ SS_{Errors} = SS_{Total} - SS_{Treatments} \] \[ A = 15.8333 - 12.5138 = 3.3195 \]

Step 3: Calculate Mean Squares

Mean Square (MS) is calculated as: \[ MS = \frac{SS}{df} \]

Calculate \( MS_{Treatments} \) and \( MS_{Errors} \): \[ MS_{Treatments} = \frac{SS_{Treatments}}{df_{Treatments}} = \frac{12.5138}{2} = 6.2569 \] \[ MS_{Errors} = \frac{SS_{Errors}}{df_{Errors}} = \frac{3.3195}{12} = 0.2766 \]

Step 4: Calculate F-statistic

\[ F = \frac{MS_{Treatments}}{MS_{Errors}} \] \[ F = \frac{6.2569}{0.2766} \approx 22.62 \]

Completed Table:

\[ \begin{array}{|c|c|c|c|c|c|} \hline \text{Source} & \text{SS} & \text{df} & \text{MS} & \text{F} & \text{(0.01)} \ \hline \text{Treatments} & 12.5138 & 2 & 6.2569 & 22.62 & E \ \text{Errors} & 3.3195 & 12 & 0.2766 & & \ \text{Total} & 15.8333 & 14 & & & \ \hline \end{array} \]

1.2: Test for Difference among Treatment Means

Using degrees of freedom \( df_{Treatments} = 2 \) and \( df_{Errors} = 12 \), we can determine the critical F-value at \( \alpha = 0.01 \). Typically from F-distribution tables, the critical value for \( F(2,12) \) at \( \alpha = 0.01 \) is approximately 5.676.

Since \( F \approx 22.62 \) > 5.676, we reject the null hypothesis. This indicates that there is a statistically significant difference among the treatment means.

1.3: 99% Confidence Interval for Mean Differences

To construct a confidence interval for the difference between two treatment means, we use the formula for the confidence interval:

\[ CI = ( \bar{Y}1 - \bar{Y}2 ) \pm t{(n_1 + n_2 - 2, 0.01/2)} \times \sqrt{ \frac{MS{Errors}}{n_1} + \frac{MS_{Errors}}{n_2} } \]

Here:

  • \( n_1 = 4, n_2 = 6 \)
  • \( \bar{Y}_1 = 8.875, \bar{Y}_2 = 6.7 \)

Calculate the difference of means: \[ \bar{Y}_1 - \bar{Y}_2 = 8.875 - 6.7 = 2.175 \]

Calculate \( t \)-value at \( \alpha = 0.01/2 = 0.005 \) for \( df = 8 \) (considering \( n_1+n_2-2 = 4+6-2 = 8 \)): From t-tables, \( t_{(8, 0.005)} \approx 4.501 \).

Calculate the standard error: \[ SE = \sqrt{ \frac{MS_{Errors}}{n_1} + \frac{MS_{Errors}}{n_2} } = \sqrt{ \frac{0.2766}{4} + \frac{0.2766}{6}} \] Calculating the numerical value, \[ = \sqrt{ 0.06915 + 0.0461 } = \sqrt{ 0.11525 } \approx 0.3395 \]

Now, construct the CI: \[ CI = 2.175 \pm 4.501 \cdot 0.3395 \]

Calculating the margin of error: \[ = 4.501 \cdot 0.3395 \approx 1.528 \] Thus the confidence interval is: \[ CI = 2.175 \pm 1.528 \Rightarrow (0.647, 3.703) \]

2. Complete the Table for the Study of Renal Dialysis Patients

Step 1: Group Total and Means

Group 1:

  • Observations: 2.33, 1.50, 2.67, 1.17
  • Calculate \( T_1 = 2.33 + 1.50 + 2.67 + 1.17 = 7.67 \)
  • \( T_i^2/n \) can be calculated, but we need the number of observations \( n \) first.

Group 2:

  • Observations: 2.83, 2.67, 2.17, 1.83, 1.67, 3.00
  • \( T_2 = 2.83 + 2.67 + 2.17 + 1.83 + 1.67 + 3.00 = 14.17 \)
  • Need \( n = 6 \) for number of observations.

Group 3:

  • Observations: 2.17, 2.33, 2.50, 2.83, 2.67
  • \( T_3 = 2.17 + 2.33 + 2.50 + 2.83 + 2.67 = 12.50 \)
  • \( n = 5 \) for number of observations.

Total Calculation

  • Sum of Observations \( n_{total} = 4 + 6 + 5 = 15 \)

Step 2: Compute \( \sum Y_{ij}^2 \)

Compute the sum of squares for each group: \[ \sum Y_i^2 = 2.33^2 + 1.50^2 + 2.67^2 + 1.17^2 + 2.83^2 + 2.67^2 + 2.17^2 + 1.83^2 + 1.67^2 + 3.00^2 + 2.17^2 + 2.33^2 + 2.50^2 + 2.83^2 + 2.67^2 \] Calculating the squares:

  • Group 1: 5.4289 + 2.25 + 7.1289 + 1.3689 = 16.1767
  • Group 2: 8.4489 + 7.1289 + 4.7089 + 3.3489 + 2.7889 + 9.0000 = 35.4235
  • Group 3: 4.7089 + 5.4289 + 6.25 + 7.9889 + 7.1289 = 31.5056 \[ \sum Y_{ij}^2 = 16.1767 + 35.4235 + 31.5056 = 83.1058 \]

Step 4: Complete the Table:

\[ \begin{array}{|c|c|c|} \hline \text{Group} & \text{Observations (n)} & \frac{T_i^2}{n} \ \hline 1 & 4 & \frac{7.67^2}{4} = 14.67 \ 2 & 6 & \frac{14.17^2}{6} = 33.33 \ 3 & 5 & \frac{12.50^2}{5} = 31.25 \ \hline \end{array} \]

2.1: Test for Difference Among Groups

You will conduct ANOVA to compare these three groups.

  1. Calculating the Group Means and overall mean.
  2. Calculate Between Group and Within Group sums of squares.
  3. Conduct the necessary calculations for F-statistic.

If your F-statistic exceeds the critical value (for \(\alpha = 0.05\) in a group of 3, typically \(F(2,12)\)) from the F-distribution table, you reject the null hypothesis, indicating a difference among the groups.