Download Analysis of Variance (ANOVA) for Different Scenarios and more Exercises Mathematical Statistics in PDF only on Docsity!
Math 143 – ANOVA 1
Analysis of Variance (ANOVA)
Recall, when we wanted to compare two population means, we used the 2-sample t procedures.
Now let’s expand this to compare k ≥ 3 population means. As with the t-test, we can graphically get an idea of what is going on by looking at side-by-side boxplots. (See Example 12.3, p. 748, along with Figure 12.3, p. 749.)
1 Basic ANOVA concepts
1.1 The Setting
Generally, we are considering a quantitative response variable as it relates to one or more explanatory variables, usually categorical. Questions which fit this setting:
(i) Which academic department in the sciences gives out the lowest average grades? (Explanatory vari- able: department ; Response variable: student GPA’s for individual courses )
(ii) Which kind of promotional campaign leads to greatest store income at Christmas time? (Explanatory variable: promotion type ; Response variable: daily store income )
(iii) How do the type of career and marital status of a person relate to the total cost in annual claims she/he is likely to make on her health insurance. (Explanatory variables: career and marital status ; Response variable: health insurance payouts )
Each value of the explanatory variable (or value-pair, if there is more than one explanatory variable) repre- sents a population or group. In the Physicians’ Health Study of Example 3.3, p. 238, there are two factors (explanatory variables): aspirin (values are “taking it” or “not taking it”) and beta carotene (values again are “taking it” or “not taking it”), and this divides the subjects into four groups corresponding to the four cells of Figure 3.1 (p. 239). Had the response variable for this study been quantitative—like systolic blood pres- sure level—rather than categorical, it would have been an appropriate scenario in which to apply (2-way) ANOVA.
1.2 Hypotheses of ANOVA
These are always the same.
H 0 : The (population) means of all groups under consideration are equal.
Ha : The (pop.) means are not all equal. (Note: This is different than saying “they are all unequal ”!)
1.3 Basic Idea of ANOVA
Analysis of variance is a perfectly descriptive name of what is actually done to analyze sample data ac- quired to answer problems such as those described in Section 1.1. Take a look at Figures 12.2(a) and 12.2(b) (p. 746) in your text. Side-by-side boxplots like these in both figures reveal differences between samples taken from three populations. However, variations like those depicted in 12.2(a) are much less convincing that the population means for the three populations are different than if the variations are as in 12.2(b). The
reason is because the ratio of variation between groups to variation within groups is much
smaller for 12.2(a) than it is for 12.2(b).
Math 143 – ANOVA 2
1.4 Assumptions of ANOVA
Like so many of our inference procedures, ANOVA has some underlying assumptions which should be in place in order to make the results of calculations completely trustworthy. They include:
(i) Subjects are chosen via a simple random sample.
(ii) Within each group/population, the response variable is normally distributed.
(iii) While the population means may be different from one group to the next, the population standard deviation is the same for all groups.
Fortunately, ANOVA is somewhat robust (i.e., results remain fairly trustworthy despite mild violations of
these assumptions). Assumptions (ii) and (iii) are close enough to being true if, after gathering SRS samples from each group, you:
(ii) look at normal quantile plots for each group and, in each case, see that the data points fall close to a line.
(iii) compute the standard deviations for each group sample, and see that the ratio of the largest to the smallest group sample s.d. is no more than two.
2 One-Way ANOVA
When there is just one explanatory variable, we refer to the analysis of variance as one-way ANOVA.
2.1 Notation
Here is a key to symbols you may see as you read through this section.
k = the number of groups/populations/values of the explanatory variable/levels of treatment
ni = the sample size taken from group i
xi j = the j th response sampled from the i th group/population.
x ¯ i = the sample mean of responses from the i th group =
ni
ni
j = 1
xi j
si = the sample standard deviation from the i th group =
ni − 1
ni
j = 1
( xi j − x ¯ i )^2
n = the (total) sample, irrespective of groups = (^) ∑ ki = 1 ni.
x ¯ = the mean of all responses, irrespective of groups =
n ∑ i j
xi j
2.2 Splitting the Total Variability into Parts
Viewed as one sample (rather than k samples from the individual groups/populations), one might measure the total amount of variability among observations by summing the squares of the differences between each xi j and ¯ x :
SST (stands for sum of squares total ) =
k
i = 1
ni
j = 1
( xi j − x ¯)^2.
Math 143 – ANOVA 4
- MS = Mean Square = SS/df : This is like a standard deviation. Look at the formula we learned back in Chapter 1 for sample stan- dard deviation (p. 51). Its numerator was a sum of squared deviations (just like our SS formulas), and it was divided by the appropriate number of degrees of freedom. It is interesting to note that another formula for MSE is
MSE =
( n 1 − 1 ) s^21 + ( n 2 − 1 ) s^22 + · · · + ( nk − 1 ) s^2 k ( n 1 − 1 ) + ( n 2 − 1 ) + · · · + ( nk − 1 )
which may remind you of the pooled sample estimate for the population variance for 2-sample pro- cedures (when we believe the two populations have the same variance). In fact, the quantity MSE is also called s^2 p.
- The F statistic = MSG/MSE
If the null hypothesis is true, the F statistic has an F distribution with k − 1 and n − k degrees of freedom in the numerator/denominator respectively. If the alternate hypothesis is true, then F tends to be large. We reject H 0 in favor of Ha if the F statistic is sufficiently large. As with other hypothesis tests, we determine whether the F statistic is large by finding a correspond- ing P -value. For this, we use Table E. Since the alternative hypothesis is always the same (no 1-sided vs. 2-sided distinction), the test is single-tailed (like the chi-squared test). Neverthless, to read the correct P -value from the table requires knowledge of the number of degrees of freedom associated with both the numerator (MSG) and denominator (MSE) of the F -value. Look at Table E. On the top are the numerator df, and down the left side are the denominator df. In the table are the F values, and the P -values (the probability of getting an F statistic larger than that if the null hypothesis is true) are down the left side. Example : Determine
P ( F 3,6 > 9.78) = 0.
P ( F 2,20 > 5 ) = between 0.01 and 0.
Example : A firm wishes to compare four programs for training workers to perform a certain manual task. Twenty new employees are randomly assigned to the training programs, with 5 in each program. At the end of the training period, a test is conducted to see how quickly trainees can perform the task. The number of times the task is performed per minute is recorded for each trainee.
Program 1 Program 2 Program 3 Program 4 9 10 12 9 12 6 14 8 14 9 11 11 11 9 13 7 13 10 11 8
The data was entered into Stata which produced the following ANOVA table.
Analysis of Variance Source SS df MS F Prob > F
Between groups 54.95 3 18.3166667 7.04 XXXXXX Within groups 41.6 16 2.
Total 96.55 19 5.
Bartlett’s test for equal variances: chi2(3) = 0.5685 Prob>chi2 = 0.
The X-ed out P -value is between 0.001 and 0..
Math 143 – ANOVA 5
Stata gives us a line at the bottom—the one about Bartlett’s test—which reinforces our belief that the vari- ances are the same for all groups.
Example : In an experiment to investigate the performance of six different types of spark plugs intended for use on a two-stroke motorcycle, ten plugs of each brand were tested and the number of driving miles (at a constant speed) until plug failure was recorded. A partial ANOVA table appears below. Fill in the missing values: Source df SS MS F
Brand 5 55961.5 11192.3 2.
Error 54 254539.26 4,713.
Total 59 310,500.
Note : One more thing you will often find on an ANOVA table is R^2 (the coefficient of determination ). It indicates the ratio of the variability between group means in the sample to the overall sample variability, meaning that it has a similar interpretation to that for R^2 in linear regression.
2.4 Multiple Comparisons
As with other tests of significance, one-way ANOVA has the following steps:
- State the hypotheses (see Section 1.2)
- Compute a test statistic (here it is F df numer., df denom.), and use it to determine a probability of getting a sample as extreme or more so under the null hypothesis.
- Apply a decision rule: At the α level of significance, reject H 0 if P ( Fk −1, n − k > F computed) < α. Do not reject H 0 if P > α.
If P > α, then we have no reason to reject the null hypothesis. We state this as our conclusion along with the relevant information ( F -value, df-numerator, df-denominator, P -value). Ideally, a person conducting the study will have some preconceived hypotheses (more specialized than the H 0 , Ha we stated for ANOVA, and ones which she held before ever collecting/looking at the data) about the group means that she wishes to investigate. When this is the case, she may go ahead and explore them (even if ANOVA did not indicate an overall difference in group means), often employing the method of contrasts. We will not learn this method as a class, but if you wish to know more, some information is given on pp. 762–769.
When we have no such preconceived leanings about the group means, it is, generally speaking, inappro- priate to continue searching for evidence of a difference in means if our F -value from ANOVA was not significant. If, however, P < α, then we know that at least two means are not equal, and the door is open to our trying to determine which ones. In this case, we follow up a significant F -statistic with pairwise comparisons of the means, to see which are significantly different from each other.
This involves doing a t -test between each pair of means. This we do using the pooled estimate for the (assumed) common standard deviation of all groups (see the MS bullet in Section 2.3):
ti j =
x ¯ i − x ¯ j sp
1 / ni + 1 / nj
To determine if this ti j is statistically significant, we could just go to Table D with n − k degrees of freedom (the df associated with sp ). However, depending on the number k of groups, we might be doing many comparisons, and recall that statistical significance can occur simply by chance (that, in fact, is built into the interpretation of the P -value), and it becomes more and more likely to occur as the number of tests we conduct on the same dataset grows. If we are going to conduct many tests, and want the overall probability of rejecting any of the null hypotheses (equal means between group pairs) in the process to be no more than α, then we must adjust the significance level for each individual comparison to be much smaller than α. There are a number of different approaches which have been proposed for choosing the individual-test
Math 143 – 2-way ANOVA 7
measured. Notice that the design includes nine different treatments because there are three levels to each of the two factors.
Comparisons of Two-way to One-Factor-at-a-Time
- usually have a smaller total sample size, since you’re studying two things at once [ rat diet example, p. 800 ]
- removes some of the random variability (some of the random variability is now explained by the second factor, so you can more easily find significant differences)
- we can look at interactions between factors (a significant interaction means the effect of one variable changes depending on the level of the other factor).
Examples of (potential) interaction.
- Radon (high/medium/low) and smoking. High radon levels increase the rate of lung cancer somewhat. Smoking increases the risk of lung cancer. But if you are exposed to radon and smoke, then your lung cancer rates skyrocket. Therefore, the effect of radon on lung cancer rates is small for non-smokers but big for smokers. We can’t talk about the effect of radon without talking about whether or not the person is a smoker.
- age of person (0-10, 11-20, 21+) and effect of pesticides (low/high)
- gender and effect of different legal drugs (different standard doses)
Two-way ANOVA table
Below is the outline of a two-way ANOVA table, with factors A and B, having I and J groups, respectively.
Source df SS MS F p-value A I − 1 SSA MSA MSA/MSE B J − 1 SSB MSB MSB/MSE A × B ( I − 1 )( J − 1 ) SSAB MSAB MSAB/MSE Error n − I J SSE MSE Total n − 1 SST
Math 143 – 2-way ANOVA 8
The general layout of the ANOVA table should be familiar to us from the ANOVA tables we have seen for regression and one-way ANOVA. Notice that this time we are dividing the variation into four components:
1. the variation explained by factor A
2. the variation explained by factor B
3. the variation explained by the interaction of A and B
4. the variation explained by randomness
Since there are three different values of F , we must be doing three different hypothesis tests at once. We’ll get to the hypotheses of these tests shortly.
The Two-way ANOVA model
The model for two-way ANOVA is that each of the I J groups has a normal distribution with potentially different means ( μ i j ), but with a common standard deviation ( σ). That is,
xi jk = μ i j ︸︷︷︸ group mean
, where i jk ∼ N (0, σ)
As usual, we will use two-way ANOVA provided it is reasonable to assume normal group distributions
and the ratio of the largest group standard deviation to the smallest group standard deviation is at most 2.
Main Effects
Example. We consider whether the classifying by diagnosis (anxiety, depression, CDFS/Court referred) and prior abuse (yes/no) is related to mean BC (Being Cautious) score. Below is a table where each cell contains he mean BC score for people who were in that group.
Diagnosis Abused Not abused Mean Anxiety 24.7 18.2 21. Depression 27.7 23.7 26. DCFS/Court Referred 29.8 16.4 20. Mean 27.1 19.
Here is the ANOVA table:
df SS MS F p-value Diagnosis 2 222.3 111.15 2.33. Ever abused 1 819.06 819.06 17.2 .0001* D * E 2 165.2 82.60 1.73. Error 62 2958.0 47. Total 67
Math 143 – 2-way ANOVA 10
Interaction Effects
Example. We consider whether the mean BSI (Belonging/Social Interest) is the same after classifying people on the basis of whether or not they’ve been abused and diagnosis.
Diagnosis Abused Not abused Mean
Anxiety 27.0 26.8 29. Depression 27.3 31.7 32. DCFS/Court Referred 23.0 37.1 26.
Mean 26.6 31.
ANOVA table:
df SS MS F p-value Diagnosis 2 118.0 59.0 1.89. Ever abused 1 483.6 483.6 15.5 .0002* D X E 2 387.0 193.5 6.19 .0035* Error 62 1938.12 31. Total 67
Since the interaction is significant, let’s look at the individual mean BSI at each level of diagnosis on an interaction plot.
Knowing how many people were reflected in each category (information that is not provided here) would allow us to conduct 2-sample t tests at each level of diagnosis. Such tests reveal that there is a significant difference in mean BSI between those who have ever been abused and those not abused only for those who have a DCFS/Court Referred disorder. There is no statistically significant difference between these two groups for those with a Depressive or Anxiety disorder (though it’s pretty close for those with an Anxiety disorder).
Math 143 – 2-way ANOVA 11
Example. Promotional fliers. [Exercise 13.15, p. 821 in Moore/McCabe]
Means: discount promos 10 20 30 40 1 4.423 4.225 4.689 4. 3 4.284 4.097 4.524 4. 5 4.058 3.890 4.251 4. 7 3.780 3.760 4.094 4.
Standard Deviations:
discount promos 10 20 30 40 1 0.18476 0.38561 0.23307 0. 3 0.20403 0.23462 0.27073 0. 5 0.17599 0.16289 0.26485 0. 7 0.21437 0.26179 0.24075 0.
Counts: discount promos 10 20 30 40 1 10 10 10 10 3 10 10 10 10 5 10 10 10 10 7 10 10 10 10
Sum Sq Df F value Pr(>F) promos 8.36 3 47.73 <2e- discount 8.31 3 47.42 <2e- promos:discount 0.23 9 0.44 0. Residuals 8.41 144
3.^ 4.^ 4.^ 4.^ 4.^
promos
mean of expPrice
1 3 5 7
discount (^4030) (^1020)
3.^ 4.^ 4.^ 4.^ 4.^
discount
mean of expPrice
10 20 30 40
promos (^13) (^57)
Question : Was it worth plotting the interaction effects, or would we have learned the same things plotting only the main effect?
