14.6 – Some other ANOVA designs

Introduction

There are several additional ANOVA models in common use. The crossed, balanced design is but one example of the two-way ANOVA. And, from a consideration of two factors it logically follows that there can be more than two factors as part of the design of an experiment. As the number of factors increase, the number of two-way, three-way, and even higher-order interactions are possible and at least in principle may be estimated.

Our purpose here is to highlight several, but certainly not all possible experimental designs from the perspective of ANOVA. Examples are provided. Keep in mind that the general linear model approach unifies these designs.

Some of the classical experimental ANOVA designs one sees include

  • Two-way randomized complete block design
  • Two-way factorial with no replication design
  • Repeat-measures ANOVA with one factor
  • Nested ANOVA
  • Three-way ANOVA
  • Split-plot ANOVA
  • Latin squares ANOVA

Put simply, these designs differ in how the groups are arranged and how members of the groups are included.

Two-way randomized complete block design

This design refers to the “textbook” design. For each, factor A and factor B, there are multiple levels, in this example three levels of Factor A and three levels of Factor B,  and subjects (sampling units) are randomly assigned to each level. However, one of the factors is, perhaps, of less interest, yet certainly accounts for variation in the response variable.

Factor B
1 2 3
Factor A 1 n_{1,1} n_{1,2} n_{1,3}
2 n_{2,1} n_{2,2} n_{2,3}
3 n_{3,1} n_{3,2} n_{3,3}

where n1,1, n2,1, etc. represents the number of subjects in each cell. Thus, in this design there are nine groups. Typically, minimum replication would be three subjects per group.

Two-way factorial with no replication design

While it may seem obvious that a good experiment should have replication, there are situations in which replication is impossible. While this seems rather odd, this scenario very much describes a typical microarray, gene expression project.

Factor B
1 2 3
Factor A 1 n_{1,1} n_{1,2} n_{1,3}
2 n_{2,1} n_{2,2} n_{2,3}
3 n_{3,1} n_{3,2} n_{3,3}

where, again, n1,1, n2,1, etc., represents the number of subjects in each cell and there are nine groups in this study. With no replication, then no more than one subject per group.

Repeated-measures ANOVA

When subjects in the study are measured multiple times for the dependent variable, this is called a repeated-measures design. We introduced the design for the simple case of before and after measures on the same individuals in Chapter 12.3. It’s straight-forward to extend the design concept to more than two measures on the subjects. The blocking effect is the individual (see Chapter 14.4), and, therefore, a random effect (see Chapters 12.3 and 14.3) in this type of experimental design.

Although straightforward in concept, repeated measure designs have many complications in practice. For example, long-term studies can expect for subjects to drop out of the study, resulting in censored data. Another complication, the assumption is that there is no carry over effect — it doesn’t matter the order different treatments are applied to the subjects. Think of this assumption as akin to the equal variances assumption in ANOVA; just like unequal variances effects Type I error rates in ANOVA, deviations from sphericity inflate Type I error rates in repeated-measures designs.

Sphericity assumption is described in two ways:

Assumption of sphericity — the ranking of individuals remains the same across treatment levels — no interaction between individual and treatment. Sphericity assumption is always met if there are just two levels of the repeated measure, eg, before and after.

Compound symmetry assumption — the variances and covariances are equal across the study: the changes experienced by the subjects are the same across the study regardless of the order of treatments.

Tests for sphericity include:

Mauchly test: mauchly.test(object)

If results of tests for violations of sphericity warrant, corrections are available. One recommended correction is called Greenhouse-Giesser correction, which adjusts the degrees of freedom and so results in a better p-value estimate. A second correction is called Huyhn-Feldt correction; this correction, too, adjusts the degrees of freedom to improve the p-value estimate.

Three-way ANOVA

It is relatively straightforward to imagine an experiment that involves three or more factors. The analysis and interpretation of such designs, while feasible, becomes somewhat complicated especially for the mixed models (Model III).

Consider just the case of a fixed-effects 3-way ANOVA. How many tests of null hypotheses are there?

  1. Three tests for main effects.
  2. Three tests of two-way interactions.
  3. A test for a three-way interaction.

Thus, there are seven separate null hypotheses from a three-way ANOVA with fixed effects! As you can imagine, large sample sizes are needed for such designs, and the “higher-order” interactions (eg, three-way interaction) can be difficult to interpret and may lack biological significance.

ANOVA designs without random assignment to treatment levels

Latin square design

We have introduced you to several ANOVA experimental designs that employed randomization for assignment of subjects to treatment groups. The purpose of randomization is even out differences due to confounding variables. However, if we know in advance something about the direction of the influence of these confounding variables, strictly random assignment is not in fact the best design. For example, the Latin square design is common in agriculture research and is very useful for situations in which two gradients are present (eg, soil moisture levels, soil nutrient levels).

Dry soil ←→ Wet soil
Soil Nutrients

low
↑
↓
high

 T1 T4 T3 T2
T3 T2 T1 T4
T2 T3 T4 T1
T4 T1 T2 T3

Split-Plot Design

Another design from agriculture research is especially useful to ecotoxicology research. We mentioned the repeated measures design in which individuals are measured more than once and each individual receives all levels of the treatment in a random order (cross-over design). However, this design assumes that there are no carry-over effects (see Hills and Armitage 1979;  For ecology/evolution definition see O’Connor et al 2014). While this assumption may hold for many experiments, we can also imagine many more situations in which this is undoubtedly false. For example, if we wish to measure the effects of ozone and relative humidity on frog behavior, we might consider using the individual as its own control. But we also wish to compare frog behavior following ozone exposure against behavior exhibited in clean air. But we are likely to violate the carry-over assumption. If a frog receives ozone then air, the effects of ozone may inhibit activity for several days after the initial exposure, which would then influence subsequent measures. The solution to this dilemma is to use what’s called a split plot design. The design combines elements of nesting.

Consider our frog experiment. There would be three factors:

Factor 1 = Exposure (air or ozone),

Factor 2 = Saturation (dry, intermediate, wet),

Factor 3 = Individual (each frog is measured 3 times).

The design table would look like

Exposure
Air Ozone
Humidity Dry Frog1 Frog2 Frog3 Frog4 Frog5 Frog6
Intermediate Frog1 Frog2 Frog3 Frog4 Frog5 Frog6
Wet Frog1 Frog2 Frog3 Frog4 Frog5 Frog6

Thus, the design is crossed for one factor (saturation), but nested for another factor (individuals are nested within Exposure factor).

Questions

  1. Which of the study designs mentioned so far are sensitive to carry-over effects?
  2. With respect to how levels of Factors are assigned, distinguish the split-plot design from the Latin square design.

Quiz Chapter 14.6

Some other ANOVA designs

Chapter 14 contents

10.1 – Compare two independent sample means

Introduction

We introduced the concept of comparing a sample statistic (mean) against a population parameter (Chapter 6.7, Normal deviate) or one-sample t-test against a specified mean (eg, from published data or from theory, Chapter 8.5).

Consider now a basic experimental design, the randomized control trial, or RCT (Fig 1), introduced in Chapter 2.4.

simple RCT

Figure 1. A two group Randomized Control Trial.

Subjects randomly selected from population of interest, then again — random assignment — once recruited into one of two treatment groups. Importantly, subjects belong to one treatment arm only: no subject simultaneously receives the treatment and the control. This is in contrast to the paired design, in which subjects receive both treatments (see Chapter 10.3).

In inferential statistics about an experiment, we are more likely trying to test if sample means are different. For example

  • two species grown in common garden, do they differ in growth rate?
  • human subjects given a new treatment have better outcomes compare to those receiving a control treatment (eg, placebo).

The equivalent null hypothesis is that two samples are pulled from the same population. We write the null hypothesis as

    \begin{align*} H_{0}\ :\ \bar{X}_{1}=\bar{X}_{2} \end{align*}

and the corresponding alternative hypothesis, HA, then must be

    \begin{align*} H_{A}:\bar{X}_{1}\neq \bar{X}_{2} \end{align*}

Question: Is this a one tailed or two-tailed hypothesis?

Answer: Two-tailed (review Chapter 8.4)

Note 1: In this day and age, there’s really no compelling reason to learn the t-test. First, it is just a special case of the one-way ANOVA, therefore, it’s a special case of the general linear model. Struggling to learn R commands? Well, one solution would be just to learn the general linear model approach — the R function lm() (OK — don’t get too excited — lm() has many options and details). Second, few experiments or observational studies are likely to have only two groups; thus, the temptation to carry out a series of t-tests, taking all groups two at a time, or “pairwise,” while tempting, actually violates a whole bunch of basic statistical rules (discussed in Chapter 12.1). It will also make statisticians go crazy when they see it. That said, if your experiment has but two groups, then by all means, the t-test is a choice. The t-test is also a statistical test that you have likely already used before so we present the discussion here to build on what you may already have learned. We also present the independent t-test as a vehicle.

Worked example

We introduce the two-sample t-test, or better, the independent sample t-test.

    \begin{align*} t=\frac{\bar{X}_{1}-\bar{X}_{2}}{s_{\bar{X}_{1}-\bar{X}_{2}}} \end{align*}

where the numerator is the difference between the two sample means and the denominator is the standard error of the differences between the two groups standard errors. The formula for

    \begin{align*} s_{\bar{X}_{1}-\bar{X}_{2}}=\sqrt{ \frac{s_1^2}{n_{1}}+\frac{s_2^2}{n_{2}}} \end{align*}

The choice of independent sample over two-sample is best because it emphasizes that the two groups (the two samples), must be comprise of independent sampling units. This is a pretty straight-forward requirement; you have randomly assigned twenty individuals to two groups, a control group (n = 10) and a treatment group (n = 10). Individuals are either in the control group or they are in the treatment group — they cannot simultaneously appear in both groups.

We will work our way through this test by example. For starters, let’s use the same lizard dataset (see Example data set, below), four body mass recordings (grams) each for house geckos (Hemidactylus frenatus, Fig 2) and the Carolina anole (Anolis carolensis, Fig 3), two of many lizard species introduced to Hawaiʻi.

Male Hemidactylus frenatus, central Oahu, M. Dohm 2018

Figure 2. Male Hemidactylus frenatus, central Oahu, M. Dohm.

Male Anolis carolinensis, ʻAkaka Falls, Big Island of Hawai`i, M. Dohm 2018

Figure 3. Male Anolis carolinensis, ʻAkaka Falls, Big Island of Hawaiʻi, M. Dohm.

Example data set

Geckos: 3.186, 2.427, 4.031, 1.995

Anoles: 5.515, 5.659, 6.739, 3.184

Question: How would you go about creating a data frame with the values in long form (stacked worksheet), including a label variable and the body mass?

Note 2: The independent sample t-test in Rcmdr requires a stacked worksheet and not in unstacked worksheet two columns implied by the above pseudo-code. If you need help with worksheet format, then see Part07 in Mike’s Workbook for Biostatistics.

Answer: At the R prompt, type

Geckos <- c(3.186, 2.427, 4.031, 1.995); Anoles = c(5.515, 5.659, 6.739, 3.184) #create two vectors
bmass <- c(Geckos, Anoles)    #combine the two vectors into a single vector holding all of the body mass records
species <- c("gecko", "gecko", "gecko", "gecko", "anole", "anole", "anole", "anole")   #create your label variable
lizards <- data.frame(species, bmass)  #create your data frame
lizards   #print your data frame  species bmass
1   gecko 3.186
2   gecko 2.427
3   gecko 4.031
4   gecko 1.995
5   anole 5.515
6   anole 5.659
7   anole 6.739
8   anole 3.184

Note also that you can enter data into the Data editor by creating the data frame first then adding values. To edit the data frame “lizards” type fix(lizards) at the R prompt, then close the data frame when you have added or changed values as needed.

As always, begin with an exploration of the data, including a graph (Fig 4).

Box plot of lizard body mass.

Figure 4. Box plot of lizard body mass.

We can see already that there’s greater spread of data for the Anoles compared to the Geckos, but the median values differ. Small sample sizes can be a problem for analyses as we can only have reduced confidence in our conclusions. However, we press on for the sake of demonstration.

Let’s test the null hypothesis, H0, ie, the two species of lizards have the same mean body mass.

Rcmdr: Statistics → Means → Independent-samples t-test…

In this next image I posted the Rcmdr menu popup for the Independent Samples t-test (Fig 5). Later versions of Rcmdr split the settings for this command into two tabs; the first tab allows for the selection of the variables and setting the hypotheses whereas the second tab, labeled Options, permits additional choices. The default selections need your attention: to actually conduct the t-test you need to answer “No” to the question, “Assume equal variance?” (Fig 6).

Screenshot Rcmdr t-test menu

Figure 5. Rcmdr menu for Independent sample t-test.

Select the Options tab (Fig 6) to select null hypothesis and to select the t-test and not the Welch-test (which is the default, ie, No to the prompt “Assume equal variances?”).

Screenshot Rcmdr t-test options

Figure 6. Rcmdr Options menu for Independent sample t-test.

Let’s look at the results and break down the parts of the test.

t.test(Body.mass~Lizard, alternative='two.sided', conf.level=.95, 
+ var.equal=TRUE, data=lizards)

Two Sample t-test

data: Body.mass by Lizard 
t = 2.7117, df = 6, p-value = 0.03503
alternative hypothesis: true difference in means is not equal to 0 
95 percent confidence interval:
0.2308685 4.4981315 
sample estimates:
mean in group Anolis mean in group Gecko 
5.27425 2.90975

Consider the R session output above and answer the following questions.

Questions for the worked example

  1. Which lizard group had the greater mean value, Anolis or Gecko?
  2. What are the assumptions necessary for you to use the independent sample t-test?
  3. What does “two-sided” mean?
  4. What was the null hypothesis?
  5. Was this a one-tailed or two-tailed test of the null hypothesis?
  6. What is the value of the test statistic?
  7. How many degrees of freedom?
  8. What is the critical value for this test?
  9. What is the value of the lower limit of the 95% confidence interval?
  10. What is the value of the lower limit of the 99% confidence interval?
  11. True or False. If the null hypothesis is accepted, then zero is a value included in the 95% confidence interval.
  12. Do you accept the null hypothesis? Explain your selection.

Try another example

DNA damage, changes in the chemical structure of nucleotide bases or breakage of the DNA chains, occurs in cells under many circumstances. The comet assay, or single-cell gel electrophoresis, is one method for visualizing and measuring DNA strand breaks in cells. Exposed cells are mixed with a low-melting temperature agarose and placed onto a microscope slide. The cells are then lysed with an alkaline detergent and high salts. When current is applied across the slide, undamaged DNA remains in the nucleus, whereas damaged DNA extends towards the anode to form a comet-like tail, with imaging assisted by including a fluorescent dye like Sybr-Green. Examples of comets are shown below (Fig 7).

example comet assay

Figure 7. Comet examples. A Intact cell, no DNA damage, B Cell with some DNA damage, a slight tail to the right is evident, C  Cell with significant DNA damage, a large tail is evident. M. Dohm.

In an experiment, immortalized lung epithelial cells were exposed to dilute copper solutions for 30 minutes then washed with PBS. The comet assay was applied to these cells and for comparison, to cells without copper exposure but otherwise treated the same way (controls). The data are available at the bottom of this page (scroll down or click here).

Again, you should begin all analyses with an exploration of the data, including a graph (Fig 8).

box plot comet tails

Figure 8. Boxplot of comet tail lengths for cells with and without (control) exposure to copper in the cell medium for 30 minutes.

Let’s look at the R output for the t-test analysis.

    Two Sample t-test

data:  CometTail by Treatment
t = -5.8502, df = 38, p-value = 9.139e-07
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
 -22.39865 -10.88213
sample estimates:
mean in group Control  mean in group Copper 
             11.14533              27.78571

Consider the R session output above and answer the following questions.

Questions for Comet assay data set

  1. Which cell group had the greater mean value, Copper-exposed or Control-exposed cells?
  2. What are the assumptions necessary for you to use the independent sample t-test?
  3. What does “two-sided” mean?
  4. What was the null hypothesis?
  5. Was this a one-tailed or two-tailed test of the null hypothesis?
  6. What is the value of the test statistic?
  7. How many degrees of freedom?
  8. What is the critical value for this test?
  9. What is the value of the lower limit of the 95% confidence interval?
  10. What is the value of the lower limit of the 99% confidence interval?
  11. True or False. If the null hypothesis is accepted, then zero is a value included in the 95% confidence interval.
  12. Do you accept the null hypothesis? Explain your selection.

T test from summary statistics

In some cases you may only have to summary statistics for data, eg, the means and the standard deviations. We can use the equations of the t test to write a simple formula, where the user provides the known means, standard deviations, and sample size. For example, create a simple function with readline for user input.

myTtest <- function() {
   mnx <- as.numeric(readline(prompt="Enter mean of x: "))
   stdevx <- as.numeric(readline(prompt="Enter sd of x: "))
   nx <- as.numeric(readline(prompt="Enter n of x: "))
   mny <- as.numeric(readline(prompt="Enter mean of y: "))
   stdevy <- as.numeric(readline("Enter sd of y: "))
   ny <- as.numeric(readline(prompt="Enter n of y: "))
   myTvalue <- abs(((mnx-mny)-0)/sqrt(((stdevx^2)/nx)+(stdevy^2)/ny))
   myDF <- as.integer(nx+ny-2)
   myPvalue <- pt(myTvalue,myDF,lower.tail=FALSE)*2
   myResults <- c(myTvalue, myDF, myPvalue)
   report <- c("T-test: ", "df: ", "two-tailed p-value: ")
   cat(sprintf("%s %3.3f, ", report, myResults))
}

then run the function by typing myTest() at the R prompt and entering the means, standard deviations, and sample size when prompted.

myTtest()
Enter mean of x: 2.91
Enter sd of x: .895
Enter n of x: 4
Enter mean of y: 5.27
Enter sd of y: 1.497
Enter n of y: 4
T-test: 2.706, df: 6.000, two-tailed p-value: 0.035

P-values from confidence intervals

While we expect certain reporting criteria for published, it is not uncommon that one or more elements are missing. For example, while means plus or minus standard deviations for two groups may be reported, the confidence interval of the difference may not be reported, or, an inexact p-value is reported like “< 0.05,” but we “need” an exact p-value for our meta-analysis.  A little math, and these missing statistics can be calculated.

We can use the lizard study as an example; while all of the expected elements were reported, including the p-value (0.03503), let’s say the p value was reported as “statistically significant at p < 0.05.”

We need

lower and upper confidence intervals: 0.2309 and 4.498, respectively.

critical value from t distribution, df = 6 (2 tailed, therefore 0.05/2). A little R code we have:

critVal <- qt(c(.05), df=6, lower.tail=FALSE)=2.447

the standard error, which can be calculated as SEliz = (upper − lower)/(2×critVal)=1.097968

and then the p-value, exp(-.717*testLiz-0.416*testLiz^2) = 0.03101

which is pretty close to the result from R (p = 0.03503). The difference is likely due to the small sample size.

Questions

  1. Don’t forget to work through the Questions for the Comet tail data set (scroll up or click here).
  2. Microsoft Excel, LibreOffice Calc, and Google sheets spreadsheet software all include t-test functions and return the p-value. Consider two variables big (100, 110, 120, 100, 110, 210, 200) and small (0,1,1,2,0,1,0). (Note — these two groups are obviously very different, calculating a t-test on their difference is silly, just for this question.)  If formatting is set to the default two decimal places for Number cell category, the p-value will return as “0.00.” How should you report the p-value in this case?

Quiz Chapter 10.1

Compare two independent sample means

Comet assay data set

Treatment CometTail
Control 17.856139
Control 16.52125
Control 14.925449
Control 14.029174
Control 13.332945
Control 8.811185
Control 14.701654
Control 9.261025
Control 21.779311
Control 6.180284
Control 9.201752
Control 5.54472
Control 6.717885
Control 2.625092
Control 7.191583
Control 5.392866
Control 11.284813
Control 15.441254
Control 17.857176
Control 4.250956
Copper 53.214287
Copper 38.92857
Copper 18.928572
Copper 30
Copper 28.928572
Copper 15.357142
Copper 17.857143
Copper 17.5
Copper 21.071428
Copper 29.285715
Copper 28.214285
Copper 16.785715
Copper 21.071428
Copper 37.5
Copper 38.214287
Copper 17.857143
Copper 29.642857
Copper 11.071428
Copper 35
Copper 49.285713

Chapter 10 contents

5.6 – Sampling from Populations

Introduction

Researchers generally can’t study an entire population. More generally, striving to study each member of a population is not necessary to arrive at answers about the population. For example, consider the question, does taking a multivitamin daily improve health? What are our options? Do we really need to follow every single individual in the United States of America, monitoring their health and noting whether or not the person takes vitamins daily in order to test (inference) this hypothesis? Or, can we get at the same answer by careful experimental design (see Schatzkin et al 2002; Dawsey et al 2014)? Supplement use is wide-spread in the United States, but both health and vitamin use differ by demographics. Young people tend to be healthier than older people and older people tend to take supplements more than younger people.

A subset of the population is measured for some trait or characteristic. From the sample, we hope to refer back to the population. We want to move from anecdote (case histories) to possible generalizations of use to the reference population (all patients with these symptoms). How we sample from the reference population limits our ability to generalize. We need a representative sample: simple to define, hard to achieve.

Statistics becomes necessary if we want to infer something about the entire populations. (Which is usually the point of doing a study!!) Typically tens to thousands of individuals are measured. But in addition, HOW we obtain the sample of individuals from the reference population is CRITICAL.

Kinds of sampling include (adapted from Box 1, Tyrer and Heyman 2016; see also McGarvey et al 2016)

Probability sampling

  • random
  • stratified
  • blocking
  • clustered
  • systematic

Nonprobability sampling

  • convenience, haphazard
  • judgement
  • quota
  • snowball

For an exhaustive, authoritative look at sampling, see Thompson (2012).

How can samples be obtained?

Sampling from a population may be convenient. For one famous example, consider the Bumpus data set. (We introduced this data set in Question 5, Chapter 5.) So the legend goes, Professor Bumpus was walking across the campus of Brown University, the day after a severe winter storm, and came across a number of motionless house sparrows on the ground. Bumpus collected the birds and brought them to his lab. Seventy-two birds recovered; 64 did not (Table1).

Table 1. A subset of Bumpus data set, summarized by sex of birds.

House sparrowsLivedDied
Female2128
Male5136

Bumpus reported differences in body size that correlated with survival (Bumpus 1899), and this report is often taken as an example of Natural Selection (cf. Johnston et al 1972). The Bumpus dataset is clearly convenience sampling. It’s also a case study: a report of a single incident. But given that is is a large sample (n = 136), it is tempting to use the data to inform about about possible characteristics of the birds that survived compared with those that perished.

Another way we collect samples from populations is best termed haphazard. In graduate school I got the opportunity to study locomotor performance of whiptail lizards (Aspidoscelis tigris, A. marmoratus, genus formerly Cnemidophorus**) across a hybrid zone in the Southwest United States (Dohm et al. 1998). During the day we would walk in areas where the lizards were known to occur and capture any individual we saw by hand. (This would sometimes mean sticking our hands down into burrows, which was always exciting — you never really knew if you were going to find your lizard or if you were going to find a scorpion, venomous spider, or …) Lizards collected were returned to the lab for subsequent measures. Clearly, this was not convenience sampling; it involved a lot of work under the hot sun. But just as clearly, we could only catch what we could see and even the best of us would occasionally lose a lizard that had been spotted. Moreover, one suspects we missed many lizards that were present, but not in our view. Lizards that were underground at the time we visited a spot would not be seen nor captured by us; individual lizards that were especially wary of people (Bulova 1994) would also escape us. In other words, we caught the lizards that were catchable and can only assume (hope?) that they were representative of all of these lizards. Applying a grid or quadrant system to the area and then randomly visiting plots within the grid or quadrant might help, but still would not eliminate the potential for biased sampling we faced in this study.

Quota sampling implies selection of subjects by some specific criteria, weighted by the proportions represented in the population. It’s different from stratified sampling because there is no random selection scheme: subjects are selected based on matching some criteria, and collection for that group stops when the sample number matches the proportion in the population. Consider our vitamin supplement survey. If student population at Chaminade University was the reference population, and we have enough money to survey 100 students, then we would want a sample of 70 women students and 30 male students, representing the proportions of the student population.

Snowball sampling implies that you rely on word-of-mouth to complete sampling. After initial recruitment of subjects, sample size for the study increases because early participants refer others to the researchers. This can be a powerful tool for reaching underrepresented communities (eg, Valerio et al 2016).

Types of Probability sampling

Random sampling is an example of probability sampling. As we defined earlier, simple random sampling requires that you know how many subjects are in the population (N) and then each subject has an equal chance of being selected

    \begin{align*} p = \frac{1}{N} \end{align*}

Examples of nonprobability sampling include

  • convenience sampling
  • volunteer sampling
  • judgement sampling

Convenience sampling (the first 20 people you meet at the library lanai); volunteer sampling (you stand in front of a room of strangers and ask for any ten people to come forward and take your survey — or more seriously, persons with a terminal disease calling a clinic reportedly known to cure the disease with a radical new, experimental treatment), and judgement sampling (to study tastes in fashion, you decide that only persons over six feet tall should be included because .…).

Random in statistics has a very important, strict meaning. As opposed to our day-to-day usage, random sampling from a population means that the probability that any one individual is chosen to be included in a sample is equal. Formally, this is called simple random sampling to distinguish it from more complex schemes. For a sampling procedure to be random requires a formal procedure for sampling a population with known size (N).

For example, at the end of the semester, I may select the order for student presentations at random. Thus, students in the classroom would be considered the population (groups of students are my sample unit, not individual students!). What is the probability that your group will be called first? Second? We need to know how many groups there are to conduct simple random sampling. Let’s take an extreme and say that all groups have a size of one; there are 26 students in this room, so

    \begin{align*} p = \frac{1}{26} \end{align*}

or

    \[p=0.0385\]

of being selected first.

Now to determine the probability of your group being selected second, we need to distinguish between two kinds of sampling:

  • Sampling with replacement — after I select the first group, the first group is returned to the pool of groups that have not been selected. In other words, with replacement, your group could be selected first and selected second! The probability of being selected second then remains

        \[p= \frac{1}{26}\]

    .

  • Sampling without replacement — after I select the first group, then I have

        \[26 -1=25\]

    groups left to select the second group, so probability that your group will be second is

        \[p= \frac{1}{25}\]

    . The first group has already been selected and is not available.

Random sampling refers to how subjects are selected from the target, reference population. Random assignment, however, describes the process by which subjects are assigned to treatment groups of an experiment. Random sampling applies to the external validity of the experiment: to the extent that a truly random sample was drawn, then results may be generalized to the study population. Random assignment of subjects to treatment however makes the experiment internally valid: results from the experiment may be interpreted in terms of causality.

R code, sample without replacement

sample(26)
[1] 1 10 11 23 17 13 8 6 19 20 21 16 14 15 5 2 12 4 26 22 9 7 25 18 3 24

R code, sample with replacement

mySample <- sample(26, replace = TRUE); mySample
[1] 9 26 24 9 3 1 23 14 26 6 5 19 17 23 11 12 3 8 7 19 2 16 3 11 1 1

How many elements duplicated?

out <- duplicated(mySample); table(out)
out
FALSE TRUE 
17 9

More on sampling in R discussed in Sampling with computers below.

Additional sampling schemes

Simple random sampling is not the only option, but in many cases it is the most desirable. Consider our multivitamin study again. Perhaps studying the entire USA population is a bit extreme. How about working from a list of AARP members, sending out questionnaires to millions of persons on this list, getting back about 20% of the questionnaires, sorting through the responses and identifying the respondent to diet categories and multivitamin use. The researchers had nearly 500,000 persons willing and able to participate in their prospective study (Schatzkin et al 2001; Dawsey et al 2014; Lim et al 2021). It’s an enormous study. But is this really much better than our described lizard experiment? Let’s count the ways: Not all older people are members of AARP (that 500,000? That’s less than 1% of the 50 and older persons in the USA); A large majority of AARP members did not return surveys; Some fraction of the returned surveys were not usable; How representative of diverse aged populations in the United States is AARP?

Simple random sampling may not be practical, particularly if sub-populations, or strata, are present and members of the different sub-populations are not available to the researcher in the same numbers. In ecology, habitat-use studies routinely need to account for the patchiness of the environment in space and time (Wiens 1976; Summerville and Crisp 2021). Thus, samples are drawn in such a way to represent the frequency within each sub-population.

For a simple example, researchers conducting a controlled breeding program of mice don’t use simple random sampling to choose pairs of mates; after all, random sampling without regard of sex of the mice would lead to some pairings of males only, or females only (Fig 1).

Sixteen mice, eight red and eight blue. Copyright 2024 Mia D Graphics

Figure 1. Sixteen mice, eight red and eight blue. Image © 2024 Mia D Graphics

R code

pairs <- c(rep("red", 8), rep("blue", 8))
replicate(2, sample(pairs, 8), simplify = FALSE)
[[1]]
[1] "blue" "blue" "blue" "red" "blue" "red" "blue" "red"

[[2]]
[1] "blue" "blue" "red" "red" "red" "red" "red" "blue"

Oops. If we interpret the output blue = male and red = female, then four out of the eight pairs were same sex pairings.

Thus, the breeding strategy is to random sample from female mice and from male mice and the stratification is sex of the mouse. Alternatively, breeders may select mice to form breeding pairs systematically: From a large colony with dozens of cages, the breeder may select one mouse from every third cage.

Stratified Random Sampling: Individuals in the population are classified into strata or sub-groups: age, economic class, ethnicity, gender, sex …, as many as needed. Then choose a simple random sample from each group, typically in proportion to the size of the sub-groups in the population. Combine those into the overall sample. For example, when I wanted a random sample of mice for my work, I called the supplier and requested that a total of 100 male and female mice be randomly selected from the five colonies they maintained. The reference population is the entire supply of mice at that company (at that time), but I wished to make sure that I got unrelated mice, so I needed to divide the population into groups (the five colonies) before my sample was constructed.

Note 1: The size of the population must be known in advance, just like in simplified random sampling. A more interesting example, the Social Security Administration conducts surveys of popular baby names by year. They post the top ten most popular names based on 1% or 5% (first strata), then by male/female (second strata).

For a simple example of the need for stratified sampling, consider our 16 blue and red mice again. If mice were randomly assigned to treatment group, then by chance unequal samples within strata may be assigned to a treatment group, potentially biasing our conclusions (Fig 2).

Sixteen mice, randomly assigned to treatment groups C and T; by chance, 75% blue in C and just 25% in T. If color was a confounding factor then our conclusions about the effectiveness of the treatment would be associated with color. Image © 2024 Mia D Graphics

Figure 2. Sixteen mice, randomly assigned to treatment groups C and T; by chance, 75% blue in C and just 25% in T. If color was a confounding factor then our conclusions about the effectiveness of the treatment would be associated with color. Image © 2024 Mia D Graphics

Block sampling: superficially resembles stratified sampling, but refers to how experimental units are assigned to treatment groups. Blocks in experimental design refers to potentially confounding factors for which we wish to control, but are otherwise not interested in. For our example described in Figure 2, mouse color may not be of specific interest for our study and therefore would be considered a blocking effect.

Agricultural research provides a host of other blocking situations: plots distributed across an elevation or hydration gradient (Fig 3), soil types, and many others.


Agricultural fields in Central Poland. NASA

Figure 3. Agricultural fields in Central Poland. NASA https://visibleearth.nasa.gov/images/146593/agriculture-fields-in-central-poland/146595w

For a microbiology example, edge effects of heated culture plates. For a growth experiment it’s unlikely we are specifically interested in the plate layout on the growth characteristics of treated cells. However, it’s also unlikely we can assume placement has no influence on growth (eg,  Mansoury et al 2021). A 96-well microplate has 60 inner wells and 36 edge wells (A1 – H1, A12-H12, plus 1st and 12th wells of columns B – H, for a total of 36 wells, Fig 4). Thus, analysis of growth effects should include a block effect for well position.

Format of 96-well plate. Red cells = "edge" wells; blue cells = "inner" wells. Plotted with ggplate package.

Figure 4. Format of 96-well plate. Red cells = “edge” wells; blue cells = “inner” wells. Plotted with ggplate package.

For long incubation periods it’s not unexpected that edge wells and inner wells may evaporate at different rates, potentially confounding interpretation of treatment effects if, by chance, treatment samples are more commonly assigned to edge wells. Block sampling implies subjects from treatment groups are randomly assigned to each block, in equal numbers if possible.

Cluster Sampling: In many situations, the population is far too large or too dispersed and scattered for a list of the entire population to be known. And, a random approach ignores that there is a natural grouping — people live next to each other, so there is going to be things in common. A multi-stage approach to sampling will be better than simply taking a random sample approach. Most surveys of opinion (when conducted reputably) use a multi-stage method. For example, if a senator wishes to poll his constituents about an issue, he may his pollster will randomly select a few of the counties from his state (first stage), then randomly select among towns or cities (second stage), to obtain a list of 1000 people to call. In some instances, they might use even more stages. At each stage, they might do a stratified random sample on sex, race, income level, or any other useful variable on which they could get information before sampling. If you are interested in this kind of work, for starters, see Couper and Miller (2009).

There are more types of sampling, and entire books written about the best way to conduct sampling. One important thing to keep in mind is that as long as the sample is large relative to the size of the population, each of the above methods generally will get the same answers (= the statistics generated from the samples will be representative of the population).

As long as some attempt is made to randomize, then you can say that the procedure is probability sampling. Nonprobability, or haphazard sampling, describes the other possibility, that is, each element is selected arbitrarily by a non-formal selecting of individual. … all the fish or birds that you catch may not be a random sample of those present in a population. For example,

  • wild Pacific salmon do not feed on the surface, hatchery salmon feed on the surface. …
  • all the individuals who respond to a survey.
  • Phone surveys, web surveys, person on the street surveys… how random, how representative?

Sampling with computers

Now to determine the probability of your group being selected second, we need to distinguish between two kinds of sampling:

Sampling with replacement — after I select the first group, the first group is returned to the pool of groups that have not been selected. In other words, with replacement, your group could be selected first and selected second! The probability of being selected second then remains

    \begin{align*} p = \frac{1}{26} \end{align*}

Sampling without replacement — after I select the first group, then I have

    \[26 -1\]

groups left to select the second group, so probability that your group will be second is

    \[p = \frac{1}{26-1}\]

. The first group has already been selected and is not available, and so on.

Sampling is usually easiest if a computer is used. Computers use algorithms to generate pseudo-random numbers. We call the resulting numbers pseudo-random to distinguish them from truly random physical processes (eg, radioactive decay). For more information about random numbers, please see random.org.

If all you wish to do is select a few observations or you need to use a random procedure to select subjects prior to observations, then these websites can provide a very quick, useful tool.

Sampling in Microsoft Excel or LibreOffice Calc

Microsoft’s Excel is pretty good at sampling, but requires knowledge of included functions. Here are the steps to generate random numbers and select with and without replacement in Excel. I’ll give you two cases.

1. For random numbers, enter the function =rand() in a cell, then drag the cell handle to fill in cells to N (in our case = 26, so A1 to A26). This function generates a random (more or less!) number between 0 and 1. We want digits between 1 and 26, not fractions between 0 and 1, so combine INT function with RAND function

=INT(27*RAND())

Note 2: To get between 1 and 9, multiply by 10 instead of 27; to get between 1 and 100, multiply by 101, etc.

In Excel to sample with replacement, simply pick the first two cells (the algorithm Excel used already has conducted sample with replacement. See next item for method to sample without replacement in Excel. You have to have installed the Data Analysis Tool Pak. Here’s instructions for Office 2010.

If you have a Mac and Office 2008, there is no Data Analysis Tool Pak, so to get this function in your Excel, install a third-party add-in program (eg, StatPlus, a free add-in, really nice, adds a lot of function to your Excel). If you have the 2011 version of Office for Mac, then the Data Analysis Tool Pak is included, but like your Windows counterparts, you have to install it (click here for instructions).

2. Let’s say that we have already given each group a number between 1 and 26 and we enter those numbers in sequence in column A.

To sample without replacement, select Tools → Data Analysis… (if this option is not available, you’ll have to add it — see Excel help for instructions, Fig 5).

Figure 10. Screenshot of Sampling in Data Analysis menu, Microsoft Excel

Figure 5. Screenshot of Sampling in Data Analysis menu, Microsoft Excel

Enter the cells with the numbers you wish to select from. In our example, column A has the numbers 1 through 26 representing each group in our class. I entered A:A as the Input Range.

Next, select “Random” and enter the number of samples. I want two.

Click OK and the output will be placed into cell B3 (my choice); I could have just as easily had Excel put the answer into a new worksheet.

Figure 11. Screenshot of input required for Sampling in Data Analysis menu, Microsoft Excel

Figure 6. Screenshot of input required for Sampling in Data Analysis menu, Microsoft Excel

To sample with replacement from column A (with our 1 through 26), type in the formula B1

=INT(27*RAND())

and the formula C1

=INDEX(A:A,RANK(B1,B:B))

then drag the cell handles to fill in the columns (first column B, then column C).

That’ll do it for MS Excel or LibreOffice Calc.

Sampling with R (Rcmdr)

It’s much easier with much more control to get sampling in Rcmdr (R) than in Excel. Sampling in R is based on the function called sample() and sample.int(). I will present just the sample() command here.

sample(x, size, replace = FALSE, prob = NULL)

Example

You want to sample ten integers between 1 and 10

sample(10)

R output

sample(10)
[1]  5  1 10  8  7  3  4  9  2  6

You have a list of subjects, A1 through A10

subjects = c("A1","A2","A3","A4","A5","A6","A7","A8","A9","A10") 
sample(subjects, 3, replace = FALSE, prob = NULL)

R output

sample(subjects, 3, replace = FALSE, prob = NULL)
[1] "A5" "A2" "A9"

Could use this to arrange random order for ten subjects

sample(subjects, 10, replace = FALSE, prob = NULL)
[1] "A9" "A3" "A8" "A2" "A1" "A4" "A10" "A7" "A5" "A6"

Try with replacement; To sample with replacement, type in TRUE after replace in the sample() function. The R output follows

sample(x, 10, replace = TRUE, prob = NULL)
[1]  5  5  3  5  3  7  3  5 10  6

R’s randomness is based on pseudonumbers and is, therefore, not truly random (actually, this is true of just about all computer-based algorithms unless they are based on some chaotic process). We can use this pseudo part to our advantage: if we want to reproduce our “random” process, we can seed the random number algorithm to a value (eg, 100), with the command in the Script Window

set.seed(100)

For 10 random integers (eg, observations), type in the Script window

sample(10)

R returns the following in the Output Window

sample(10)
[1] 4 5 7 3 10 9 2 1 6 8

Sampling was done without replacement. Here’s another selection round,

First without setting a seed value

sample(10)
[1] 10 4 7 3 8 1 2 9 6 5

Try again to see if get the same 10

sample(10)
[1]  1  4  8  2  9  6  5 10  3  7

Now to demonstrate how setting the seed allows you to draw repeated samples that are the same.

Note 3: If I need to precede the sample command with a set.seed() call, then sampling is repeatable.

set.seed(100)
sample(10)
[1] 4 3 5 1 9 6 10 2 8 7

and try again

set.seed(100)
sample(10)
[1] 4 3 5 1 9 6 10 2 8 7

Additional R packages that help with sampling schemes include: sampling() and spatialsample, which is part of the BiodiversityR package, which is available as a plugin for R Commander.

Quiz Chapter 5.6

Sampling from populations

Questions

  1. For our two descriptions of experiments in section 6.1 (the sample of patients; the sample of frogs), which sampling technique was used?
  2. What purpose is served by set.seed() in a sampling trial?
  3. Consider our question, Does taking a multivitamin daily improve health? Imagine you have a grant willing to support a long-term prospective study to follow up to one thousand people for ten years. List at least three concerns with proposed solutions about how sampling of subjects for the study.
  4. Imagine you wish to conduct a detailed survey to learn about student preferences. Your survey will include many questions, so you decide to ask just ten students. Student population is 70% female, 30% male.
    1. Assuming you select at random (simple random sampling), what is the chance that no male students will be included in your survey?
    2. You are able to increase the number of surveys to 20, 30, 40, or 50. What is the chance that no male students will be included in your survey for each of these increased sample numbers?
      1. What can you conclude about the effects of increasing survey sample size on representativeness of students for the survey?
  5. Discuss how you could apply a stratified sampling scheme to this survey and whether or not this approach improves representativeness.
  6. Why are random numbers generated by a computer called pseudo random numbers?

Chapter 5 contents

5.1 – Experiments

How to put an experiment together

With the background behind us, we outline a designed experiment. Readers may wish to review concepts presented previously, in

Experiments have the following elements

  1. Define a reference population.
    1. All patients with similar angina symptoms — shortness of breath, chest pain, dizziness — but who have not suffered a heart attack, or myocardial infarction (Wu et al 2017);
    2. All adult frogs in a pond near industrial waste runoff (eg, Sowers et al 2009).
  2. Define sample unit from the reference population (eg, as many patients as can be screened by the physician are recruited; all frogs visible to the biologist are captured).
  3. Design sampling scheme from population (eg, random sampling vs. convenience/haphazard sampling).
  4. Agree on the primary outcome or endpoint to be measured and whether additional secondary outcomes are to be observed and collected.
    1. For the patients with similar symptoms, we follow
    2. For the frogs, we count adults with developmental malformations.
  5. Separate the sample into groups so that comparisons can be made
    1. We may be tempted to provide the same treatment and follow the group for improvement — that is the clinically valid approach; to be an experiment, we would use random assignment of individuals to one of two groups — half will receive a placebo, half will receive a beta blocker;,
    2. The frog example doesn’t follow this scheme — the event has already occured, it’s an observational study.
  6. Devise a way to exclude or distinguish from the possible explanations, or alternative hypotheses.
    1. eg, for the physician, she will record how many of her patients fail to respond to treatment;
    2. for the biologist, noting presence or absence of parasites between the normal and abnormal frog defines the groups).

Note 1: The studies cited are examples of work done on similar problems, they are not offered to evaluate or critique any research efforts that were done or described in these reports.

The frog study, as described by me for pedagological purposes, lacks any design elements that would allow for a test of the hypothesis that industrial runoff causes developmental malformations in frogs — I’m not saying or implying anything about the actual studies!

A basic experimental design looks like this.

  1. One treatment, with two levels (eg, a control group and an experimental group)
  2. A collection of individuals recruited from a defined population, esp. by random sampling.
  3. Random assignment of multiple individuals to one of the two treatment levels.
  4. The treatments are applied to each individual in the study.
  5. A measure of response (the primary outcome) and additional features (secondary outcomes) are recorded for each individual.

Of course, there are other considerations, and by now, students reading this text should be able to identify what is missing from the list. (See quiz questions.)

Contrast with design of observational studies

Importantly, for observational studies, no matter how sophisticated the equipment used to measure the outcome variable(s), key steps outlined above are missing. Researchers conducting observational studies do not control allocation of subjects to treatment groups (steps 4 and 3 in the two above lists, respectively). Instead, they may use case-control approaches, where individuals with (case) and without (control) the condition (eg, lung cancer) are compared with respect to some candidate risk factor (eg, smoking). Both case and control groups likely have members who smoke, but if there is an association between smoking and lung cancer, then more smokers will be in the case group compared to the control group.

A cohort study, also a form of observational study, is similar to case-control except that the outcome status is not known. A cohort study includes, in our example, smokers and non smokers who share other characteristics: age, medical history, etc. The other kind of design you will see is the cross-sectional study. Cross-sectional studies are descriptive studies, and, therefore also observational. Primary outcomes along with additional characteristics and outcomes are measured for a representative subsample, or perhaps even an entire population. Cross-sectional studies are used to absolute and relative risk rates. In ecology and evolutionary biology cross sectional studies are common, eg, comparisons of species for metabolic rate (Darveau et al 2005) or life history traits (Jennings et al 1998), and specialized statistical approaches that incorporate phylogenetic information about the species are now the hallmark of these kinds of studies.

A word on outcomes of an experiment. Experiments should be designed to address an important question. The outcomes the researcher measures should be directly related to the important question. Thus, in the design of clinical trials, researchers distinguish between primary and secondary outcomes or endpoints. In educational research the primary question is whether or not students exposed to different teaching styles (eg, lecture-style or active-learning approaches) score higher on an knowledge-content exams. The primary outcome would be the scores on the exams; many possible secondary outcomes might be collected including student’s attitudes towards the subject or perceptions on how much they have learned.

Note 3: While it’s logical to include secondary outcomes — they may offer support or complement the primary results — there needs to be caution in interpreting the test results of multiple secondary outcomes. By definition, the study is designed to test hypothesis about the primary or main outcome. While there are a number of concerns about testing secondary outcomes, from a statistical point of view, the study may lack the sample size — the power, see Chapter 11 — Power analysis — to test many additional hypotheses without the risk of increasing chance of false positives because of the multiple comparisons problem. We address this scenario in Chapter 12.1 – The need for ANOVA.

About Primary and Secondary Outcomes.

Consider a evolutionary physiology project on species differences and effects of body size on burst running speed of terrestrial lizards (eg, Bauwens et al 1995; Dohm et al 1998; Labonte et al 2024). Our independent variables, the things we believe influence running performance — and therefore we wish to control for — species, body size, core body temperature, body condition, race quality (see cited papers for these and other examples).

What do we measure? What are the likely primary vs. secondary outcomes for a this kind of project?

Primary outcomes are the direct, definitive measures of maximal locomotor performance. Examples may include

  • Maximum Burst Speed: The absolute highest velocity achieved by the lizard over a short distance or time interval (eg, across a 0.5-meter interval).
  • Maximum Acceleration: The highest rate of speed change during the start of the burst or during a sudden sprint increase.
  • Stride Length: The physical distance covered during a single full stride cycle at maximum speed.
  • Stride Frequency: The number of strides taken per second at maximum speed.

And what other measures may be of interest — secondary outcomes — those that may help explain why or how the primary performance metrics change across the independent variables. To list a few,

  • Stance and Swing Duration: The individual time durations spent with feet on the ground (stance) vs. in the air (swing).
  • Gait Transition Thresholds: The speed at which a lizard shifts from a walk/trot to a high-speed gallop or bipedal running (if applicable to the species).
  • Fatigue Rate: The decline in burst speed across successive trials or over a sustained treadmill run.
  • Trial-to-Trial Repeatability: The statistical consistency (intraclass correlation coefficient) of a lizard’s performance across multiple bursts.

In a prospective cohort study, we may track participants over time to observe how baseline factors predict the development or worsening of type 2 diabetes (eg, Gao et al 2026). Progression over a two year followup would be measured by incident clinical events or the longitudinal rate of change in biomarkers. Possible primary outcomes are

  • Glycated Hemoglobin (HbA1c): A biomarker reflecting average blood glucose levels over the past 2 to 3 months (glycemic control).
  • Fasting Plasma Glucose (FPG): Glucose concentration in the blood after an overnight fast (acute baseline glycemic regulation).
  • T2D Complication Stage: A clinical score or categorization indicating the presence and severity of microvascular or macrovascular damage.

Secondary outcomes may include

  • Fasting Lipid Profile
  • C-Reactive Protein
  • Body Mass Index
  • Estimated Glomerular Filtration Rate

If one or more primary outcomes are significant, this reveals whether an intervention or exposure (like a vegan diet or income level) changes the clinical course of Type 2 Diabetes. Secondary outcomes are critical for explaining how and why it happens. Secondary outcomes provide mechanistic context.

Some caution is warranted — imagine if, after conducting the study, we find diet, age, or income did not statistically affect your primary outcome, but did significantly alter a secondary outcome (eg, improved lipid profiles). What to report? We state in your Abstract and Discussion that the study failed to meet its primary objective, and that any significant secondary findings are exploratory. We cannot claim that a diet “successfully manages diabetes progression” based solely on a secondary marker (cf discussion and suggsted guidelines in Butcher et al 2022).

Example of an experiment.

We’ll work through a familiar example. The researcher is given the task to design a study to test the efficacy in reducing tension headache symptoms by a new pain reliever. There are many possible outcomes: blurry vision, duration, frequency, nausea, need for and response to pain medication, and level of severity (Mayo Clinic). The drug is packaged in a capsule and a placebo is designed that contains all ingredients except the new drug. Forty subjects with headaches are randomly selected from a population of headache sufferers. All forty subjects sign the consent form and agree to be part of the study. The subjects also are informed that while they are participating in the research study on a new pain reliever, each subject has a 50-50 chance of receiving a placebo and not the new drug. The researcher then randomly assigns twenty of the subjects to receive the drug treatment and twenty to receive the placebo, places either a placebo pill of the treatment pill into a numbered envelope and gives the envelopes to a research partner. The partner then gives the envelopes to the patients. Both patients and the research partner are kept ignorant of the assignment to treatment.

We can summarize this most basic experimental design in a table, Table 1.

Table 1. Simple formulation of a 2X2 experiment (aka 2×2 contingency table).

Did the subject get better?
Yes No Row totals
Subject received
the treatment

 Yes a b a + b
No c d c + d
Column totals a + c b + d N

where N is the total number of subjects, a is number of subjects who DID receive the treatment AND got better, c is number of subjects who DID NOT receive the treatment, but DID get better, b is number of subjects who DID receive the treatment and DID NOT get better, and d is number of subjects who DID NOT receive the treatment and DID NOT get better.

Note 4: We owe Karl Pearson (1904) for the concepts of contingency — that membership in a (or b, c, or d for that matter) is a deviation from any chance association or independent probability (Chapter 6) — and the contingency table. Contingency is an important concept in classification and is central to machine learning in data science (Michie et al 1995). We return at some length to contingency and how to analyze contingency tables in Chapter 7 and Chapter 9.2.

By convention, the standard format for an epidemiological 2×2 contingency table organizes exposure on the rows and disease outcome on the columns as follows:

  • Rows (Exposure): The top row represents Exposed / Treatment (Yes), and the bottom row represents Non-exposed / Control (No).
  • Columns (Outcome): The left column represents Disease / Case (Yes), and the right column represents No Disease / Control (No). See Chapter 7.

And our basic expectation is that we are testing whether or not treatment levels were associated with subjects “getting better” as measured on some scale.

One possible result of the experiment, although unlikely, all of the negative outcomes are found in the group that did not receive the treatment (Table 2).

Table 2. One possible outcome of our 2X2 experiment.

Subject improved
Yes No
Subject received
the treatment		 Yes 20 0
No 0 20

Results of an experiment probably won’t be as clear as in Table 2. Treatments may be effective, but not everyone benefits. Thus, results like Table 3 may be more typical.

Table 3. A more likely outcome of our 2X2 experiment.

Subject improved
Yes No
Subject received
the treatment		 Yes 7 13
No 2 18

Must a control group always be included in an experiment?

To be an experiment, yes. However, it is not true that the control treatment must only be applied to separate, individual subjects. For example, a common experimental design would be repeated measures of an outcome, like behavior (eg, feeding response) before and after an intervention is applied. Thus, the individual is its own control or reference point. For example, Dohm et al (2008) measured feeding and locomotor behavior of adult toads (Rhinella marina, formerly Bufo marinus) prior to and after ozone exposure. We concluded that a single 4-h exposure to O3 depressed toad feeding behavior, but had little effect on voluntary locomotor behavior. Typically, the design randomly assigns individuals to the  treatments — some individuals get the exposure first, then the control exposure second; others, again selected at random, receive the control exposure first, then the active exposure.

The within subjects design is an important method in biological research; these designs offer control for individual differences as confounding variables and require fewer participants compared to separate control group designs. This design is effective provided there is little risk of carryover effects; if the effects of a previous condition linger, then it may influence a participant’s response in subsequent conditions, confounding the results. For our study (Dohm et al 2008) we used a modified within-subjects design by employing a split plot design. We return to within-subjects designs in Chapter 14.4 – Randomized block design.

Questions

  1. When I updated this page in 2020, we were in our fifth month of the coronavirus pandemic of 2019-2020. Daily it seemed, we heard news about “promising coronavirus treatments,” but no study had at the time been published that would meets our considerations for a proper experiment. However, on May 1, 2020, the FDA issued an emergency use authorization for use of the antiviral drug remdesivir (Gilead), based on early results of clinical trials reported 29 April in The Lancet (Wang et al 2020). Briefly, their study included 158 treated with does of the antiviral drug and 79 provided with a placebo control. Both groups were treated otherwise the same. Improvement over 28 days was recorded: 62 improved with Placebo and 133 improved receiving dose of remdesivir.
    1. Using the background described on this page, list the information needed to design a proper experiment. Using your list, review the work described in The Lancet article and check for evidence that the trial meets these requirements.
    2. Create a 2×2 table with the described results from The Lancet article.
  2. Consider our hazel tea and copper solution experiment described in Chapter 3. The outcome variables (Tail length, tail percentage, Olive moment) are quantitative, not categorical. Create a table to display the experimental design.
  3. For the migraine example, identify elements of the design that conform to the randomized control trial design described in Chapter 2.4.

Quiz Chapter 5.1

Experiments

Chapter 5 contents