There are certain themes that crop up in Critical Reasoning questions so often that it’s worthwhile to treat these problem types as their own sub-categories. One category that shows up with greater frequency in each new edition of the Official Guide is one that I’ll christen, “The tainted experiment.”
The logic of these arguments is always rooted in the notion that we can only trust the results of the experiment if we have a legitimate control group, and there aren’t any other confounding variables that we’ve failed to account for. Spoiler alert: typically in GMAT questions, we will find such confounding variables tainting the experiment’s predictive value.
Imagine, for example, that you’re testing a drug designed to alleviate headaches. You have two groups of subjects: a control group that takes a placebo and an experimental group that receives the drug. The results of the experiment show that the control group has a higher rate of headaches than the group receiving the medication. Time to rejoice, notify the delighted shareholders, and move this drug to market as quickly as possible? Well, maybe.
But now imagine that the control group consisted largely of stressed-out, sleep-deprived college students living near construction sites, and the experiment group consisted of retired yoga instructors. Suddenly we’ve got other variables to contend with. Yes, it’s possible that the effectiveness of the drug is what accounts for the differential in headache incidence between the two groups. But it’s just as likely that other environmental factors are responsible. A good experiment would have controlled for these factors.
The upshot: whenever you see a question that involves an experiment with a control group, always ask yourself if there are variables that the experimenters have failed to account for.
Here’s a good example of such an argument:
In Colorado subalpine meadows, nonnative dandelions co-occur with a native ﬂower, the larkspur. Bumblebees visit both species, creating the potential for interactions between the two species with respect to pollination. In a recent study, researchers selected 16 plots containing both species; all dandelions were removed from eight plots; the remaining eight control plots were left undisturbed. The control plots yielded significantly more larkspur seeds than the dandelion-free plots, leading the researchers to conclude that the presence of dandelions facilitates pollination (and hence seed production) in the native species by attracting more pollinators to the mixed plots.
Which of the following, if true, most seriously undermines the researchers’ reasoning?
A) Bumblebees preferentially visit dandelions over larkspurs in mixed plots.
B) In mixed plots, pollinators can transfer pollen from one species to another to augment seed production.
C) If left unchecked, nonnative species like dandelions quickly crowd out native species.
D) Seed germination is a more reliable measure of a species’ ﬁtness than seed production.
E) Soil disturbances can result in fewer blooms, and hence lower seed production.
This is a classic experiment argument. There are two populations: plots that contain both dandelions and larkspurs, and plots that have had all the dandelions removed, and thus contain only larkspurs. We’re told that the plots containing both types of flowers produced more larkspur seeds than the plots containing only larkspurs, thus validating the contention that the presence of dandelions has a positive benefit on larkspur seed yields.
Fortunately, the GMAT is pretty predictable. If we’re trying to weaken the conclusion derived from an experiment comparing two populations – a control group and an experimental group – we’re looking for a confounding variable. The initial hypothesis is that the presence of dandelions promotes seed production in larkspurs. An alternative hypothesis is that an environmental factor we haven’t yet considered accounts for the differential in larkspur seed production in the two groups, so that’s what we’re on the lookout for when we examine each of the answer choices.
A) Which flower bees prefer sheds no light on the validity of the experiment. A is out.
B) This answer option would be entirely consistent with the hypothesis that dandelions promote larkspur seed production. We’re trying to weaken the argument. B is also out.
C) This answer choice makes no sense. We’ve already been told that the plots containing both types of flower produce more larkspur seeds – we never want to contradict a premise. C is no good.
D) This tells us nothing about whether it is the presence of dandelions that’s helping promote larkspur seed production. D gets kicked to the curb.
E) If removing the dandelions disrupts the soil, perhaps it’s the disrupted soil, rather than the absence of dandelions, that accounts for the lower larkspur production in the plots where the dandelions have been removed. We’ve got our confounding variable – E is the answer.
Takeaway: On Critical Reasoning questions on the lookout for the tainted experiment. If you’re trying to weaken an argument regarding an experiment containing a control group and an experimental group, the key will be determining which answer choice provides a confounding variable, and thus, an alternative explanation for the conclusion given.