Jump to content
Toggle search
Toggle menu
75
257
179
8.5K
Sense & Sensibility & Science
Toggle preferences menu
Toggle personal menu
Not logged in
Your IP address will be publicly visible if you make any edits.

6.2 Hill's Criteria

From Sense & Sensibility & Science
More actions
Revision as of 18:37, 25 May 2023 by Gpe (talk | contribs) (// Edit via Wikitext Extension for VSCode)


Useful Links

Readings and Assignments

Learning Goals

After this lesson, students should

  1. Identify cases in which "ideal" RCT experiments are not possible, due to ethical or practical constraints.
  2. For a given scenario in which a causal hypothesis/claim is being made, identify plausible alternative hypotheses that could be consistent with the data.
  3. Identify additional sources of evidence that could be used to help mitigate flawed experiments, including prior plausibility, dose-response relationships, specificity, temporal ordering, and consistency across contexts.
  4. Recognize when causal evidence in the absence of an RCT can be fairly compelling, especially if there are many different types of evidence combined.

Definitions

Hill's criteria:

  • Prior plausibility
    Can a plausible mechanism be constructed, or is there some other basis for interpreting the current evidence in terms of one causal structure over another, such as data from other studies?
  • Temporality/Temporal Sequence
    Did the hypothesized cause precede the effect?
  • Specificity
    Specific predictions for specific consequences that have come true are less likely to be caused by other factors.
  • Dose-response Curve
    Do the quantities of the hypothesized cause correlate with the quantity, severity, or frequency of the hypothesized effect across ranges?
  • Consistency Across Contexts
    Does the correlation appear across diverse contexts?
It is not necessary for all of Hill's criteria to be satisfied to infer causation. Each criterion adds to the case for causation. Some criteria are not applicable in certain situations (e.g. dose-response curve in whether light switches cause the light to turn on and off).

Examples

  • Scrotal cancer in chimney sweeps, example by John.

  • The causal connection between leaded gasoline across the world and violent crimes in many countries. (See video above, 17:57.)
    • Prior Plausibility: High levels of lead are known to cause cognitive damage. It is conceivable that extended exposure to lower levels of lead could have similar effects.
    • Temporality: In the graph shown in the video, the levels of violent crime correlate with the use of leaded gasoline, but delayed by about 20 years.
    • Specificity: Within the same demographic group, delinquents are 4 times more likely to have elevated bone lead concentrations than non-delinquents.
    • Dose-response Curve: See temporality.
    • Consistency Across Contexts: The delayed rise in violent crime after increased use of leaded gasoline is observed in many industrialized countries.

Common Misconceptions

  • Since we can't run a randomized controlled trial on whether CO2 emissions cause global warming, we can't ever know whether it does.
    A non-RCT study can still be serve as potentially weaker, but sometimes just as strong, evidence for causation.
  • Students are quick to notice small sample size, slower to notice problems with experimental design.
  • Students struggle to generate non-RCT types of evidence for causality, although they are better at recognizing it.
  • Identifying natural experiments is also difficult, probably because many students find the principles of RCTs slippery.

Context

This is a discussion-based lesson that familiarizes students with the concept of Hill's criteria, which are used when an ideal experiment (e.g. RCT) could not be done due to resource or ethical considerations. The criteria themselves are not difficult, but students typically have trouble remembering their names, and they would benefit from a diverse range of examples.

Before

2.2 Systematic and Statistical Uncertainty
When coming up with alternative explanations to an apparent correlation between two variables, it helps to consider factors that contribute to systematic and statistical uncertainties.
3.1 Probabilistic Reasoning
In the upcoming lesson on Probabilistic Reasoning, students will explore how to use quantified credence (i.e., confidence) levels to track and communicate their certainty about claims and predictions. For example, a .50 credence level indicates a 50-50 chance the claim is false or true, whereas a 1.00 credence level indicates complete certainty. Because it is difficult to impossible to conclusively infer causation without an RCT, Hill's criteria are typically used to infer causation with varying degrees of confidence, such that each additional piece of evidence (especially of new types) increases the confidence that the causal claim is true.
6.1 Correlation and Causation
RCTs are introduced as ideal experiments. These are often not possible due to resource or ethical considerations, and we must resort to Hill's criteria to examine the plausibility of a causation.
Causation is defined as correlation under manipulation. When manual manipulation is not possible, Hill's criteria can still help establish causation.

After

11.2 When Is Science Suspect
In this upcoming lesson, students will explore how science has sometimes been used to justify the oppression of certain human groups, e.g. by inferring from current differences in achievement that there are fundamental differences in capacity, when in fact these differences in achievement can be fully explained by differences in opportunity. Because it is impossible to randomly assign individuals to human groups such as "African-Americans" or "women," all of this science is observational, and subject to the limitations of observational research. In evaluating non-RCT evidence, it is important to remember these limitations of observational as opposed to experimental research, especially when data may be used to exacerbate existing injustices.

Recommended Outline

Before Class

  • Prepare a seating chart.
  • Review PlayPosit and discussion questions and ask faculty, Gabriel, or Emlen any questions you have.
  • (Optional) Prepare a presentation.

During Class

Lesson Content

Concept Review

  1. (8 min) Review: What are the three essential elements of a Randomized Controlled Trial?
    Experimental intervention, control condition, and random assignment.
    Note that random sampling is not essential to the validity of an RCT. It merely affects the generalizability of the results of the RCT.
  2. (? min) For both of the following hypotheses, what RCT would you use to test it? Why would it be challenging to perform?
    1. For a child, does spending more than 22 hours in their home per day over their childhood increase their risk of developing myopia?
      It is not hard to imagine an ideal RCT for this, but it is unethical/impractical to perform the intervention.
    2. On the level of individual towns, does implementing universal basic income reduce the rate of violent crimes in 10 years? In 150 years?
      It is conceivable to do an RCT over 10 years, albeit expensive. However, it is hard to do this experiment on a large enough sample. It is also much less practical to track such an experiment over 150 years. It is difficult to randomise the towns.

Paper Analysis (Part 2)

Respond to Emlen's concerns: 6.2 Hill's Criteria: Two studies on natural experiments in section, not really focused on hill's criteria. Perhaps better to do just one natural experiment, and more practice w/ Hill's. This is related to the funkiness of teaching RCTs before Hill's criteria. Switch out one natural experiment for something similar to the Covid & masks case, but where it's really impossible to do an RCT, maybe Chinese restaurant syndrome. Need a little more structure for the discussion of Hill's criteria on Sunset Time and Social Jetlag, i.e. instruct students explicitly to extract each of Hill's criteria from the studies. Concept Review currently RCTs, not Hill's criteria.

This is a continuation of last lesson's paper analysis, in which we looked at a classic example of a proper RCT. Here, we will look at two more papers in which an RCT is difficult to conduct, but a strong case can still be made about causation based on experimental results.

Strictly speaking, the experiments in the following two papers are not RCTs, since the experimenter never performed any intervention. However, their causal conclusions are nearly as strong as an RCT. Hill's criteria, on the other hand, are useful when even this kind of experiment is impossible to do. We will go through an example of Hill's criteria in #Covid-19 below.

For each of the two papers, students should try to answer the following.

  1. What causal relationship is this paper trying to study? What is the hypothesis?
  2. What, if it exists at all, is the experimental intervention? (i.e., What is the independent variable that is being manipulated)?
  3. What is the dependent variable that is being measured (i.e., the variable the researchers anticipate may be affected by the experimental intervention)?
  4. How is the independent variable manipulated? Are there control and intervention groups? Are those randomly assigned? Is the control condition a good one (only the independent variable is different, with all else kept equal)? Is this an RCT?
  5. What is the result of the experiment?
  6. Can the causal relationship in question 1 be concluded from the experimental results? If not, what, if anything, can be concluded? How confident are you in this conclusion?
  7. Can you think of an alternative explanation for the data?

Vietnam War Draft Lottery

Example of natural experiment, where the intervention and assignment are not performed by the experimenter but by environmental factors: Lifetime Earnings and the Vietnam Era Draft Lottery

  1. Does being drafted to serve in the military cause one's income to change later in life?
  2. Whether one has served in the military (veteran status).
  3. One's lifetime income.
  4. The veteran status is manipulated by the Vietnam War draft lottery, which randomly selected young men to serve in the military. The control group consists of young men who were not drafted by the lottery. This seems like a good control, as the selection is supposedly not based on any factor that may introduce systematic bias to one's future income. There are other factors that allow people to get out of the draft. This paper attempts to account for this. This is a natural experiment.
    @Gabriel write this more carefully.
  5. Veterans earn 15% less than non-veterans long after the draft and the war ended.
  6. Being drafted to serve in the military causes a long term reduction in one's income.
  7. Well-connected or wealthier young men may have had an easier time getting out of the draft, which is a possible confound. Unknown confounds are more likely in a natural experiment, where the differences between experimental group and control group are not as tightly controlled as in a deliberately created RCT.

Upshot: This is a natural experiment. All the aspects of an RCT are still satisfied, but the randomized selection and intervention process is not performed by the experimenter but by an external agent/natural process.

Sunset Time and Social Jetlag

Not technically an RCT but could be argued to have comparable strength: Sunset Time and the Economic Effects of Social Jetlag

  1. Does an extra hour of natural light in the evening (according to local time) cause a change in one's sleep duration?
  2. The time of sunset in one's local time zone.
  3. Sleep duration.
  4. People who live on either side of a time zone boundary are considered, since they live geographically close to each other but have a 1-hour difference in sunset time. We may treat those on the east side of a boundary to be the intervention group and those on the west side to be the control group. The people are certainly not randomly assigned to one group or the other, but it is presumed that there are no systematic differences in the populations on either side that would cause a big difference in sleep patterns, other than the time zones. This is not an RCT, but its strength in implying causal relationship may be argued to be comparable to that of an RCT.
  5. Those living in a time zone with one extra hour of daylight in the evening (an earlier sunset time) sleep 19 minutes less on average.
  6. Living in a time zone with one extra hour of daylight in the evening causes a 19-minute reduction in average sleep duration.
  7. This is a pretty convincing study, what with the huge representative Census sample and the arbitrary cutoff. However, it's possible that some existing geographic power/wealth differential enabled people one one side of the time zone split to set it at their advantage (e.g. people in cities), leaving the less powerful/wealthy on the darker side (e.g. people in more rural areas). If this occurred, the difference in sleep and health could be due to the pre-existing power or wealth differential, not the time zone.

Discussion Questions

In your small group, explain Hill's criteria using aspects of the Sunset Time and Social Jetlag study as examples.

  • Hill's criteria is essentially a list of criteria that can add to the case of determining causality, when you can only observe correlation. The criteria are used when RCT's can't be performed. See definitions above.
  • Prior plausibility: The presence of sunlight and one's daily schedule are well known to affect sleep patterns. If the work schedule is shifted by one hour due to time zone, it should affect sleep patterns as well.
  • Temporality/temporal sequence: People have lived on one side of the time zone border before they developed this sleep pattern.
  • Specificity: The observed effect in sleep times is drastic across the time zone borders, and this effect is not observed elsewhere. This effect was also only observed in people that had to get up early for work.
  • Dose-response curve: The further away one lives from a time zone border (more accurate their sunset/work time is to the natural circadian rhythm), the less pronounced the effect on sleep times becomes. The effect is also absent among unemployed people. (Fig. 6)
  • Consistency across contexts: The effect on sleep times is observed across many different time zone borders scattered throughout the US.

Covid-19

There has been a lot of controversy regarding mask wearing policies to combat the spread of COVID-19. Suppose it were very difficult to conduct a randomized controlled trial to study the effect of community mask wearing on COVID spread. (Such RCTs have in fact been conducted; see this paper.) However, we'd still like to find out whether there is such a causal connection using Hill's criteria.

Work in small groups. For each of the Hill's criteria, what evidence, if observed, would support that criterion?

  1. Plausible mechanism
    Masks physically blocks one's mouth from emitting/receiving virus-ridden aerosols.
  2. Temporal sequence
    Do COVID cases reduce after a mask mandate has been instituted in an area/business/school?
  3. Consistency across contexts
    Is this reduction in cases after a mask mandate observed in many countries/areas/businesses/schools?
  4. Dose-response curve
    Does the extent of decrease in COVID cases correlate with the percentage of mask wearers in an area?
  5. Specificity
    Is it the case that counties/businesses/schools that implemented a mask mandate have a dramatically reduced case rate compared to those that did not?

Overflow

Chinese Restaurant Syndrome

(This has been a good quiz question when it wasn't a discussion topic.)

Starting from the late 1960s, some people in the USA noticed that they had a headache after eating at Chinese restaurants. They attributed their symptoms to the consumption of monosodium glutamate (MSG), a common food additive in East Asian cuisines. "Chinese Restaurant Syndrome" was coined as a term referring to the negative health effects of MSG.

"Chinese restaurant syndrome" was the popular term historically used and it highlights the underlying racism and the bias associated with the claims. We will be discussing prejudices and biases in society clouding our sound scientific judgment in a future topic.
  1. (4 min) Describe a randomized controlled trial (RCT) that can test the hypothesis that "consumption of MSG causes headaches."
    Randomly split the participants into two groups (random assignment). Give one group a MSG-filled meal (experimental condition). Give the other group a meal low in MSG (control condition). Observe if the participants experience a headache after the meal.
  2. (10 min) Suppose that for some reason a RCT could not be done, but after some research, the following facts are known.
    • MSG also appears in Western cuisine, naturally in tomatoes and artificially added in many processed foods, and Americans who reported Chinese Restaurant Syndrome did not experience such symptoms after eating a Western meal containing MSG.
    • People in East Asian countries, who consistently consume meals containing high amounts of MSG, do not report such symptoms.
    • No biological pathway is known that connects MSG consumption and headaches.
    1. (2 min) Why might a RCT not be able to be conducted?
      An RCT might not be able to be conducted for a variety of ethical or practical reasons.
    2. (6 min) Each of the three points above can be used as evidence against a causal relationship between MSG consumption and headaches. For each point, discuss which of Hill's criteria for causality is not satisfied as a result of that fact.
      (1) Consistency across different contexts. (2) Dose-response curve and consistency across contexts. (3) Plausible mechanism.

Based on Hill's Criteria we can say that there is not a general causation between MSG and the symptoms claimed by "Chinese Restaurant Syndrome." See the quote at this link:

"These adverse event reports helped trigger FDA to ask the independent scientific group Federation of American Societies for Experimental Biology (FASEB) to examine the safety of MSG in the 1990s. FASEB's report concluded that MSG is safe. The FASEB report identified some short-term, transient, and generally mild symptoms, such as headache, numbness, flushing, tingling, palpitations, and drowsiness that may occur in some sensitive individuals who consume 3 grams or more of MSG without food. However, a typical serving of a food with added MSG contains less than 0.5 grams of MSG. Consuming more than 3 grams of MSG without food at one time is unlikely."

Causal Link Brainstorming

  1. (5 min) Brainstorm three causal links that you are pretty confident are real without RCT evidence.
  2. (4 min) Why are you confident about these causal links?
    Think about causal links in everyday life, like allergies or breaking a glass.
    There are any number of answers to this question (i.e. Rain clouds cause rain, The sun heats up objects, Exercise causes muscle growth, Heat will melt butter, Pressing the unlock button on your keys unlocks your car etc.)
  3. (5 min) Can you come up with a situation where someone might come to the wrong conclusions about these causal links?
    Temporality may cause you to draw the wrong conclusions, especially when it is immediate (i.e. In the PlayPosit John talked about the power outage and someone in a family argument and pounding the table and the lights go out; Homeopathic remedies can lead to the belief that the remedy caused the healing; if you are in a parking garage and you unlock with your fob and another car unlocks at the same time or you slam your car door and another car alarm goes off.)
  4. (5 min) How might using just one of Hill's criteria lead someone to a wrong conclusion?
    For each causal link you can work through the different criteria (i.e. Every time you press the unlock button on your car keys it only unlocks your car, this mechanism works at home, in a parking garage, and on a street, the mechanism works in any context such as rain/sun/night/day, other people can do it, etc.)
Retrieved from "https://sensibility.gpe.fyi/index.php?title=6.2_Hill%27s_Criteria&oldid=3100"