Please enable JavaScript.
Coggle requires JavaScript to display documents.
practical problems 1: ethics, participant recruitment, and random…
practical problems 1: ethics, participant recruitment, and random assignment
ethical and legal issues with experiments
the ethics of experimentation on human beings
the ethics of withholding a potentially effective treatment from control or comparison participants
the ethics of random assignment compared with alternatives such as assignment based on needs
the conditions under which experiments might be discontinued for ethical reasons
some legal problems that bear on experiments
horrible experiments were done during WWII and in the US scientists withheld cures for syphilis to see the long term effects
ethical codes
informed consent
institutional review boards
post WWII several review boards came up with 3 rules to follow when designing an experiments
respect for persons
: that individuals are autonomous agents with the right to decide whether to enter a study (hence the need for informed consent) and, if their autonomy is diminished by a disability, have the right to be protected from harm
Beneficence
: the researchers must maximize benefits and minimize harms to participants (hence the need to reveal potential harms and benefits)
3.
Justice
:that the benefits and harms of treatment be distributed fairly (hence the need to recruit participants fairly), and that persons not be deprived of efficacious treatments to which they would otherwise be entitled (hence the need to inform of alternative treatments)
informed consent and experiments
participants have to write that they consent to the experiment rather than to say they did not consent later
a statement that the study involves research, an explanation of the purposes of the research, the expected duration of the participant's participation, a description of the procedures to be followed, and identification of any procedures that are experimental
a description of any reasonably foreseeable risks or discomforts to the participant
a description of any benefits to the participant or to others that may reasonably be expected from the research
a disclosure of appropriate alternative procedures or courses of treatment, if any, that might be expected from the research
a statement describing the extent, if any, to which confidentiality of records identifying the participant will be manipulated
for research involving more than minimal risk, an explanation as to whether any compensation and any medical treatments are available if injury occurs and, if so, what they consist of or where further information may be obtained
an explanation of whom to contact for answers to pertinent questions about the research and research participant's rights and whom to contact in the event of a research-related injury to the participant
a statement that participation is voluntary, that refusal to participate will involve no penalty or loss of benefits to which the participant is otherwise entitled, and that the participant may discontinue participation at any time without penalty or loss of benefits to which the participant is otherwise entitled
The IRB monitors research using human beings
withholding a potentially effective treatment
the question becomes how do we decide who gets the treatment if people are deserving of it
if we do not have treatment and control groups side effects, both short term and long term may not be found until they have effected people taking the drug
variants in treatment can be implemented. this means that different types of effective treatment can be switched in and out so participants are all receiving treatment but they dont know which treatment
options when withholding treatment is problematic
using dose-response designs
offering all participants an intervention prior to randomization
using a "treatment-on-demand" control
the ethics of random assignment
need, merit and randomization are the three things we cycle through
some argue that randomization is only appropriate when conditions are being compared to be therapeutic and if no better treatment exists
an argument against randomization is that the participants may not know their rights or understand what randomization actually is
other partial remedies
use a regression discontinuity design to assign based on need or merit instead of randomly
use a safety valve mechanism for treating the most needy participants
divide participants into strata by need, assigning proportionally more of the most needy to treatment
assign proportionately more participants to the treatment highest in demand
use a dose-response design
use an adaptive design to increase the proportion of assignment over time to the most successful condition
use the informed-consent procedure to ensure that participants are willing to be assigned to all conditions
use public lotteries for assignment to increase perceived fairness
if in a serious case, like aspirin studies to prevent heart attacks, the treatment is more effective than the control it may be unethical to withhold the findings and the experiment may have to be discontinued
improving random assignment
1. simple random assignment
: - any procedure for assigning units to conditions by chance with nonzero probability (without replacement) -- typically done with table of random numbers or computer-generated random numbers
2. restricted random assignment to force equal sample sizes
: particularly useful with small sample sizes to prevent severely unequal splits over conditions. -- equal sample sizes tend to minimize power for testing treatment main effects in many (but not all) designs. -- preferred method is to assign from matches or strata
3. restricted random assignment to force unequal sample sizes
: can be done to cope with practical limitations, such as restriction in the number of units that can receive treatment or ethical objections to depriving many participants of treatment. -- can increase power to test certain limited hypothesis (see optimal design theory)
4. batch randomization
: refers to cases in which small groups with more units than experimental conditions, but not the whole sample, are available to be randomized to conditions
5. trickle process randomization
: refers to cases in which units trickle in slowly and assignment must be made from batches that are smaller than the number of conditions. -- the key problem is to ensure that desired proportions are assigned to conditions over time
6. adaptive randomization strategies
: methods for correcting imbalances in the desired proportions assigned to conditions by changing the proportions over time. -- can be unbiased if the units characteristics remain stable over time; if not, certain analytic adjustments may be required
7. random assignment from matches or strata
: placing units into groups and then assigning separately for each group. -- in matching, groups contain as many units as conditions; in stratifying, groups contain more units than conditions. -- always helps control proportion assigned to conditions. -- if matching or stratifying variable is related to outcome, this can also increase the power of the design. -- useful to match or stratify on variables expected to interact with treatment
7 lessons about implementing random assignment
plan in advance how to explain the nature and purpose of randomization to those who will be affected, how to respond to various arguments about why randomization could or should not be done, and how to provide incentives for doing randomization
pilot test the randomization procedure to discover problems that can be remedied with further planning
develop clear procedures for implementing, controlling, and monitoring the randomization process throughout the entire experiment
have meetings at which to negotiate the randomization procedures with those who will be affected by them
develop fallback options that can be used to bolster estimates of program effects in the event that randomization fails
take advantage of naturally occurring opportunities that facilitate the conduct of randomization
carefully examine the match between the proposed design and those factors that will make randomization more likely to be successful in the particular context of the experiment
before an experiment be conducted it should go through these phases
the present conditions need improvement
the proposed improvement is of unclear value
only an experiment could provide the necessary data to clarify the question
the results of the experiment would be used to change the practice or policy
the right of individuals would be protected in the experiment
recruiting participants to be in the experiment
pre-experimental surveys to locate and characterize potential participants
pipeline studies to follow what happens to them over time
pilot test of the solicitation procedures to see who will learn about the experiment and who will attend if eligible
trained outreach specialists who aggressively recruit potential participants
efforts to learn what features of the intervention cause people to decline to enroll
if the number of participants is lower than expected a research can so the following
extend the time frame for the experiment if time and resources are available and if the programs staff is willing
divert additional resources to intensified outreach efforts
alter eligibility requirements so that more participants are eligible, though this may require testing interactions between eligibility requirements and treatment
reduce the proportion assigned to treatment if power analysis suggests it would still be feasible to find an effect
terminate the experiment, which is sometimes better than spending funds on an experiment that cannot find the expected treatment effect, especially if the decision is made early in the life of the experiment
when recruiting participants to be in random and not random experiments follow these three steps
a lottery condition in which participants are randomized to both treatment and control groups
an invitation to treatment in which participants are offered the chance to participate in treatment
an invitation to control in which participants are offered the chance to participate in the control