If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

### Course: Statistics and probability>Unit 6

Lesson 5: Experiments

# Introduction to experiment design

Introduction to experiment design. Explanatory and response variables. Control and treatment groups.

## Want to join the conversation?

• I have never heard the terms Explanatory and response variables. They are usually called predictor\independent variable and dependent or outcome variable.
• According to wiki there are many terms used for the same variables. One of the factors is the context in which these terms are used.
• What exactly is a lurking variable?
• If I understand Sal correctly, it's just what he calls a variable that you didn't plan for in your experiment. For example, say you're trying to see if birds' feather colors make them faster, and you have a group of blue birds and a group of red birds. But you don't take into account that in some species of birds the females are larger and therefore slower, and so the "lurking variable" of the sex of the bird throws off your experiment.
• What´s the difference between a replicate and a repetition?
Are there any definition of a "replicate"?
• Replication is the strict repetition of an experimental condition so that the variability associated with the phenomenon can be estimated. It assumes that we can repeat this experiment in every detail. In formal definition "the repetition of the set of all the treatment combinations to be compared in an experiment. Each of the repetitions is called a replicate."
• What if...a Drug Company wants to test a new (expensive & difficult to deliver) drug. They decide to do this:
Phase 1: give all subjects a placebo. Remove any responders.
Phase 2: give remaining subjects the drug. Analyze effect.

Is this a legitimate design because all placebo responders have been removed at the start of the study?
Is this a more ethical design?
• Sadly, the placebo effect doesn't work like that.
It's not like an on/off switch where you either get a response or not.
Also the effect tends to get stronger over time, so the subjects that didn't show a strong response in phase 1 may develop one in phase 2.
• So what is the difference between a block design and a SRS? Is it just that the block is used in an experiment and the SRS in used in a survey?
• you can use an SRS in an experimental design. Block design are for experiments and a stratified sample is used for sampling. Blocking implies that there is some known variable that can affect the response variable or the overall experiment. In the video the example would have been gender because maybe there were more men in the treatment group than the control group and women would react differently than men to the pill.
(1 vote)
• At , Sal talks about a triple-blind experiment where neither the people taking the pill, the person giving them the pill, nor the ones analyzing the data know which one it is. But if nobody knows which pill is given to each person, how can you test for causality between the medicine and the A1c levels?
(1 vote)
• In a triple-blind experiment, neither the participants, the individuals administering the interventions, nor the researchers analyzing the data know which group received the treatment or control condition. Despite the lack of awareness about the assigned interventions, researchers can still test for causality by comparing the outcomes between the treatment and control groups. The key is to ensure that the only systematic difference between the groups is the intervention being studied. By controlling for other variables and randomizing group assignment, researchers can infer causal relationships between the intervention and the outcomes of interest.
(1 vote)
• what are the differences between grouping and sampling? should get a sample from a population, and then do an experimental and control group
(1 vote)
• Grouping and sampling are different concepts in experimental design. Sampling involves selecting a subset of individuals or items from a population to participate in the study. Grouping, on the other hand, refers to the division of participants into different categories or conditions, such as control and treatment groups, based on the research design. While sampling ensures that the study sample is representative of the population of interest, grouping allows for comparison between different conditions to assess the effects of interventions.
(1 vote)
• "...you could do really a version of stratified sampling that we've talked about in other videos, which is you could do what's called a block design for your random assignment where you actually split everyone into men and women..." Sal is rambling and this is confusing. To say that something is a "version of" is to say it is a synonym. This audio doesn't make a clear distinction between stratified sampling and block design. Could you please fix this video?
(1 vote)
• Sal's explanation of stratified sampling and block design could be clarified to differentiate between the two concepts. Stratified sampling involves dividing the population into homogeneous subgroups (strata) based on certain characteristics (e.g., gender) and then randomly selecting samples from each subgroup. Block design, on the other hand, involves grouping participants into blocks based on specific variables (e.g., gender) and then randomly assigning treatments within each block to ensure balance across treatment groups. Clarifying these distinctions would enhance understanding.
(1 vote)
• Can you not have a placebo group but rather another drug which previously had placebo controlled trials?
(1 vote)
• While a placebo group is commonly used in clinical trials to control for the placebo effect and assess the efficacy of a new treatment, it's possible to compare the new treatment to an existing standard treatment instead of a placebo. In such cases, participants in the control group receive the standard treatment, which has already undergone placebo-controlled trials to establish its efficacy. This approach allows researchers to evaluate the relative effectiveness of the new treatment compared to the established standard, providing valuable insights for clinical practice.
(1 vote)
• An explanatory variable is what you manipulate or observe changes, while a response variable is what changes as a result.
(1 vote)
• An explanatory variable is indeed the variable that is manipulated or observed to see changes, while a response variable is the variable that changes as a result of the manipulation or observation. This distinction is crucial in experimental design as it helps researchers identify causal relationships between variables and understand the effects of interventions or treatments on outcomes.
(1 vote)