This question deals with blinds/double blinds. When a study is blinded, it means that information is either withheld from the participants or from the researchers in order to correct for certain biases. In this case, the lab assistant is NOT blinded; they know which coffee went into which cup. You can imagine a scenario where the lab assistant could influence the participant's reaction towards each coffee by subtly grimacing at a coffee brand the lab assistant doesn't like.

---------------------------

This question deals with causality. There are many observations you can make that happen together by coincidence, these observations are said to be correlated. However, when one thing directly causes something else to be affected, that relationship is causal.
A researcher finds that larger parks tend to have fewer endangered species. Did the researcher prove that increasing park size directly leads to fewer endangered species? In other words, did the researcher actually do an experiment, say by increasing park size and counting the number of endangered species over time?

At this stage, there is insufficient information to conclude that this relationship is causal. You could imagine a scenario where larger parks require larger budgets to maintain, and having better funded parks could lead to endangered species over time. There's just no way to know for sure without doing some testing.

---------------------------

This question deals with confounding variables. In an ideal experimental set up, you would only change one aspect of the experiment at a time, so that you can properly attribute an experimental outcome to that aspect.

In this case, what the researchers care about is puzzle solving speed. The researchers changed both the room temperature AND the presence of music. Was it the room temperature or music or both that lead participants to solve the puzzle faster? We don't know. Next time, keep room temperature constant and change whether a radio is present, or just change the room temperature.

---------------------------

This question deals with control groups. A control group refers to a group of research participants who are treated the exact same as participants who are given an intervention or treatment, except for ACTUALLY getting that intervention/treatment. This allows for treatment and non-treatment groups to be compared in a clear way.

You can imagine a scenario in which Scientist A's decision to give ALL the participants acne cream could make it difficult to actually tell if the cream worked. If all of Scientist A's participants had lower acne after treatment started, we would not be able to tell, for example, whether the acne would have gone away on its own over time. Scientist B does not have this problem, as their experiment is controlled. 

---------------------------

This question deals with ecological validity. In an ideal world, our experiments in the lab would match perfectly to real life. If only things were that simple. You can imagine a scenario where participants in a research study value tokens less than they would say, their hard-earned real cash. While results from this experiment may be useful, it would be an extremely gutsy claim to say that this experiment reflects real life.

---------------------------

This question deals with history. Generally, the closer you get to studying something resembling real life, the harder it is to parse what affects what. In this case, a famous celebrity dying of disease A could have contributed to public awareness of disease A. Unless a survey specifically is designed to figure out what factors contributed to increased awareness of disease A, there is no way to know the total contribution of the media campaign. Maybe the media campaign was 100% the reason why awareness is increased. Maybe 50%. Maybe none at all. If all the survey says is that awareness is up, it's not possible to tell how much a sudden change in history could have affected things.

---------------------------

This question deals with maturation. Maturation refers to the experience a participant has in a study over time. One possible explanation as to why participants struggle to perform later on in this experiment is that the experiment does in fact get harder. But another possible explanation as to why participants struggle later on is that the experiment may be dragging on for too long and causing people to get bored or tired, causing more errors.

---------------------------

This question deals with validity. Validity refers to how well evidence supports conclusions. In this case, the conclusion the researcher makes is that general math ability can be measured by administering a geometry test. You can imagine a scenario where there are geometry whiz kids who can't seem to figure out precalculus, or the reverse - kids who can't grasp geometry, but absolutely demolish precalculus. Is it fair to either one of these types of kids to have their general math ability tested by a geometry test?

---------------------------

This question deals with random assignment to condition. Remember that the definition of a control group is that they are treated exactly the same as the treatment/intervention group, except for the treatment/intervention itself. Here, by assigning all the overweight children to the intervention group, another confounding variable is introduced.

You can imagine a scenario where a health intervention, say, daily speedwalking at recess, has a weight loss effect for overweight children, but not in school children overall.

---------------------------

This question deals with reliability. Two ways that you can describe the quality of your data are its accuracy and its precision.

Accuracy refers to whether your data, on average, reflect reality. You can imagine a bullseye target. If I fired 3 arrows that landed in a perfect, regular triangle in the outermost ring, the average position of the arrows would be the bullseye.

Precision refers to whether your data points are close to each other. In the previous example, the data is accurate, but NOT precise, because if you are not consistently aiming at the same spot, your data points won't be close to one another.

If I shot three arrows at a target, and all of them landed in the outer ring, but all of the arrows landed right next to one another, then my data would be precise but NOT accurate - as all of my arrows, despite being so close to one another, are all far away from the bullseye.

Just because a method is accurate, doesn't make it precise. Just because a method is precise, doesn't make it accurate.

---------------------------

This question deals with response bias. There are many subtle factors that can play a role in how we, as participants, act in experiments. You can imagine a scenario in which framing someone's mind to think positively about something (I am satisfied) may cause them to respond to a survey differently than framing their mind to think negatively about something (I am dissatisfied).

---------------------------