True experiment requirements

C

ceperras

Guest
#1
I know this is more of a R&D question, but it is for a statistics class.

If subjects are randomized to 1 of 4 groups, none of which is a control group, is it still a true experiment? I'm trying to distinguish between a true experiment and a quasi-experiment and need to know if the presence of a control groups is a necessity for a true experiment (rather than just typical).

THANK YOU!
 

Dragan

Super Moderator
#2
I know this is more of a R&D question, but it is for a statistics class.

If subjects are randomized to 1 of 4 groups, none of which is a control group, is it still a true experiment? I'm trying to distinguish between a true experiment and a quasi-experiment and need to know if the presence of a control groups is a necessity for a true experiment (rather than just typical).

THANK YOU!

Essentially, the difference between an experimental design and a quasi-experimental design is that the latter lacks randomization. Thus, even though there is no control group per se, given your example, it's still considered to be an experimental design - because subjects were randomly assigned to the various treatment levels (or groups).
 
C

ceperras

Guest
#3
Follow-up

Thank you, Dragan! That helped immensely. I have one further question along the same line.

Say there is an experiment with four conditions. Each subject is exposed to each condition but the order of exposure is randomized. Even though participants are not randomized to one of the four groups, is this considered a "true experiment"?

Thank you!
 

Dragan

Super Moderator
#4
Thank you, Dragan! That helped immensely. I have one further question along the same line.

Say there is an experiment with four conditions. Each subject is exposed to each condition but the order of exposure is randomized. Even though participants are not randomized to one of the four groups, is this considered a "true experiment"?

Thank you!

Yes.

Look, what you're describing is a basic repeated measures experimental design.

Let me provide an example to (perhaps) help.

Suppose I have random sample of N subjects.

And, each of the N subjects is exposed to the following 4 treatment conditions (exposure to temprature): -30 degrees below freezing, -10 degrees below freezing, +10 degrees above freezing, +30 degrees above freezing....Mkay.

Now, the dependent variable is the measure of Pulse Rate on each of the N subjects.

All that is needed is to assume (classicly) that the N subjects observations are independent, normally distributed, and the variance-covariance matrix is equal (sphericity) accross the 4 treatment conditions.

If this is not enough, then you might want to provide some more specific details.
 
Last edited:
C

ceperras

Guest
#5
true randomized experiment - example provided

Thank you so much. You have been so helpful to me! I see what you are saying but have a further concern regarding the importance of the order. If you don't mind, I will paste below an example from a study with all the relevant study information. Would you consider this a true randomized experiment?

Each participant heard three cross-species adoption stories, each of which was followed by a series of questions about the offspring, including its inheritance of properties, kindhood, and the effect of a blood transfusion. To familiarize children with the format of this task, the experimenter first told them a warm-up story, explaining, “A baby deer grew up with other deer in the forest. When the baby is all grown up, will it drink water or coffee? Will the baby be brown or green?”

Each of the three cross-species adoption stories involved a baby animal that had been separated from its birth parent and raised exclusively by an adoptive parent from a different species. These stories involved the following species pairs: turtle–toad, cow–pig, and pigeon–turkey. For each participant, the animal serving as the birth parent was assigned randomly.

Participants were asked explicitly to judge the kindhood of the offspring: “Now that the baby is all grown up, what kind of animal is it? Point out the kind of animal the baby grew up to be” [drawings of the two parent animals are shown].

Next, the experimenter introduced the blood transfusion scenario, saying, "When the baby was growing up, it became sick. A doctor came and, with a needle, took out all of the old blood that the baby got from its mother [the drawing of the birth parent is shown] when it was born. The doctor then went to the animal that was taking care of the baby [the drawing of the adoptive parent is shown] and took some of its blood to give to the baby. So the baby got all new blood like the blood of the pig."

The experimenter then asked, “Now that the baby is all grown up, what kind of animal is it? Point out the kind of animal the baby grew up to be” [drawings of the birth and adoptive parents are shown].

Each participant completed this procedure for all three cross-species pairs.


In this situation, comparisons are not being made between the different species pairs, so does the random assignment make this a true randomized experiment?

Thank you again for taking the time to help me!!!! I appreciate it more than I can express!
 
Last edited by a moderator: