The Classical Experiment

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:00

      [MUSIC PLAYING][The Classical Experiment]

    • 00:10

      RYAN C MELDRUM: Hello.My name is Professor Ryan Meldrumand I'm going to assistant professorin the Department of Criminal Justice at FloridaInternational University.[Dr Ryan C Meldrum, Assistant Professor]In this tutorial, I'm going to be discussingwhat the purpose of a classical experimentis, some examples of classical experiments,and then I'll explain the componentsof the classical experiment for you.[Purpose of the Classical Experiment]

    • 00:36

      RYAN C MELDRUM [continued]: The main purpose of the classical experimentis to be able to establish a cause and effectrelationship between two variables, whatwe call the independent and dependant variable.Although there are several other types of research designsthat can be used to help researchers establish causeand effect, most researchers considerthe classical experimental design

    • 00:57

      RYAN C MELDRUM [continued]: to be the gold standard to be able to establishcause and effect.The reason why this is such a powerful research design--and why it can help researchers to truly establish a causeand effect relationship-- is that itcan help us to satisfy the three requirements of causation,the first being you need to have a correlation establishedbetween your independent and dependent variable.

    • 01:19

      RYAN C MELDRUM [continued]: The second requirement that needs to be establishedis that the independent variable proceeds, or comes before,the dependent variable.And third, we have to be able to establishthat the correlation between our independent and dependentvariable is not due to some third variable.[Examples of the Classical Experiment]

    • 01:43

      RYAN C MELDRUM [continued]: So now that we've briefly just discussedthe purpose of the classical experiment,let's talk about some examples without usingany of the technical language.So let's begin by thinking about an experiment in which we wantto try to establish whether or not a new experimental dietpill can help people to lose weight.In each of the examples that I go through,

    • 02:04

      RYAN C MELDRUM [continued]: I'll be drawing up on the board some diagramsof what the research design of a classical experimentwould look like.And so in the example of wanting to establish whether or nota diet pill helps people to lose weight,researchers would start by seeking outindividuals who would be interested in this typeof study, who might be overweightor might have been clinically diagnosed as obese.

    • 02:26

      RYAN C MELDRUM [continued]: So we'll start by just drawing a simple pool of individuals.And so in this case, you might have,let's say, 200 individuals who are all wanting to lose weight.Well, the first thing that the researchers would do

    • 02:46

      RYAN C MELDRUM [continued]: is they would split out this large group of individualsinto two smaller groups.And so they'll split that group outinto roughly two groups that are about the same size.

    • 03:09

      RYAN C MELDRUM [continued]: Now this is where the experiment starts to take place.What will happen is that the researchers will assignone of the groups to receive the experimental diet pilland the other group will not receive the diet pill.

    • 03:38

      RYAN C MELDRUM [continued]: So we have our top group here thatis receiving the diet pill and the bottom grouphere who is not going to be receiving the diet pill.And let's imagine for the sake of this illustrationthat the top group is going to be taking the diet pillfor a period of six months.So that is the length of the experimental condition.

    • 04:09

      RYAN C MELDRUM [continued]: So what we've done is we've followed the two groupsfor a period of roughly six months,with the top group receiving the diet pill and the bottomone not.The thing that we would be measuring to see whether or notthe diet pill helped people to lose weightwould be how much they weighed at the beginningof the study and then how much they weighedat the end of the study.So at the very beginning of the study,

    • 04:30

      RYAN C MELDRUM [continued]: we would measure everyone's weightbefore they did or did not receive the diet pill.And what we would probably find--if we've split these individuals out into two different groups,

    • 04:53

      RYAN C MELDRUM [continued]: based upon a process called randomizationthat we'll discuss in just a little while, that the twogroups would look on average about similar to oneanother in terms of their starting weight.Now let's imagine that we found the diet pill was effective.This is what we might find by the end of the study.

    • 05:20

      RYAN C MELDRUM [continued]: What we see here in this example datais that by the end of the six month period,the members in the top group, the members whoreceived the diet pill, on averagehave lost a significant amount of weight.On average, they started out at 177 pounds,and by the end of the six month period,they now weigh on average 160 pounds,while the individuals who did not receive the diet pill

    • 05:42

      RYAN C MELDRUM [continued]: started out on average 175 pounds.And what we see over that six month period isthose individuals actually, on average,gained a couple pounds.So there are two comparisons that wecan make to help us to establish that the diet pill waseffective.The first comparison we can make is between the two groupsat the end of the study.We see that there's a 17 pound difference between the group

    • 06:09

      RYAN C MELDRUM [continued]: individuals who received the diet pill and thosewho did not.And then the second comparison that we can makeis we can actually make the comparison for those people whoreceived the diet pill and how much that group of peoplechanged on average from the start to the end.

    • 06:30

      RYAN C MELDRUM [continued]: And we see in this example data that the group whoreceived the diet pill decreased in weightfrom on average from 177 pounds down to 160 pounds.So this would provide us with some evidenceof a causal effect of the diet pillon helping people to lose weight.Let's move on to a second example.

    • 06:52

      RYAN C MELDRUM [continued]: Let's imagine that we're wanting to evaluate whether or notparticipation in the Big Brothers Big sisters programhelps to reduce antisocial behavior among teenagers.We will still continue with an examplewhere maybe we have roughly 200 individuals that would berelevant to this type of study.In this case, maybe we have 200 teenagers whohave been defined as at-risk.

    • 07:13

      RYAN C MELDRUM [continued]: Maybe they already have some historyof involvement in the juvenile justice system,or they have some police contact,or they've been suspended or expelled at different schools.So in this case, we start out with 200 teenagerswho would be relevant for this type of study.Just as in the first example, we would

    • 07:33

      RYAN C MELDRUM [continued]: split that group of teenagers out into two smallergroups of roughly 100.In this case, we're not evaluatingthe effect of a diet pill, but rather participation in the BigBrother Big Sister program.So the 100 teenagers in the top group

    • 07:54

      RYAN C MELDRUM [continued]: would be those who would be assigneda big brother or a big sister whomwe think would help to mentor them and perhaps maybe reducetheir involvement in antisocial behavior,while the teenagers in the bottom grouphere would just essentially not receive anything at all.Let's imagine we're evaluating this Big Brother Big Sisterprogram over the period of 12 months.

    • 08:15

      RYAN C MELDRUM [continued]: So in this case, at the very beginning of the study,we would measure their involvementin antisocial behavior and then we would measure it againat the end of the study.So let's imagine that we are measuringtheir involvement in antisocial behaviorby having the teenagers self-report to usthe number of delinquent acts that they've committedin the previous 12 months.

    • 08:37

      RYAN C MELDRUM [continued]: So at the beginning of the study here,we would be asking them to reportto us how many delinquent acts they'vecommitted in the 12 month period priorto the beginning of the study.And so let's imagine that we findthe following to be the data that getsreported to us by the teenagers.

    • 09:01

      RYAN C MELDRUM [continued]: At the beginning of the study, wesee that the teenagers in the top grouphave reported to us that they've committedon average 5.4 delinquent acts.And if this is a group of at-risk teenagers,that might be a very reasonable average that we would find.In the bottom group, we see those teenagershave reported on average 5.1 delinquent acts in the last 12

    • 09:21

      RYAN C MELDRUM [continued]: months.So we would make that exact same measurementat the very end of the study, once the 12 month period haspassed and the members in the top grouphave received the Big Brother Big Service program.So if we were to find that the Big Brother Big Sisterprogram was effective, we might see this type of result.

    • 09:50

      RYAN C MELDRUM [continued]: At the end of the 12 month period,we see that the teenagers who were assignedto big brother or big sister reportedthat they had committed on average 3.1delinquent acts in this 12 month period of time.And we see that the teenagers who did notget assigned a big brother or big sisterreported on average 4.9 delinquent acts

    • 10:12

      RYAN C MELDRUM [continued]: in the last 12 month period.So, just as in our first example with the diet pill,we can again make two comparisonsto help us to establish whether or not this type of programhas reduced involvement in antisocial behavior.The first comparison we make is the differencebetween the two groups in their overall involvement

    • 10:34

      RYAN C MELDRUM [continued]: in antisocial behavior at the end of the study.And we can see the teenagers who receivedthe big brother or big sister reported3.1 delinquent acts, where those teenagers who did notare reporting 4.9.And so we can see that there's a significant differencein the overall amount of delinquency being committed

    • 10:55

      RYAN C MELDRUM [continued]: between the two groups.And then the second comparison that we can makeis, again, how much has that delinquent behavior changedamong those teenagers who got the big brother or big sister.And we can, again, see in this datathat the teenagers have dropped from 5.4 delinquent acts per 12

    • 11:17

      RYAN C MELDRUM [continued]: months down to 3.1 delinquent acts per 12 months.So let's talk about one final examplethat comes from a very emerging area of research knownas biosocial criminology, where researchers have startedto reinvestigate the biological, andneurological, and genetic correlatesof delinquent behavior.

    • 11:39

      RYAN C MELDRUM [continued]: So let's imagine for this particular examplethat researchers are wanting to determine whether or nottaking in omega-3 fatty acids as a supplementhelps teenagers to reduce their overall aggressive behavior.So in this case, we could again startwith a hypothetical pool of teenagers of approximately 200.

    • 12:01

      RYAN C MELDRUM [continued]: We would randomly divide that group into our two smallergroups of about 100.And in this case, the top group wouldreceive the omega-3 supplementation--which can be done through something like a pill--

    • 12:22

      RYAN C MELDRUM [continued]: while those individuals in the bottom groupwill not be receiving the supplementation.So, to a certain extent, this exampleis more similar to the first example with the diet pill,but we're not looking at the effect of a diet pill,but rather omega-3 supplementation.So we would again want to have a measurement of aggression

    • 12:43

      RYAN C MELDRUM [continued]: or delinquent behavior at the beginning of the study.And then let's imagine for this examplethat the teenagers are getting the omega-3 supplementationfor a period of six months.And let's say that we're going to measure aggressionby having the parents of the teenagersprovide us a self-report by filling out some type of survey

    • 13:04

      RYAN C MELDRUM [continued]: questionnaire that could then give usan overall indicator of the amount of aggressive behavioror aggressive tendencies that each of the participatingteenagers has.So let's imagine that at the very beginning of the studywe find that the parents report that the individualsin the top group have a score ranging from zero to 10--

    • 13:26

      RYAN C MELDRUM [continued]: the average is at 5.5-- and that the teenagersin the second group, who are not going to be receivingthe supplementation, their average score, as reportedby their parents, is 5.2.The top group of teenagers receives

    • 13:47

      RYAN C MELDRUM [continued]: the omega-3 supplementation, the bottom group does not,and we are tracking them over a six month period of time.At the end of the six months, let's imagine that the datapanned out this way.

    • 14:07

      RYAN C MELDRUM [continued]: We see at the end of the study that the individuals whoreceive the omega-3 supplementationhad an average aggression score of 3.5and that among the individuals who did notreceive the omega-3 supplementationthey have an average score of 4.7.We again can make our two comparisonsto determine whether or not the omega-3 supplementation helped

    • 14:31

      RYAN C MELDRUM [continued]: to reduce aggressive tendencies in this sample of individuals.We see the average score of 3.5 is significantly smallerthan the average score 4.7 between the two groups.And we see a decrease in the overall average score

    • 14:52

      RYAN C MELDRUM [continued]: for aggression among those teenagerswho received the omega-3 supplementation from 5.5all the way down to 3.5.[Components of the Classical Experiment]Now that I have covered some examples of studiesthat would take the form of a classical experiment,let's break down the components of a classical experiment

    • 15:16

      RYAN C MELDRUM [continued]: and discuss the specific terminology that we associatewith each one of these components.In a classical experiment, the independent variableis commonly referred to as the experimental stimulusbecause the researchers believe that it will help to stimulatechange in the outcome.The dependent variable is measured

    • 15:37

      RYAN C MELDRUM [continued]: at the beginning of the study and thenat the end of the study.And these two points in time are calledthe pretest and the posttest.

    • 16:05

      RYAN C MELDRUM [continued]: So in the three examples we worked through,there were three different independent variables, or threedifferent experimental stimuli, thatwere discussed-- the diet pill, the big brother or big sisterprogram, and the omega-3 supplementation.And so you'll see I put in this boxthe placeholder, noting that thatis the independent variable or the experimental stimulus.

    • 16:28

      RYAN C MELDRUM [continued]: With regard to the dependent variable,I was making mention of the measurement beingat the beginning of the study and thenat the end of the study.And then at the beginning of the study, we call it the pretest,and at the end of the study, we call it the posttest.And in our three examples, we had both a pretest and posttestmeasurement of how much someone weighed,

    • 16:50

      RYAN C MELDRUM [continued]: their involvement in delinquency,and their overall score on aggression.Now with regard to the individualsthat we studied in the classical experiment,I've described in each one of the examplesthat there is a top group that receivesthe experimental stimulus and a bottom group of participantsthat does not receive anything.

    • 17:10

      RYAN C MELDRUM [continued]: We refer to these two groups as the experimental groupand the control group.It makes sense that we would refer to this top groupas the experimental group, because they're receivingthe experimental stimulus.This bottom group we refer to as the control group.And they are important for us in helpingto establish cause and effect, because that second group

    • 17:31

      RYAN C MELDRUM [continued]: allows us to see what happens in termsof the dependent variable from our pretestto our posttest measurement in the absenceof an experimental stimulus.Because the control group membersare not receiving the independent variableor are not being exposed to the experimental stimulus.

    • 17:53

      RYAN C MELDRUM [continued]: The final component of the classical experimentis how individuals get assigned to either the experimentalgroup or the control group-- whatI've drawn here in terms of the arrows assigning peopleto one of the two groups.This process is known as randomizationor random assignment.We start out with individuals who are relevant to our study

    • 18:17

      RYAN C MELDRUM [continued]: and we use some random process to determinewho is going to become an experimental group memberand who is going to become a control group member.That can be done through flipping a coin,using a random number generator, or even usingsomething like a lottery ball tumblerthat you might see being used like Powerball.So just to recap the main components

    • 18:37

      RYAN C MELDRUM [continued]: of the classical experiment, we start outwith a large group of individualswho are relevant to the study.They get randomly assigned to eitherbe part of the experimental group or the control group.We take a pretest measurement on the dependent variablefor all of the participants in the study.

    • 18:57

      RYAN C MELDRUM [continued]: The experimental group members are exposed to--or receive-- the experimental stimulus,while the control group members do not.And then at the very end of the study,we take our posttest measurement of the dependent variable.And then we make two assessments.Is there a meaningful difference between the two groupsat the posttest?

    • 19:18

      RYAN C MELDRUM [continued]: And has there been a significant change from the pretestto the posttest for our experimental group members?And so these are the components and examplesof the classical experiment.[MUSIC PLAYING]

The Classical Experiment

View Segments Segment :

Abstract

Professor Ryan Meldrum gives an introductory lesson on classical experimentation. His explanation also includes a series of hypothetical set-ups to demonstrate the principles of classical experimentation.

SAGE Video Tutorials
The Classical Experiment

Professor Ryan Meldrum gives an introductory lesson on classical experimentation. His explanation also includes a series of hypothetical set-ups to demonstrate the principles of classical experimentation.

Copy and paste the following HTML into your website

Back to Top