Skip to main content
Search form
  • 00:00


  • 00:11

    STEVE TAYLOR: Rural North Carolina.Researchers from Duke University were working thereon a long-term study of mental health problemsin a sample of nearly 1,500 children,about a quarter of whom were from the Cherokee Reservation.

  • 00:25

    CLAIRE PARSONS: In 1996 during the study,the Cherokee Indians in North Carolina's Great SmokyMountains opened a casino, but thiswas a casino with a difference.It was different, because some of the profits from this casinowere distributed equally amongst families on the reservation.[Claire Parsons] And this lifted many of them out of poverty.

  • 00:43

    STEVE TAYLOR: And this was an absolutely fantasticopportunity for the researchers.[Dr. Steve Taylor] Because this new money was independentof anything the families had done for themselves.So it was literally a brand new factor.And this gave the researchers, Jane Costelloand her colleagues, a unique chanceto test if ending family poverty hadany effect on the mental health of the children.

  • 01:03

    STEVE TAYLOR [continued]: [Jane Costello]

  • 01:04

    CLAIRE PARSONS: The findings surprised even the researchersthemselves.In just four years after the study began,children whose families had moved out of povertyshowed a 40% reduction in conduct and oppositionaldisorders.["Those moved out of poverty showed 40% reduction in conductand oppositional disorders."]But there was no significant changein conduct of Cherokee children whohad not been living in poverty.["No significant change for those who had not been livingin poverty."] And this led researchers to conclude that

  • 01:25

    CLAIRE PARSONS [continued]: poverty was a factor in some child mental health problems.

  • 01:29

    STEVE TAYLOR: As the title of the paper shows,this is an example of what's called a natural experimentin psychology.

  • 01:34

    CLAIRE PARSONS: One of the ways youcan show you have a good understanding of what'smeant by a natural experiment is to compare it to other methods.[natural experiment]

  • 01:42

    STEVE TAYLOR: Well, in a way, natural experimentsare a bit like naturalistic observation.[natural experiments and naturalistic observation]The researchers are just observing stuff as it happens.But the crucial difference is that natural experimentsare also a specific test of the effectof one variable on another.

  • 02:02

    STEVE TAYLOR [continued]: So in Costello's experiment, the independent variablewas a sudden improved income of many families.In the dependent variable, the effect on their children'sbehavior and their mental health.[Great Smoky Mountains Study, Costello et al.,independent variable-- improved income,dependent variable-- children's mental health]

  • 02:14

    CLAIRE PARSONS: So in that way, it'ssimilar to a field experiment. [natural experiments and fieldexperiments] But the difference hereis that in natural experiments, the researcher has notmanipulated the independent variable.It's just happened.So from here, we can give a definition.

  • 02:27

    STEVE TAYLOR: A natural experimentis a type of experiment conductedin the environment of the participantswhere the independent variable is notmanipulated by the researcher but occurs naturally.["A natural experiment is a type of experiment conductedin the environment of the participants wherethe independent variable is not manipulated by the researcherbut occurs naturally."]

  • 02:39

    CLAIRE PARSONS: It's also important here,to show you know, there are different typesof natural experiments.For example, they can be longitudinalwhere the dependent variable is observed over time for changes.[longitudinal]Or they might be cross-sectional.[cross-sectional] For any one of two broadly similar populationsexperiences the independent variable.Or as in Costello's study, for example, there

  • 03:00

    CLAIRE PARSONS [continued]: might be a combination of longitudinal andcross-sectional data. [longitudinal and crosssectional]

  • 03:04

    STEVE TAYLOR: So when you're writingabout natural experiments, it's important to rememberthat they don't just happen on the spur of the moment.For example, knowing TV was going to be introducedinto Fiji in 1995, researchers studying eating disorders wereable to study a sample of girls' and women's perceptionsof their body images before and after TV.[Becker et al., 2002, "Eating behaviours and attitudesfollowing prolonged exposure to television among ethnic Fijian

  • 03:25

    STEVE TAYLOR [continued]: adolescent girls"]Alternatively, like Jane Costello and her colleagues,you're doing a piece of research when some dramatic change justtakes place.So she had lots and lots of data on the familiesbefore the casino opened.And she was then able to monitor changes in these familiesafter the casino.That's why the study was experimental and notjust correlational.

  • 03:45

    CLAIRE PARSONS: Natural experimentsare often used in psychology.So what do they give researchers that other methods don't?What are their strengths?

  • 03:54

    STEVE TAYLOR: Well, the main strength of natural experimentsis they enable researchers to do things,that for practical and ethical reasons,they couldn't possibly do in a lab,or even in a field experiment. [enables research that couldn'tbe done in lab or field]

  • 04:06

    CLAIRE PARSONS: For example, looking at Costello's research,you couldn't get hold of lots of families,dump them into poverty, and then take them out again,just to see what would happen.But as it was happening anyway, researchershad a great opportunity to observe the effects.But natural experiments aren't just done out of necessity.They've got other strengths.

  • 04:26

    STEVE TAYLOR: Another major one is high ecological validity.[high ecological validity] Not onlydo they take place in real, everyday settings,but nothing has been manipulated by the researcher.There's nothing artificial.Things are just as they are.

  • 04:39

    CLAIRE PARSONS: There's also less chanceof demand characteristics, as participants are usually eitherunaware the research is taking place, or unaware of its aim.[less chance of demand characteristics]

  • 04:49

    STEVE TAYLOR: For example, the families in Costello's researchwere not aware of the researchers'interest in the casino money.Also, natural experiments often involve large numbersof participants who didn't have to be recruited,reducing time and cost. [large numbers, less time and cost]

  • 05:03

    CLAIRE PARSONS: So there's an opportunist aspectto natural experiments.As with many things in life, opportunity comes along,and you have to recognize it, and grab hold of it.It's been described by one researcheras like finding gold in a stone.But, of course, like all methods,it comes with possible limitations or weaknesses.

  • 05:22

    STEVE TAYLOR: The major weakness of natural experimentsis that it's much harder to establisha clear relationship of cause and effect,as the researcher has so little controlover possible confounding variables.[lack of control over confounding variables]

  • 05:33

    CLAIRE PARSONS: And for some psychologists,natural experiments aren't really experiments at all.They lack reliability, as they'revirtually impossible to replicate. [cannot replicate]

  • 05:42

    STEVE TAYLOR: And as participants are usually eithernot aware they're taking part in experiment, or at leastwhat its real purpose is, there were usuallyethical issues involved. [ethical issues][MUSIC PLAYING]

Video Info

Series Name: Experimental Research Methods

Publisher: ShortCutstv

Publication Year: 2017

Video Type:Tutorial

Methods: Experimental design, Quasi-experimental designs

Keywords: mental health; psychology

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Natural experiments are experiments that are not done in a lab. These experiments take place with nothing manipulated by the researcher.

Looks like you do not have access to this content.

Natural Experiments

Natural experiments are experiments that are not done in a lab. These experiments take place with nothing manipulated by the researcher.

Copy and paste the following HTML into your website