Skip to main content
SAGE
Search form
  • 00:00

    [MUSIC PLAYING]

  • 00:11

    CLAIRE PARSONS: When you're revising correlationsthink relationships.This is the key word or image to keep in your head,because that's what correlations are.They're types of relationships.Correlations measure the extent to which two or more variablesare related, and correlational analysis usuallystarts with a hypothesis, something you'd expect to find.

  • 00:33

    CLAIRE PARSONS [continued]: So for example, I might expect to finda relationship between the amount of timemy students spend on their schoolwork and their grade.

  • 00:42

    SPEAKER: Well, a correlation would find outlooking at the time students spend revising, T,and their grades in class, G. Data are displayed on a scattergraph, and two key words here are direction and strength.Direction is about the nature of the relationship.A positive correlation shows that both variables increase

  • 01:02

    SPEAKER [continued]: or decrease at the same time, while a negative correlationshows that if one variable increases, the other decreases,or their may be no correlation.Correlation coefficients are calculationsthat measure the strength of a relationship,and the closer the number is 2 plus or minus 1,the stronger the correlation.

  • 01:22

    SPEAKER [continued]: So the sign plus or minus tells youthe direction of the relationshipand the number tells you it's strength.A common mistake here is to assumethat negative correlations are somehowless strong or less significant than positive ones,but they're not.For example, a negative correlation of 0.73is a stronger correlation than a positive one of 0.45.

  • 01:44

    SPEAKER [continued]: The strength of the correlation is given by the figure and notby the direction.

  • 01:49

    CLAIRE PARSONS: So correlational analysishas a lot going for it.It gives us precise measurable dataon the relationship between two or more factors,so it can tell us if there is a relationship between revisiontime and grades.And if so is it positive or negative?And what's the strength of that relationship?

  • 02:09

    CLAIRE PARSONS [continued]: But correlations also have important limitations.

  • 02:14

    SPEAKER: Correlations are dependent on the qualityof the data they're analyzing.So for example, if the data had been obtainedfrom a non-representative sample then generalizationis limited no matter how impressive the correlation.

  • 02:26

    CLAIRE PARSONS: But there's an even more important limitationto correlation, one that's often forgottenby the media, government, students, and even sometimesby the researchers themselves.

  • 02:38

    SPEAKER: Correlations just cannot demonstrate causeand effect.For example, even if you get a strong positive correlation,it doesn't prove time spent revising causes the examgrades.

  • 02:51

    CLAIRE PARSONS: So let's illustrate thisby looking at a real and widely reported correlational study.

  • 02:57

    SPEAKER: A few years ago in the United Statesparents were being warned about the dangers of lettingtheir kids watch too much TV, and there seemedto be good reason for this.A lot of TV is violent, and a large scale longitudinalstudy by Johnson and colleagues hadshown the relationship between time spent watching TVas a kid and the likelihood of committing

  • 03:17

    SPEAKER [continued]: an aggressive act as a young adult.The data were clear and consistent.There was a strong positive correlation.The more time watching TV, the more likelyit was that the person would commit an aggressive act.Now, that's a very interesting finding,but despite what some people were suggesting,it doesn't show that TV watching was a causal factor

  • 03:40

    SPEAKER [continued]: in the aggression.Remember the mantra correlation doesn'tmean causation, because both excessive TVwatching and aggression might be the result to another factor,such as an aggressive personality, family dynamics,or socioeconomic status.

  • 03:57

    CLAIRE PARSONS: But if researcherswant to go a stage further and see if there's actuallya causal relationship, then they have to do sounder controlled conditions.That means designing an experiment.

Video Info

Series Name: Non-Experimental Research Methods

Publisher: ShortCutstv

Publication Year: 2017

Video Type:Tutorial

Methods: Correlation, Experimental design

Keywords: practices, strategies, and tools

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:

Keywords:

Abstract

Research correlations can show the relationship between factors, but can not be used to show cause and effect.

Looks like you do not have access to this content.

Correlations

Research correlations can show the relationship between factors, but can not be used to show cause and effect.

Copy and paste the following HTML into your website