Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
22 Cards in this Set
- Front
- Back
correlational research
|
independent variable is not manipulated by the experimenter. Instead research relies on finding natural variation in the independent variable or dependent variables. Not defined by whether you work with correlation coefficients.
|
|
one reason for doing experimental design
|
to stop variables correlating
|
|
collecting data
|
questionnaires, archives and official statistics, observations. Psychological scales are often used
|
|
analyzing the data:
correlation coefficients |
Spearman rank correlation coefficient (Ρ (Greek Rho)) & the Pearson coefficient (r)
|
|
key concept of correlation
|
variance accounted for (r²) sometimes written as a %
|
|
linear regression
|
the line that produces the smallest squared error
Y= A+BX Y= dependent variable X= independent variable A= a constant B= a constant (slope of line) Y'= predicted value |
|
standardized formula
linear regression |
means subtracted from them and then divided by their standard deviation.(written in small letters y and x)
y = βx β=r |
|
multiple regression
|
better prediction more than 1 independent variable (Xi)
Y/Y'= A+B1X1 + B2X2 + ... +BnXn |
|
multiple regression
Standardised |
y/y' = β1x1 + β2x2+ ... + βnxn
|
|
main reasons to use multiple regression
|
separate out the effect of the different independent variables so tell which are causally important. attempt to help internal validity (third variable) statistically. used in field setting where you have less experimental control.
|
|
beta weights
|
measures of what the correlations would be if all the other variables in the equation were held constant
|
|
assumptions of multiple regression
|
1)independence
2)linearity 3)normality 4)nature of independent variables 5)identify key variables |
|
assumptions of multiple regression
1)independence |
observations/cases are independent of eachother
|
|
assumptions of multiple regression
2)linearity |
the relationship between the dependent variable and each of the independent variables should actually be linear than than peaked.
|
|
assumptions of multiple regression
3)normality |
dependent variable should be normally distributed
|
|
assumptions of multiple regression
4)nature of the independent variables |
multiple regression is used when independent variables are continuous or they are are a mixture of categorical and continuous variables. (categorical 1,0 code etc how you code, makes a diff)
|
|
assumptions of multiple regression
5)identify key variables |
in general beta weights will change every time you put in new independent variables or take out old ones. If you miss some out then you will not be measuring the true real world beta weights etc.
|
|
causal modelling
|
correlation does not = causation. some correlation can only go one way (time)
|
|
guidelines for causal modelling
|
1.time
2.A doesn't have to measured before B 3.causing-> caused not caused-> causing |
|
multiple regression and causal modelling
|
helps us distinguish between causal scenarios. by signif beta weights
|
|
path analysis
|
one variable can be dependent and independent
|
|
moderation
|
e.g. contraceptive sex
| = ----->pregnancy |