research methods

quantitative psychologcal categories

difference, association, prediction

sample

a group of people taken from the population

inference

a conclusion formed by combining known facts, evidence, and background reasoning rather than relying on direct observation

distribution of data

values of data in an experiment (measured by central tendency)

normal distrubition

2s rule of thumb

in a normal distribution, 95% of scores fall between 2 sd of the mean, everything inside this is normal

how can a distribution of data be fundamentally described

central tendency and variability

how to know if entire sample is extreme compared to population

1. find sample mean

2. create distribution of sample means

3. use 2sd rule of thumb to find whether typical or extreme

central limit throem

tells us the precisde characterismtics of any distribution of sample means, mean of sample means is same as population mean, sample size over 30 = normal shape

standard deviation of distribution of standard means (standard error)

as sample size increases, standard error decreases (pop mean becomes more precise)

hypothesis

a statement that predicts an effect (difference or association)

the null hypothesis (H0)

predicts there is no effect, only either null or experimental hypothesis can be supported

null hypothesis significance testing

1. propose a null hypothesis that a pop mean has a particular valye

2. proceed assuming null hypothesis is true

3. determine probability of sample mean occurring if null hypothesis is true

4. is probability is small, reject null, if large, do not reject null

alpha level - level of significance

5% (0.05), how we determine which sample means are expected/typical if null hypothesis is true, corresponds to 1.96sd from the mean

single sample z test

determes in a single sample mean differs from specified value (requires population sd), tells us if findings can be generalised to population, if z score above 1.96, reject null, if before, do not reject null

single sample t test

same as z test, used when sd not known, use sample sd instead of population, if t score not extreme for degrees of freedom, do not reject null, if extreme, reject

degree of freedom

sample size - 1 (used for t test)

research designs for independent t test

participants are assigned to multiple groups e.g control/treatment, do the groups have a difference?

independent groups t test

when comparing 2 groups that did same thing, determines if they are the same or extremely different, tells us if these results can be expected in population

repeated measures t test

one group completes thing over 2 time periods, determines if results are the same or extremely different, tells us if this can be expected in population

research designs for repeated measures t test

participants are measured on 2 or more occasions, is there change across time?

research designs for corrolation test

relationship between 2 variables, each participants provides 2 pieces of data, observation

corrolation r

determines if scores on 2 variables are related, and strength of observed findings, is there a linear and symetrical association (pos = both y/x increase)

strength of r

0.1 = weak, 0.3 = moderate 0.5 = strong