intro to SPSS Flashcards
how do you use SPSS to perform a correlation test?
1) have 1 column for each variable (2)
2) analyse
3) correlate
4) bivariate
5) click both variables and press arrow to get into the right hand column
6) choose whether spearman’s or pearsons
7) choose whether one or 2 tailed
8) look at r value and corresponding p value (where it says sig. 2 tailed)
9) if p value less than 0.05 then significant
how do you change the headings of the variables?
click variable view and then change the headings
how to do a scatterplot graph?
click graphs legacy dialog scatter plot simple scatter choose x and y axis
how to perform a related t-test?
1) have a column for each variable and a row for each participant
2) analyse
3) compare means
4) paired-samples t-test
5) click 1st variable and arrow to put in ‘variable 1’
6) click 2nd variable and arrow to put in ‘variable 2’
7) find t value and p value
8) if p value is less than 0.05 then significant
how to perform an unrelated t-test?
1) label first column as ‘name of variable 1_ name of variable 2’
2) label second column as the DV
3) give score of 1 to 1st condition and score of 2 to 2nd condition in 1st column
4) analyse
5) compare means
6) independent samples t-test
7) put DV in ‘test variable’ box
8) put IV in ‘grouping variable’
9) click ‘define groups’
10) group 1 put ‘1’ (as they have a score of 1) and group 2 put in ‘2’
11) look at ‘equal variances assumed’
12) look at sig. to see if p value less than 0.05
how to perform an analysis of variance test? (ANOVA)
1) label first column as the IV condition and assign a score to each group (can handle more than 2 groups)
2) label second column as DV
3) analyse
4) general linear model
5) univariate
6) put DV in ‘dependent variable’
7) put IV in ‘fixed factor’
8) see if p value is less than 0.05
how to calculate and compare the means of data sets?
analyse
compare means
means
put in IV and DV
what to say when interpreting significance with ANOVA?
if significant, say singificant but unspecified as unable to single out a group due to multiple groups
how do you plot ANOVA data?
analyse general linear model univariate plots put IV in horizontal axis click OK
how come an ANOVA test can be used instead of a t-test?
as F=t^2
how to perform a simple contrast?
e.g to compare means of the 2nd and 3rd groups out of 4
type into syntax: /contrast(IV column name) = special(0,-1,1,0)
analyse
general linear model
univariate
paste
add in line above into first row of code under UNIANOVA
green arrow
interpret results (if significant then can use means and plots to determine direction of singificance)
how to compare 2 groups to another group in an ANOVA test?
do the same format as a simple contrast
for the groups that you want to group together put as the same minus number e.g -1
for the group you want to compare them to, do a posiitve number which sums with the minus numbers to make 0 e.g 2
what happens if multiple contrasts are performed on information which isn’t independent?
contrasts will reveal greater differences between conditions than you actually have
e.g A vs C then A&B vs C as comparing A &C both times
what are Orthogonal contrast sets?
ones which are independent of each other
how to test if contrasts are independent?
if the slot-by-slot inner product calculation comes to 0
so times 1st slot value by 1st slot value in other contrast and do the same to all others then sum together (should = 0 if independent)
how to test if more than 2 contrasts are independent?
need to do 0 calulcation of slots for each pair individually
e.g for 1&2, 2&3, 1&3
and all have to be 0 for it to be orthogonal
what does a linear contrast tell us?
whether the best-fitting straight line slopes upor downwards
what does a quadratic contrast tell us?
whether the best-fitting quadratic curve has a bend in it
indicates whether the data plot departs from linearity
e.g 1,-1,-1,1
what does a cubic contrast tell us?
picks up evidence for an S-shaped bend (double inflection)
e.g 1,-3,-1,3
what do the ‘main effects’ tell you about each variable?
whether each variable has an indepdent effect by itself
what do the ‘interactions’ tell you about each variable?
how the effect of each variable is itself altered by the influence of others
how to conduct a two-way ANOVA?
have the 1st variable in the 1st column and label each 1 or 2
have the 2nd variable in the second column and label each either one or two
have the dependent variable in the 3rd column and write each set of data
analyse general linear model DV in DV IVs in fixed factor OK IV*IV means influence in one condition changes depending on what the other IV is
what does df1 and df2 tell us in a 2 way ANOVA?
df1 - how many levels there are in the factor being examined (-1) - numerator degrees of freedom
df2 - number of data points collected under each level of the factor under examination - denominator degrees of freedom
what to do when interpreting results?
do a standard ANOVA but press ‘plots’
put the IV with more levels in ‘horizontal’ and the other IV in ‘separate lines’
press continue
see graphical display of means to compare conditions
how to know from a graph when there is no interaction between conditions?
when the lines are parallel
what is significance (p value) effected by?
both sample size and effect size
what is effect size?
independent of sample size
the larger the effect size, the larger the difference between groups
how to work out effect size?
do standard ANOVA
options
estimate of effect size
continue
on results will say ‘partial eta squared’ and if more than 0.14 then large effect size shown, medium if more than 0.06 and small if more than 0.01
CHECK THIS COHEN’S CRITERIA
how to calculate a 3 way ANOVA?
have a column for each IV and label as level_level
put DV in last column and enter the data sets
put all IVs in fixed factor and DV in DV
how and when to conduct a Scheffe test?
when comparison devised after data viewed (unplanned/a posteriori)
1) put in data set the same way
2) run contrast in the usual way
3) do (number of conditions -1) x table value of F (degrees of freedom)
4) compare to F value and if F value is greater, then significant
