EDS Flashcards
(130 cards)
1
Q
what is nominal data
A
categorical data which cant be ordered e.g. gender, blood group
2
Q
what is ordinal data
A
categorical data which can be ordered
3
Q
what is discrete data
A
numerical data which only contains whole numbers
4
Q
what is continuous data
A
numerical data which has no limitations on the value e.g. height or weight
5
Q
what is prevalence
A
the number of people with that condition at that point in time
6
Q
what is incidence
A
the rate of new events in a time period
7
Q
how do you calculate variance
A
sum of squares of residuals/degrees of freedom
8
Q
how do you calculate degrees of freedom
A
sample size - number of constraints
9
Q
how do you calculate standard deviation
A
square root of variance
10
Q
what is internal validity
A
refers to whether the effects observed in a study are due to manipulation of the independent variable and not another factor
11
Q
what is external validity
A
looks at whether the results can be generalised across the whole population
12
Q
how do you calculate the standard deviate (z score)
A
x-mean/SD
13
Q
between what range of standard deviations from the mean do 68% of the population lie
A
-1 and +1
14
Q
between what range of standard deviations from the mean do 95% of the population lie
A
+/- 1.96
15
Q
between what range of standard deviations from the mean do 99.9% of the population lie
A
+/- 3.29
16
Q
what is pseudoreplication
A
where multiple measurements are taken from one individual
17
Q
what is the standard error of the mean
A
the measure of the variability of the distribution of means
18
Q
how do you calculate SEM
A
SD/square root n
19
Q
how does sample size affect SEM
A
greater sample size gives a smaller SEM
20
Q
how do you calculate 95% confidence internals
A
CI = mean +/- (1.96 x SEM)
21
Q
what is a one tailed hypothesis
A
where there is a specified direction
22
Q
what is a two tailed hypothesis
A
where there is no specified direction
23
Q
what is a type 1 error
A
a false postiive
24
Q
what is the error rate for type 1 errors
A
5%
25
how does p value affect the type 1 error rate
high p value increases type 1 error rate
26
what causes type 1 errors
publication bias, lower power, poor understanding of statistics, p-hacking
27
what are type 2 errors
false negatives (not rejecting the null hypothesis even though its false) - incorrectly concluding that there's no effect
28
what is the rate of type 2 errors
1-power
29
what is power
this is the probability of rejecting the null hypothesis when its false
30
what factors increase power
large sample size, decreased variability, larger p value
31
how do you calculate the t value
t = (sample mean-hypothesised mean)/SEM
32
what test would you use to test the mean of 1 sample to the null hypothesis
one sample t test
33
when would you use a one sample t test
to test the mean of 1 sample to the null hypothesis
34
what are the assumptions of a one sample t test
normal distribution of data
35
when are paired/unpaired t tests used
to test the differences between the means of 2 samples
36
what is the difference between a paired and unpaired t test
paired t test = where the 2 sets of data are collected from 1 individual
unpaired t test = where the 2 sets of data are collected from difference individuals
37
what test is used for normal distribution
Shapiro wilk test
38
what test is used for variance
levene's test
39
is levene's test is failed what can you use
welch's correction
40
when is one way ANOVA used
when you are comparing the means for 3 or more sets of unrelated data from a population
41
what test is used when you are comparing the means for 3 or more sets of unrelated data from a population
one way ANOVA
42
how do you calculate the F ratio
sum of squared of group means/sum of squares of residuals
43
why is doing a one way anova better than 3 unpaired t-tests
3 unpaired t-tests would increase the type 1 error rate
44
what are the assumptions for one way anova
random sampling, normal distribution, homogeneity of variance, independence of data
45
what test is used for looking at the independence of data
durbin-watson test
46
what is the non-parametric alternative for one way anova
kruskal-wallis
47
which error rate is increased by post hoc tests
type 2 error rate increases due to low p value
48
which post hoc test is used for comparing a control against other groups
dunnett
49
when would you use dunnett's post hoc test
to compare a control against other groups
50
when can you used planned contrasts
when there is a specific hypothesis to test
51
name 2 types of orthogonal contrasts
helmert and difference
52
when do you use two-way anova
when making comparisons between the means of 3 or more groups of data where 2 independent variables are considered
53
what test would you use when making comparisons between the means of 3 or more groups of data where 2 independent variables are considered
two-way anova
54
what is the non-parametric alternative for 2 way anova
there is no alternative
55
when would you do a simple effects analysis
when there is a significant interaction between independent variables
56
what type of multiple comparison test is applied to a simple effects analysis
sidak
57
when is a repeated measures anova
when comparing 3 or more group means but where the participants are in the same group (have multiple data taken)
58
what test is used for sphericity
mauchy's test
59
what extra assumption is required for repeated measures anova
sphericity
60
what do you use if sphericity cant be assumed
greenhouse-geisser correction
61
which post hoc test is used if mauchy's test is failed
bonferroni
62
if the Pearson correlation coefficient is +1 then what does this say about the direction of the slope
positive
63
what does the R2 value mean
it is the proportion of the total variability which can be explained by the model
64
what is the non-parametric alternative for Pearson correlation coefficient
spearman rank correlation
65
what is homoscedascitiy
this is the assumption that the variance around the regression line is the same for all values
66
what does the DFFit tell us
what the fit of the model would be after removing a data point
67
what does the DFBeta tell us
what the slope of the data would be after removing a data point
68
what value indicates a significant DFFit or DFBeta
>1 shows the data point has substantial influence on the model
69
which is more important in determining outliers - cook's distance or leverage
cook's distance
70
when is multiple linear regression used
when you are looking at the relationship between the dependant variable and multiple independent variables
71
what is multicollinearity
where difference variables in a regression model are correlated
72
how can you test for multicollinearity
VIF > 10 indicates multicollinearity
73
how big should the sample size roughly be compared to variable number for multiple linear regression
10 x the size
74
how do you calculate odds ratio
AD/BC
75
what test is used to compare categorical data
chi squared test
76
what is the non-parametric alternative for chi squared test
fishers exact test
77
how do you calculate a proportion
number of people with the characteristic / total number of people
78
how do you calculate CIs for proportions
proportion +/- (1.96 x SE)
79
how do you calculate SE for proportions
square root of p(1-p)/n
80
how do you calculate expected frequency
row total x column total / grand total
81
when would you use fishers exact test
when 20% of cells in the table contain expected counts less than 5
82
when is power analysis used
used to determine the right sample size
83
what is Simpsons paradox
shows that when combining 2 treatments has different results than when using the treatments individually
84
what test would you use to look for an association
chi squared test
85
what test would you use to look for a trend
linear by linear association test
86
what test would you use to look for a dependance
regression analysis
87
what does the adjusted R2 value use
the variability explained by the models when its applied to another population sample
88
what value should the Durbin Watson test be
close to 2
89
what value of cook's value shows a significant effect of a data point
>1
90
at what value should the leverage value be investigated
>0.01
91
what is the hierarchical entry of variables
when independent variables are added in order of theoretical importance
92
what is the non-parametric alternative for repeated measures anova
Friedman's test
93
what test makes no assumptions of the underlying data
bootstrapping
94
what is the likelihood of a type 1 error in an ANOVA with 3 groups?
15%
95
what is the dependant variable
the variable that you measure
96
what is the independent variable
the variable that you change
97
what is the b0 coefficient
y intercept
98
what is the b1 coefficient
the slope of the line
99
what values of DFFit and DFBeta indicate that a data point is influential
>1
100
what is the z score
the units for normal deviation
101
what proportion of the probability curve is found between 1 SEM above and below the mean
68%
102
if the SEMs of 2 groups overlap then what is true of the p value
it is greater than 0.05
103
how is the type 1 error rate increased
increasing the p value
104
how is the type 2 error rate increased
decreasing power, decreasing sample size, decreasing p value
105
what is a type 1 error
false positive
106
what is a type 2 error
false negative
107
how do you increase the power on a statistical test
increase the p value, increase sample size, increase effect size, decrease the standard deviation
108
how do you calculate F ratio
model variability/residual variability
109
does a low f ratio give a better or worse significance
worse
110
why are post hoc tests used
to control the type 1 error rate after making multiple comparisons using an ANOVA
111
when would you use Tukey's post hoc test
with equal sample sizes
112
which post hoc test would you use for unequal group sizes
Gabriels or Hochberg's GT2
113
which post hoc test would you use if equality of variance can't be assumed
Games-Howell
114
which post hoc test would you use if you need a high level of confidence
Bonferroni
115
when would you use Dunnett's post hoc test
when comparing groups to a control
116
when applying multiple post hoc tests what happens to the type 1 and type 2 error rates
type 1 error rate decreases but type 2 error rate increases
117
when are planned contrasts used
to test a specific hypothesis
118
what are helmert contrasts
an orgothonal contrast where each category is compared to the mean of the subsequent categories
119
what are difference contrasts
an orthogonal contrast were each category is compared to the mean of the previous categories
120
what are deviation contrasts
non orthogonal contrast where each category is compared to the overall effect
121
what are simple contrasts
non orthogonal contrast where each category is compared to either the first or last
122
what are repeated contrasts
where is category is compared to the previous category
123
when do you use Friedman's test
when the parametric assumptions for a 2 Way ANOVA is not met
124
when do you perform a simple effect analysis
when there is an interaction between 2 variables in a 2 Way ANOVA
125
which multiple comparison test is used in a simple effects analysis
sidak
126
what is a repeated measures ANOVA
where multiple readings are being taken from one individual
127
what test is done instead of Levene's test in a repeated measures anova
Mauchly's test for sphericity
128
what multiple comparisons test is used in repeated measure anova
bonferroni
129
if mauchly's test is failed then what is used
greenhouse-geisser correction
130
what multiple comparisons test is used in mixed model anova
bonferroni
