Flashcards in Mock Questions Deck (52)
What does the mean describe?
How data is concentrated
What does standard deviation describe?
How data is separated
How is the mean calculated?
Sum of all the samples / no. of samples
How is SD calculated?
Square root of:
Sum of (x - mean) squared / n-1
What shape is ND graph?
Where is the mean on a ND graph?
How much data is included in the ranges of mean +/- 1, 2 and 3 SD's?
+/- 1 SD = 68%
+/- 2 SD = 95%
+/- 3 SD = 99%
What forms can be inputted to SPSS?
What are the 3 types of data?
What is nominal data? Give an example
Categories without rank
(e.g. gender, colour weekday, place)
What is ordinal data?
Categories with rank
(e.g. pain, satisfaction, comfort)
What data characteristics can be shown using a histogram?
Frequency and distribution
In SPSS, what can users do on variable view?
In SPSS, what can users do with data view?
Input and edit data
What is the median?
Middle number when the data is ordered
What does the crosstab function do?
Arranges 2 variables into a table
Can calculate hypothesis and chi square directly
What is an error bar?
Mean with: SD, SEM or 95% CI
When should an error bar be used compared to a box plot?
Error bar - when data is ND
Box plot - when data is not ND
What is a box plot?
The median (50%)
Quartiles (25% and 75%)
Extreme values (min and max)
What does a scatter/dot graph show?
Data trends/patterns between 2 variables
How is SEM calculated?
SD / square root of n
What does 95% CI mean?
Confidence that the range will include 95% of local mean
When is the median more useful than the mean?
When the data is not normal distribution
How is variance calculated?
The square of SD
What is a pie/bar chart useful for showing?
Proportions and percentages in data
In SPSS how what can data be exported as?
If p is LESS THAN 0.05 what does this mean?
If p is GREATER THAN 0.05 what does this mean?
NO SIGNIFICANT DIFFERENCE
What do * and ** mean?
* p <0.05
** p 1,0.01
When would you accept the null hypothesis?
If p < 0.05
What are the key indices used in test of hypothesis?
Significant level (p)
What is the difference between dependent and independent data?
Independent - measurements done once for each individual
Dependent - samples taken two/multiple times for each individual
What are the 3 indications to use an independent t test?
Small sample size (big still ok)
What test should be used for dependent data?
Paired t test
What is an alternative if data is not suitable for t test?
What result from ANOVA would indicate the need for a post-hoc test?
p < 0.05 to check which pair ARE SIGNIFICANTLY DIFFERENT
What assumption is made when using a Chi square?
All groups have the same chance (equal no. from each group)
What test methods can be used for data that is not numerical or normal distribution?
Wilcoxon signed-ranked test
Mann-Whitney U test
When is a Wilcoxon signed-ranked test used?
For non-numerical or normal distribution data that is related/paired/dependent
When is Mann-Whitney U test used?
No requirements on data distribution
Usually ordinal data
Also scale if the data is not normal distribution
What is the correlation coefficient used to quantify?
How variables are linearly correlated
What results would be expected for correlation and what do they mean?
Value between -1 and 1
0 is the worst value for correlation
Closer the value to 1 or -1 the more correlated
When should linear regression be used?
If 2 variables have a significant linear correlation
What is the purpose of linear regression?
To construct an equation to describe the relationship between the 2 variables
What form does a linear regression equation follow?
y - b1 + b2x
(y = mx + c)
What does b1 mean in a linear regression equation?
What does b2 indicate in a linear regression equation?
The gradient of the slope
What is the purpose of survival analysis?
To analyse how a ratio of occured to inon-occured cases changes with time
What is needed for survival analysis to work?
Large sample size
What is the purpose of meta analysis?
To compare different opinions by analysing multi source data to find which is the favourite by most studies
What is needed for meta analysis to work?
Data needs to be in similar format