Brainscape's Knowledge GenomeTM


Top Flashcards (Certified)
All Flashcards

Statistics > Stats 1 > Flashcards

Stats 1 Flashcards

(58 cards)

Start Studying
1
Q

What does PICO stand for?

A

People, patients or pop
Intervention
Control or comparison
Outcome

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is categorical data? 2 types….

A

Nominal (no units) e.g. Male/female

Ordinal (order to the group-e.g. Tumour stage)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Two types of numerical data

A

Discrete (counted units-no of children)

Continuous (e.g. Height in cm)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

35% of people smoke- how much in pie chart?

A

0.35*360= xdegrees

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Difference between histogram and bar chart?

A

Histo has no spaces between boxes

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What are the two main elements used to describe data? (Descriptive statistics)

A

1) location- where on average do the data fall?

2) spread- how much variation is there in the data?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Give 3 examples on how to describe the location of data?

A

Mean- sum all values /no. Of values
Median- central value when all observations have been ordered
Mode- most commonly occurring value in the data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

How do you calculate the median of a set of 11 observations in ascending order?

A

1/2 (11+1)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Benefit of median over mean?

A

More resistant to outliers

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Name 2 ways to describe the spread of data?

A

1) standard deviation- a measure of how far each observation deviates from the MEAN
2) interquartile range = quartiles separate data into four

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What does IQR describe and how do u calculate it?

A

IQR = where the middle 50% of the data lies

Lower quartile (QL) = 1/4 (n+1)
Upper quartile (Qu) =3/4 (n+1)
Interquartile range = Qu-QL
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Normally distributed data should fall within what SD from the mean?

A

68% within 1 SD
95% +/- 2 SD
99.7% +/- 3 SD

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

How to calculate sample variance?

A

Subtract mean from each number and square the result (the squared difference). Then work out the average of those squared differences

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How do u analyse a large pop of to big to sample?

A

Sample many smaller pop and plot the mean results. Then used the standard error (SE) to decide how well sample mean reflects the unknown population mean.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

How do you calculate standard error?

A

Standard deviation / square root of number of subjects

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What does the 95% confidence interval represent?

A

If repeated samples were taken and the 95% CI for each sample was calculated then 95% of the CIs would contain the population mean.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

How do you calculate the CI ?

A

The mean +/- 1.96 X SE

1.96 as it represents the number of SD from the mean that encompasses 95% of pop

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

How do you assess CI’s for two means that overlap?

A

Easier to interpret the dif between the two means and the CI for that difference.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is a cross sectional study?

A

Observational study that analyses data collected from a population at a specific point in time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What type of study can you use risk statistics on?

A

Cross sectional ONLY

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

How do you calculate risk ratio?

A

Risk of diseases in exposed / risk of disease in unexposed

E.g (15/78)/ (5/122)= 0.192/0.041 = 4.68 = smokers 4.68 more likely to have disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

How do you calculate risk difference?

A

Risk of disease in exposed- risk of disease in unexposed

E.g. 19-4 =15%
Smokers have 15% more disease than on smokers

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What do risk ratios of 1, less than 1 and more than 1 mean?

A 
1= no change in risk
<1 = decrease in risk
>1 = in crease in risk

RR= 3 means three times more likely

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

When dealing with risk ratios how do you compute the percent relative effect( the percent change in the exposed group)?

Study These Flashcards
A

When rr >1 %increase = (rr-1) x 100
When rr<1 %decrease = (1-rr) X 100

You regard the unexposed group as having 100% of the risk and express the exposed group relative to that.

25
How do you calculate odds ratio?
Odds of disease in exposed/odds of disease in unexposed
26
How do you calculate odds?
No. Of people with disease/ no of people without disease
27
What is the rare disease assumption?
For rare diseases (<10%) the odds ratio and risk ratio will be similar ( for common diseases the odds ratio overestimates the risk ratio)
28
What type of ratio do you calculate for case control studies?
Odds ratio
29
Which ratio calculation is best when correcting for other factors?
Odds ratio
30
How to describe the comparison of continuous outcome variables?
1) draw histogram-to determine if outcome variable is evenly distributed- 2) if normal, compute the mean and SD of the two groups. If not normal, calculate the 2 medians and their 25th and 75th percentile / IQR 3) compute a measure of effect - if normal this = mean in exposed group- mean in unexposed. If not normal= median in exposed group-median in unexposed.
31
How do you describe the comparison of binary outcome variables?
1) compute descriptive statistics- summarise data e.g. Percentages or risk 2) compute a measure of effect - e.g. Risk differences or ratios and 95% CI or odds ratio
32
How do you describe the descriptive comparison of a categorical outcome variable? E.g. Mild moderate or severe
Compare outcome between two exposure groups through a cross tabulation table (e.g. Treatment, no treatment, total vs mild moderate or severe)
33
What type of ratio is used in epidemiological studies?
Risk ratio
34
When advising on the increased risk of a disease associated with an exposure should you use risk difference or risk ratio?
Risk difference E.g. Risk of cancer from the pill increased 5 fold- but increase only from 0.001 to 0.005- dif of 0.004% sounds better
35
What is PAR?
Population attributable risk- the additional risk of a disease in the whole population due to exposure
36
What is PARF?
Population attributable risk fraction- % of disease in the whole population due to exposure
37
What do risk ratios depend on?
The number of cases or controls in the study which is fixed by the researcher
38
What do odds ratios depend on?
The proportion of cases or controls with the exposure.
39
What type of ratio is only appropriate in studies where participants are recruited independently of disease?
Risk ratio
40
What do you look at to assess the role of chance in your results?
1) CI around measure of effect 2) p-value 3) sample size
41
What is reverse causation?
Did the exposure cause the disease or did the disease cause the exposure
42
In what type of studies is reverse causation worst?
Cross sectional Case reports/series Case-control Most likely a problem when timing of onset of disease in relation to exposure is difficult to establish
43
In what type of studies is reverse causation not a problem?
Clinical trials | Cohort/longitudinal studies
44
What is a confounder?
An exposure which is independently associated with the outcome and is associated with the exposure of interest
45
What should you look at to deal with confounding?
Design- randomisation, restriction, matching | Analysis- stratified, multivariate regression
46
Describe randomisation to avoid confounding
People randomised to exposure or not- to muddy relationship between exposure and confounding
47
Describe restriction to prevent confounding
Study one level of confounding variable only - e.g. Non-smoking minor and bus drivers who get lung cancer May make results less generalised
48
Describe matching to avoid confounding ?
In case control studies match case to control based on the potential confounder. E.g- same age and gender
49
Describe stratification to avoid confounding
Analyse data for each level (strata) of the confounder
50
Describe how multivariate analysis can be used to avoid confounding?
1) Can use odds ratio to look at many confounders in a study | 2) look at unadjusted odds and adjusted odds
51
What is bias?
A systematic error in an estimate arising from problems in the study design/execution.
52
How does selection bias occur?
If criteria for inclusion of cases or controls is related to the exposure of interest
53
Explain loss of follow up bias.
This is a problem in cohort studies- is drop out related to the exposure- DEATH
54
Describe information bias
Occurs when study subjects are misclassified either according to their disease status, exposure status or both
55
Explain recall bias
Healthy baby mothers vs mothers of babies with congenital abnormalities more likely not to remember any drugs taken during pregnancy
56
What are the two MAIN types of bias?
Selection bias- arises from differences between those included in a study and those not Information bias- arises from error in individual measurements of exposure or disease
57
What are the two types of information bias?
Non-differential = when errors in classification of exposure status affect the diseased and non-diseased equally. Eg. Measurement of height will always have error. Differential= when errors in classification of disease status are dependent upon exposure e.g. Recall bias
58
How do you avoid bias?
Good study design etc. Can consider blind interviewers

Statistics (3 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions