Research methods + Stats

Question 1

Odds =
(FROM PROB)

Answer 1

probability / 1- probability

Question 2

Cluster sampling

Answer 2

divide population into clusters, often basis of geography

Question 3

Convenience sampling

Answer 3

who is easily accessible

Question 4

Quota sampling

Answer 4

fixed number of unit in each of a number of categories

Question 5

Snowball sampling

Answer 5

Asking people to pass on details to others

Question 6

Random sampling

Answer 6

Everyone in population has equal chance pf being selected

Question 7

stratified sampling

Answer 7

divide population into groups on basis of some suspected confounding characteristic

Question 8

systematic sampling

Answer 8

choosing every nth item from a list. beginning random point

Question 9

null hypothesis

alternate hypothesis

Answer 9

opposite you want to disprove e.g. if you want to prove A and B are different, say that no difference exists between A and B.

If null gets rejected = try to prove often actual research question

Question 10

Error
Type 1 error = alpha

type 2 error = beta

Answer 10

Incorrect rejection of null hypothesis = FALSE POSITIVE / chance finding / p-value (alpha)

Type 2 - FALSE NEGATIVE most likely due to small sample size or large variance

traditionally
a = 5%. B = 20%, power is 80%

Question 11

power =

Answer 11

ability of a study to detect a difference between 2 groups if such difference exists

Question 12

Effect size

Answer 12

Difference in outcomes between intervention and controls, divided by standard deviation
= IS a measure of the difference in point estimates

Question 13

Reliability tested by

Answer 13

Test-retest correlation (long enough to avoid practice effect, short enough to mean thing doesnt change enough eg patients depressive state) often 2-12 days in psychiatry

Question 14

Cronbachs alpha

Answer 14

measures internal consistency of a test = correlating each item with total score and averaging correlation coefficient

Question 15

face validity

construct validity

Answer 15

face = The extent to which a test appears to measure what it claims to measure, based on a superficial assessment.

The extent to which a test actually measures the theoretical construct it’s intended to measure. eg does becks depression inventory measure depression actually

Question 16

beyond chance agreement = kappa

Answer 16

kappa indicates level of agreement that cbe expected beyond chance

useful on agreement of categorical valuables eg presence/ lack of diagnosis

Question 17

Calculating kappa

Answer 17

Kappa = observed agreement beyond chance / max agreement beyond chance

OR
kappa = (observed agreement - agreement expected by chance / 100% - agreement expected by chance

Question 18

intention to treat analysis

Answer 18

Participants counted in groups to which they were initially randomised , regardless of whether they were compliant with allocation. more realistic real life + more generalisable

Question 19

N of 1 trial

Answer 19

doctor and patient do own personal test eg which medication best.
not generalisable
pharamcist blinds to both dr and patient

Question 20

how to control confounding?

Answer 20

Matching = make confounders equally distributed

Restriction = avoid including group with significant **confouner influence **

Randomisation = helps distribute confounders

Question 21

Types of bias = selection bias

Berkonson

neyman

response

unmasking

Verification bias

Answer 21

Berkson = admission bias (Birkinstocks on in hospital)

neyman = incidence prevalence bias - eg if unmasks something that correlated to incidence (eg sepsis), will be low prevelance if measuing this due to deaths (neyman ur a deadman)

response = peolpe who resond are different people by nature than those who dont

unmasking - risk factor unmasks rather than causes an event

verification bias (work up bias)

Question 22

Emic perspective

ETIC perspective

Answer 22

emic = MINE = internal observer

etic = outside observer

Question 23

Positive skew

negative skew

Answer 23

pulled +ve –> leg pulled to the R

pulled -ve–> leg pulled left

Question 24

variance

Answer 24

sum of squared differences of individual observations from mean / number of observations

Question 25

standard deviation =

Answer 25

square root of variance

Question 26

Normal distribution 1 SD 2 SD 3 SD

Answer 26

68% data lie in 1SD 95% lie within 2SD 99% lie within 3 SD, irrespective of the mean

Question 27

standard deviation of distribution of the means =

Answer 27

standard error

Question 28

Odds =

Answer 28

odds = probability / 1- probability

Question 29

Purposive sampling

Answer 29

participants are selected on purpose because the researcher already know that they have characteristics that are of interest for the stud

Question 30

Triangulation:

Answer 30

compares the results from either two or more different methods of data collection (for example, interviews and observation) or, more simply, two or more data sources

Question 31

Respondent validation (aka member checking):

Answer 31

techniques in which the investigator's account is compared with those of the research subjects to establish the level of correspondence between the two sets.

Question 32

Bracketing:

Answer 32

a methodological device of phenomenological inquiry that requires deliberate putting aside ones own belief about the phenomenon under investigation or what one already knows about the subject prior to and throughout the phenomenological investigation.

Question 33

Reflexivity:

Answer 33

sensitivity to the ways in which the researcher and the research process have shaped the collected data, including the role of prior assumptions and experience, which can influence even the most avowedly inductive inquiries.

Question 34

what graph detects publicaiton bias?

Answer 34

Funnel plot checks publication bias

Question 35

absolute risk reduction (ARR) =

Answer 35

subtracting the event rate (ERT) in the treatment group from the event rate in the control group (ERC) Event rate - Control event rate

Question 36

Survival curves

Answer 36

Kaplan-Meier curves.

Question 37

formulate research question PICO

Answer 37

Patient (P) What is the patient group of interest? Intervention (I) What is the intervention of interest? Comparison (C) What is the comparison? Outcome (O) What is the primary outcome?

Question 38

positive predictive value

Answer 38

Proportion of those scoring positive have a condition (i.e. The chance that a positive result will be correct). PPV = True Positive /(True Positive +False Positive)

Question 39

LR+ Likelihood ratio for a positive test result (LR+)

Answer 39

= sensitivity / (1 - specificity) aka = probability of a patient with a disease having a positive test divided by the probability of a patient without the disease having a positive test result

Question 40

likelihood ratio for a negative test result

Answer 40

The likelihood ratio for a negative test (LR-) is calculated by using the formula: LR- = (1-sensitivity) / specificity.

Question 41

fundamental criterion in Susser’s framework for causation

Answer 41

Time order Association Direction

Question 42

first quartile The IQR (interqualie range) is obtained

Answer 42

median of the lower half of values. The lower half is those values below the median. The IQR is obtained by subtracting Q3 from Q

Question 43

he GRADE approach is used for

Answer 43

Assessing the quality of evidence. The GRADE (Grading of Recommendations Assessment, Development and Evaluation

Question 44

Steps for identifying the appropriate test:

Answer 44

Step 1 → Define the research question (correlation or difference) Step 2 → Establish how many groups there are and if dependent or independent Step 3 → Identify the variables and data types Step 4 → Establish if parametric or non-parametric tests should be used Step 5 → Select the correct statistical test (as per chart below)

Question 45

RR defintion? What value can relative risk be

Answer 45

Relative Risk (RR) Relative risk compares the risk (or incidence) of a disease in the exposed group to the risk in the unexposed group. It can have values less than 1, but not less than 0 0. to infinity

Question 46

what is attributable risk? what value can attributable risk be

Answer 46

Attributable risk measures the absolute difference in disease risk between the exposed and unexposed groups --> that is, how much of the disease risk in the exposed group is due to the exposure. Value = it’s bounded between –1 and +1 (

Question 47

what can cross sectional surveys estimate

Answer 47

Prevenance only

Question 48

minimisation scheme -

Answer 48

next allocation depends on characteristics of those already allocated

Question 49

what is an effect modifier?

Answer 49

Impacts magnitude of cause effect analysis

Question 50

fixed effect model? random effects model?

Answer 50

Do we assume every study is estimating the same true effect? → Fixed-effect model Or do we assume each study estimates a different, but related, true effect? → Random-effects model

Question 51

discrete variable vs ratio variable

Answer 51

Ratio variables must be continuous AND have a true zero

Question 52

dependent vs independent

Answer 52

The independent variable is what you change; the dependent variable is what changes because of it / what you measure / changes based on independent

Question 53

transformation methods for skewed data

Answer 53

Log Transformation (log, ln) = good for R skewed- compress large values square root transformation - less extreme than log reciprocal transformation- good for highly skewed R data

Question 54

Coefficient of Variation

Answer 54

Sample standard deviation divided by sample mean of the data set

Question 55

Odd = probability =

Answer 55

odd = frequency one event occurring / frequency of another event occurring AKA favorable outcomes divided by unfavorable outcomes odds rolling a 2 on dice = 1: 5 prob = frequncy on evenet / frequnecy all events probability of rolling a 2 on dice = 1 / 6

Question 56

if accidentally get smaller sample size--> which is more likely t1 vs t2 error ?

Answer 56

t2 error (false negative more likely than false positive) eg easier to say no blue rabbits exist because youve never seen one, than it is to incorrectly say there is a blue rabbit that youve labelled wrong as its just got blue paint on ot

Question 57

NOMINAL

Answer 57

NAMES (NOT NUMBERS)