Stats+Research Flashcards
What values can the correlation coefficient range between?
Correlation coefficient can range from -1 to +1
What value is the correlation coefficient represented by?
r
What are the values of the categories of degree of correlation?
0-0.2 negligible
- 2-0.5 weak
- 5-0.8 moderate
- 8-1 strong
What does standard deviation measure?
Spread of a set of observations - variability of the obseervations
What is the formula for standard deviation?
square root of variance
What does a low standard deviation mean?
Data points tend to be very close to the mean
What does a high standard deviation mean?
Data points tend to be spread out over a large range of values
Is variance expressed in the same units as the data?
No
As the SD increases what happens to the SEM?
increases
As sample size increases what happens to the SEM?
Decreases
What’s the SEM?
A measurement of the precision of the sample mean as an estimate of the population mean
What is snowball sampling?
identifying participants with certain characteristics and getting them to identify more participants
What is Absolute risk reduction?
Absolute risk reduction is the difference in risk of a given event between two groups.
Formula = CER - EER
What is the confidence interval?
Confidence interval is the range of values within which one can be 95% certain that the true reference population value lies
What is the NNT?
Number needed to treat is the number needed to treat to get one good outcome or to prevent one additional outcome
Formula = 1/ARR
What is the relative risk?
Relative risk is the ratio of the risk of a given event in one group of subjects compared with another group
Relative risk = EER/CER
Relative risk reduction = ARR (CER-ERR)/CER
What is the relative risk reduction
Relative risk reduction is the proportion of the initial or baseline risk that was eliminated by a given treatment or by avoidance of exposure to a risk factor
What are case-control studies ideal for?
Rare diseases/outcomes
-partcularly useful for situations where there is a long time period between exposure and outcome
What type of bias are case-control studies at risk of?
Recall bias
What is face validity?
The extent to which a test measures what it’s supposed to measure
What is reliability?
How consistent a test is on repepeated measurements
What is predictive validity?
The extent to which a test is able to predict something it should theoretically be able to predict
What is inter rate reliability?
The level of agreement between assessmentsmade by two or more raters at the same time
What is contruct validity?
The extent to which a test measures a theoretical connstruct by a specific measuring device or procedure
What is Cronbach’s alpha?
It indicates the internal consistency of a test (the reliability)
-infinity to 1
What is criterion validity?
predictive and concurrent validity together.
What randomisation technique minimises contamination between intervention and control group?
Cluster
What is an independent t test used for?
Used to compare the means of two independent different sample populations - parametric test so assumes samples are normally distributed
What is a paired t test used for?
Used for matched studies
What is a crossover study?
longitudinal study in which subjects are allocated to sequences of treatment with purpose of studying differences between individual treatments - all subjects eventually receive all the same treatments
What is continuous scale + an example?
Data with no break in values and includes all possible data
e.g weight
What is an interval scale + an example?
Scale without an absolute zero, there are meaningful intervals between measurements
e.g. temp in farenheit
What is a nominal scale + an example?
Classification of observatiosn into unordered qualitative categories
e.g. bood group, ethnicity
What is a ratio scale + an example?
Interval scale with a true zero point
- thye are useful in regression
e. g. temp in Kelvin
What is an ordinal scale + an example?
Classiification of data into ranked categories
- data have order but the interval is meaningless
e. g. Hamilton Rating Scale for Depression (HAM-D) score
What is volunteer bias?
Those who agree to participate in the study tend to be healthier and more compliant than those who dont
What is selection bias?
Presence or absence of exposure influences allows allocation to particular study groups
What is ascertainmnent bias?
If data is more accurate or complete in one group than the other
What method in qualitative research can increases the validity of the study?
The use of multiple coders for validating emerging themes provides a mode of triangulation
How might using a single interviewer in qualitiative research be a weakness?
Using a single interviewer can affected the relevant experiences reported by the interviewed subjects
What type of sampling is used in most qualitative studies?
purposive sampling
What is moderator bias?
The moderator’s nonverbal and verbal attitude, expressions, body language, approach etc. may introduce bias. Similarly, the moderator’s demographics such as gender, age, social status, race and accent can also affect the interview content.
Why is purposive sampling used in qualitative studies?
maximises the chance of obtaining a range of views relevant for a qualitative study
The use of topic guides along with a time-limited interviewing is what type of interview?
semi-structured approach.
What’s the formula for the likelihood ratio of positive test?
Sensitivity / (1-Specificity)
What’s the formula for the likelihood ratio of negative test?
1 - sensitivity / specificity
What’s the formula for the false positivity rate?
1-specifity
What’s the formula for the NNT?
1/ARR
ARR = response in control group - response in exposure group
What’s a sequenced parallel RCT and what’s the purpose?
The sequenced parallel comparison design has been recently proposed by Fava et al. to reduce the impact of placebo response issue in drug trials. It is a design that identifies and removes early placebo responders.
TWO methods that help address potential bias in a meta-analysis
- considering papers not published in english
- using wider literature search
Choose TWO methods that help address heterogeneity in a meta-analysis
- random effect analysis
- Chi-square test
Which method is most useful to assess publication bias?
funnel plot
What is a measure of heterogeneity in meta analysis?
What % should heterogeneity be?
The I² statistic describes the percentage of variation across studies that is due to heterogeneity rather than chance. The rule of thumb is that you want the I² to be less than 50%. You will notice there are other statistics there like x² and z. Just focus on the I² as this is the most useful in interpreting a forest plot
How to work out 95% CI of NNT from 95% CI of ARR
Low limit = 1/higher limit of CI of ARR
Upper limit = 1/lower limit of CI of ARR
In a Gaussian distribution what is the expected value of kurtosis ?
3
What can be used to quanitify reliability?
Cronbach’s alpha
Cohens statistic
Kappa
Intraclass correlation coefficient
Which tests can show normality?
Lilliefors corrected test
Anderson Darling test
Kolgomogorov-Smirnov test
Shapiro-Wilk test
Efficacy vs effectiveness
Efficacy - research trials
Effectiveness - real life
Cohen’s effect size values
Cohen’s d is an appropriate effect size for the comparison between two means.
- 2 = small
- 5 = medium
- 8 = large
Do RCTs have highexternal validity?
No
This is about applicability in the real world
Perfect cut off point on ROC curve
0,1
X axis on ROC curve
false positive
Y axis on ROC curve
true positive
Purpose of using composite endpoint
Power of study is increased
What bias does allocation concealment help to reduce?
Allocation concealment helps to reduce selection bias
What bias does blinding help to reduce?
Blinding helps to reduce ascertainment/measurement bias.
What will be the percentage of type 1 error if the results of a trial are precisely meeting the conventional level of significance?
5%
When the same psychiatrist applies screening tool and carries out diagnostic assessment in a research study that evaluates the diagnostic tool, it can lead to what type of bias?
Information
Which analysis converts outcomes into monetary units?
cost benefit
Which analysis converts outcomes into Clinical
measurement scales?
cost effectiveness
What analysis converts outcomes into any natural
units (days saved,number of deaths prevented, or QALYs etc.)?
Cost consequences
What analysis converts outcomes into life
utility units e.g. DALY, QOL, etc.?
Cost utility
What analysis is used when both interventions produce same outcome and hence just the costs are compared?
Cost Minimisation
What is opportunity cost?
Opportunity cost is defined as the benefit that a resource would yield in its best alternative use.
The benefits of next best foregone alternatives when the available resources are utilised in one way
Which factor is not important in determining the quality of a meta-analysis?
Magnitude of outcome reported from individual studies
What is the non-parametric alternative to paired t-test?
Wilcoxon signed rank test
What is the the non-parametric alternative of the unpaired t test?
Mann-Whitney U test
What is Kruskal-Wallis test ?
Kruskal-Wallis test compares between the medians of two or more samples to determine if the samples have come from different populations (non parametric)
What is ANOVA?
ANOVA test compares between the means of two or more samples to determine if the samples have come from different populations (parametric)
What is an agreed method of assessing quality of conducting and reporting systematic reviews and metaanalyses?
QUORUM
What are measures of internal consistency of a test?
Cronbach’s α
Split half reliability
The ability of a test to correctly classify subjects as being with or without a disease is called?
Likelihood ratio
What’s the lifetime prevalence of depression in those with anorexia?
75%
The pre-test probability is equal to:
Prevalence
When are confounders important?
Confounding factors are only important if they differ between the arms of a study
What is an evidence-based, minimum set of recommendations for reporting RCTs
CONSORT
The guideline most suited to Non-randomised controlled trials
TREND
The guideline most suited to Meta-analyses
QUORUM
What is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses
PRISMA
What is a proposal for reporting a meta-analysis of observational studies in epidemiology
MOOSE
The guideline to improve the reporting of observational studies in the field of epidemiology
STROBE
an international, collaborative initiative of epidemiologists, methodologists, statisticians, researchers and journal editors involved in the conduct and dissemination of observational studies, with the common aim of STrengthening the Reporting of OBservational studies in Epidemiology
The consensus checklist recommended when reporting formal studies of healthcare quality improvemen
SQUIRE
The guideline most suited to Diagnostic studies
STARD
The guideline most suited to Microarray studies
MIAME
The guideline most suited to Qualitative studies
COREQ
The medical database where you can obtain abstracts published since 1890 is
Pychinfo
Medline has items from which year?
1949
Cohort studies are especially prone to this bias
Attrition bias
A structured approach whereby research ethics committees review proposals for RCTs
ASSERT
A Standard for the Scientific and Ethical Review of Trials
Number of deaths during a calendar year expressed in terms of midperiod population of the area
Crude mortality rate
Any study comparing incidence rates must have
a specified denominator
Exposure data are available for groups but not for their individual members in these studies
Ecological
What is ecological fallacy?
Ascribing the characteristics of a group to all individuals in that group
Disadvantages of carrying out a cohort study?
Time resources, manpower and monetary costs
Which study design is best suited to evaluate properties of a screening tool?
Cross sectional design
What is work up bias (verification bias)?
a systematic error in the assessment of the validity of a diagnostic test
What is family aggregation bias?
type of information bias
–> presence of a case in the family increases awareness of risk factors and exposure leading to more family members of cases reporting an exposure though controls have family members with less awareness.
When is Per-protocol analysis often carried out?
in biological explanatory RCTs and as the secondary analysis in effectiveness trials
Which trials have high external validity?
Pragmatc trials
What test can be used for adjusting for confounding factors when analysing the relationship between a dichotomous outcome and a dichotomous risk factor? (odds ratios)
The Mantel-Haenszel technique
Matching is most suitable for which studies?
Smaller studies i.e. small case control
*It is not needed for RCTs and extremely difficult for large cohort studies. It is not possible for ecological studies and cross-sectional surveys.
What is pygmalion effect?
The Pygmalion effect is also called the experimenter expectancy effect. Pygmalion is Bernard Shaw’s play where a professor bets and succeeds in teaching a poor flower girl to speak and act like a higher-class lady. Mere awareness of the experimenter’s expectations regarding treatment effect may result in a change in subject behavior in the direction of experimenter hypothesis.
In grounded theory method of qualitative research, the term theoretical saturation is used to describe?
Point at which no new categories can be developed from the available data
In which research method might the clear cut demarcation between researcher” and “researched” may not be so apparent?
Action research
Theoretical sampling is associated with
Developing a theory for further research
What can be used to demonstrate the validity of a qualitative study?
Reflexivity in reporting results
Triangulation
respondent checking
Deviant case analysis
What is reflexivity?
Recognition of the researcher’s preconceptions and their influence on interpretation and analysis
In qualitative studies, semi-structured interviews are often chosen over other forms of interviews. This is because
Good balance between resource limitations and amount of data collected
What is the most important factor to consider for reliability oof a qualititative study?
triangulation
member checking
What is a problem inherent in all qualitative studies?
Generalisability is a problem inherent in all qualitative studies.
One of the advantages of using a constant comparative approach for for data-analysis is?
The emerging categories can be refined in a flexible manner
Which research technique used during a qualitative study can inform the number of participants necessary for a given sample?
Saturation
Weighting refers to the level of significance attached to the outcome of individual studies considered in a meta-analysis. Which factors are considered for weighting?
Impact factor of the published journal
Sample size of studies
Methodological rigour of studies
Precision of results
In which situation is a random effects analysis is indicated in a meta-analysis?
Presence of statistical heterogeneity
What is the effect size measured in meta-analyses?
Difference in outcome between intervention and control groups in terms of standard deviation of the outcome in the population
Who coined the term meta-analysis?
Glass
Incremental cost-effectiveness ratio (ICER) formula
ICER = Cost of proposed intervention / effect of proposed intervention.
Which two parameters are required for constructing a cost-effectiveness analysis curve?
Willingness to pay and probability of cost effectiveness
In a cost-effectiveness plane which quadrant indicates best cost effectiveness of an intervention?
Southeast
In economic analysis what is the cost of pain and suffering from an intervention that cannot be quantified in monetary terms?
Intangible costs
Which method of quality improvement is based on 80/20 rule of management?
Pareto Chart
Reducing waste is a critical part of which of models of quality improvement?
Lean Thinking
What is age specific mortality rate used for?
Used to compare mortality between same age groups in different regions
Mediation model rules
- The predictor (i.e. treatment) must be significantly related to the outcome (i.e. response)
- The predictor must also be significantly related to the potential mediator
- The mediator must be significantly related to the outcome when the effect of the predictor on the outcome is controlled for
- The relationship between predictor and outcome must be reduced when controlling for the mediator (i.e. pathway c’ having effect lower than the pathway c).
If predictor remains significant when the mediator is controlled for
the mediation is considered partial
If predictor is non signifcant when the mediator is controlled for
mediation is considered complete
Most common level of prevention currently employed in psychiatric practice is
Tertiary
The incidence of schizophrenia is three times higher in migrants, but this drops to 1.8 times higher than natives when point prevalence rates are considered. This can be explained by
Shorter duration of illness in migrants
The phenomenon of relative deprivation is explained by
Jarmen index
According to the pathways of care model, illness behaviour is a strong determinant at which of the following filters between disease and secondary psychiatric care?
First stage - self recognition
What is it when two or more experimental interventions are not only compared separately but also in combination and against a control in an RCT?
factorial RCT
What is a a Guttmann Scalogram useful in measuring?
hierarchical constructs e.g. social attitudes, political views, developmental stages,
Which data is better suited for weighted kappa compared to Cohen’s Kappa for measuring agreement?
Ordinal
Which factors affect the external validity of a study?
- Representativeness of the sample
- Reactive effects of setting (is the research setting artificial)
- Effect of testing (if a pre-test was used in the study that will not be used in the real world this may affect outcomes)
- Multiple treatment inference (this refers to studies in which subject receive more than one treatment, the effects of multiple treatments may interact)
What is an example of a filtered resource?
The Database of Abstracts of Reviews of Effect (DARE)
The Kolmogorov-Smirnov and Jarque-Bera tests are used to test for
Normality
The chance that the patient does not have the condition if the diagnostic test is negative
NPV
How much the odds of the disease decrease when a test is negative
LR-ve
