Statistics Flashcards
14866 – To monitor the incidence of cancer in patients as a result of differing degrees of severity of ulcerative colitis, the preferred method would be
A. prospective survey
B. retrospective survey
C. therapeutic trial
D. double blind trial
E. meta-analysis
A: prospective survey
Refer to Motulsky, page 183-191
8603 – ‘Values for plasma sodium concentration are normally distributed in a population of normal young adults’. From this statement it can be inferred that
A. values in 95% of this population will lie within +1 standard error of the whole-population mean
B. the values in about two-thirds of this population will lie within +2 standard deviations of the whole-population
mean
C. about 2.5% of this population will have values that lie above the population mean +2 standard deviations
D. the variance of values is described as the square-root of the standard deviation
E. none of the above inferences can be made
C
Motulsky, Ch 4
19725 – With respect to the correlation coefficient (r) for two variables, x and y
A. a correlation coefficient of -1 implies that there is no correlation between x and y
B. a negative correlation would be expressed numerically as r=0
C. when expressed graphically, r can be derived from the tangent of a curved line
D. a positive correlation between x and y does not imply that x causes y
E. the correlation between x and y always lies within the range -10 to +10
D
Motulsky Page: 155-159
21983 – The standard error of the mean
1: is an estimate of the standard deviation of the means of a large number of samples from a population
2: is larger than the standard deviation
3: may be used to study the significance of the difference between means of two samples
4: is not related to the number of observations in the sample
TFTF
Motulsky Page: 44
8615 – You wish to investigate under what circumstances a surgeon should advise inguinal lymph node dissection in patients with melanoma on the leg. You should begin your investigation by
A. requesting your hospital research committee for permission to undertake a clinical trial
B. reading the relevant literature reports and the protocols of previous studies of the question
C. designing a randomized, controlled clinical trial of dissection/no dissection
D. surveying the clinical records covering your experience and that of your colleagues/hospital retrospectively
E. designing a prospective survey to follow up all such patients with melanoma on the leg to relate their outcome to whether lymph nodes were dissected or not
B
Motulsky, Ch 20
7633 – From an epidemiological viewpoint, what factors are known to be associated with the development of prostate cancer?
1: country of origin
2: smoking
3: diet
4: family history
5: presence of lower urinary tract symptoms
TFFTF
There is a significant variation in the incidence of prostate cancer depending on the country of origin of the patient, with Asian countries typically having an extremely low incidence of prostate cancer and most developed western countries having a high incidence (1 True). The incidence, however, of microscopically detected cancers at autopsy are similar throughout all countries; and hence prostate cancer in a western country tends to progress to a clinical stage as opposed to Asian countries. Migration affects this incidence, as a patient moves from a country of low incidence to high incidence they adopt the incidence of their host country. Obviously, diet has been implicated - but not proven - as a possible explanation for this differing incidence depending on country of origin (3 False?). Food substances that have been implicated include dietary fat, isoflavin ingestion (commonly found in soy products), and lycopenes (commonly found in tomato base products). Other possible explanation include exposure to ultraviolet light as those countries with minimal exposure to ultraviolet light such as the Scandinavian countries or those races that absorb vitamin D poorly (eg African-Americans) have a significant higher incidence of prostate cancer compared to their counterparts.
There is no evidence that smoking is related to prostate cancer (2 False). Family history has been conclusively shown to be associated with the incidence of prostate cancer with a two-fold incidence of prostate cancer with one first degree relative affected and up to eight to nine times increased risk if two first degree male relatives are affected (4 True). The presence of symptoms is not associated with the development of prostate cancer (5 False) as most studies have identified equal number of men with and without symptoms who have been diagnosed with the disease, despite the fact that symptoms was the most common reason why men sought to have their prostates checked.
12542 – At present, the most useful method of epidemiological investigation of staphylococcal infections is to determine
A. colony and colour variation
B. M protein production
C. coagulase production
D. specific bacteriophage typing
E. specific antihaemolysin titre
D
The classical method of ‘typing’ staphylococci (both Staphylococcus aureus and coagulase negative species) in epidemiological investigations is phage typing (D true) although other molecular techniques are slowly being developed (eg plasmid analysis, chromosomal DNA analysis) as are antibiogram analyses for coagulase negative strains. Colony and colour variation is of little use as most colonies look the same (A false), while the presence or absence of the enzyme coagulase simply serves to delineate the heterogenous species S. aureus from the remaining 20 or so species (C false). Serological procedures have not been adopted for investigating outbreaks of staphylococcal sepsis (E false) except perhaps for determining the presence of a high level of antibody against cell wall teichoic acids as an indication of chronic staphylococcal infection (eg osteomyelitis). M proteins are antiphagocytic virulence factors in streptococci; they have been used in streptococcal typing schemes but are not present on staphylococci (B false).
7156 – A null hypothesis represents
A. the absence of an hypothesis
B. an unsatisfactory hypothesis
C. an hypothesis that no relationship exists
D. the evidence that no relationship exists
E. an hypothesis that a negative relationship exists
C
A null hypothesis is a fundamental statistical tool comparing two events; and represents the hypothesis that no relationship exists between the two events. The hypothesis is then tested and either confirmed or refuted.
