SJSU Final Exam Bio 155/55
questions and answers
2 x 2 Table✔✔2 variables:
-1 variable: treatment/ Intervention/ Factor
-2nd variable: outcome/ Event
There are frequencies analyzed in a 2 x 2 contingency table
Risk✔✔-Probability an event will occur
-Risk in one group: # of individuals with the outcome/ total # of individuals in that
group
Relative Risk:✔✔-Aka risk ratio
-Ratio of risks for 2 groups
-RR = (risk of intervention group/risk for reference group)
-RR = 1 means risk is the same
-RR > 1 patients with paclovid are more likely to be hospitalized than placebo
-RR<1 patients with paxlovid less likely to get hospitalized than placebo
Relative Risk reduction (RRR):✔✔-proportionate reduction in risk for the intervention
group relative to refrence group
-RRR = ( risk for reference group - risk for intervention -group / risk for refrence
group ) x 100
-Interpretation example: RRR= 88.8837%, paxlovid groups risk of HOD is reduced by
88.9% relative to the placebo group
-Reference = placebo group
-Intervention = paxlovid
How was alpha =0.05 determined?✔✔-R.A. Fisher wrote a book in 1925 he put
statistical tables used to determine statistical significance
-Needed a separate table for each df which was too many
-He chose alpha= 0.05
-Is arbitrary, feels reasonable, it works, socially accepted
-intended as a first look
-If alpha = 0.02 that means there is a 2% chance of making a type one error.
Decreasing type one error we increase the chance of a type two error ( missing
significant findings )
Using Alpha and Null Hypothesis significance Testing( NHST)✔✔-Caused by the p-
value there is dichotomy( 2 outcomes ) which are significant p< alpha and non
significant p> alpha
-Significant also called positive findings ( most published)
-non significant also called negative findings (put in the file drawer)
Importance of negative findings✔✔-In scientific knowledge they are important
-Scientists may repeat studies because they may not know about the unpublished
studies (negative findings usually dont get published)
, -They are scientifically significant
Publication Bias:✔✔positive findings are more likely to be published (95%)
Is Alpha = 0.05 meaningful?✔✔-Depends on how you look at it
-It seems like a rational number and it works, but we can adjust it depending on what
we are looking for.
-Medically, we may be wanting a p value of 0.01 in order to have most accurate
results
-For ecology, we may change the p value to 0.1
-For many statisticians it is valuable, and they use it as a determining factor of
wether or not they want to publish their data.
-The problem with this is that a lot of the time scientist repeat experiments due to
unpublished negative findings
-Another outlook is that p value was meant to be a first look of findings
Since many scientists see p value as meaningful, they stop their studies when they
get negative results even if there is a correlation. R.A Fishcher meant for it to be
more of a first look at your findings
-Many statisticians in the community have began no longer using a p value for the
reasons listed above, and therefore the alpha can be meaningful, it varies from
statistician to statistician, along with the field of work.
P-hacking✔✔-When a statistician is trying to get positive results they add or take
away data to get the positive results they want.
-This makes the data no longer accurate/ should no longer be published ( not
meaningful)
-Therefore focusing on p values makes us ignore trends that are still important
-The trend could also possibly be due to chance
-Managing data can be p-hacking depending on the study
biased
Why is focusing on p-values problematic?✔✔-Scientific significance/significance
meaning ( + results wanting to be published, there is still knowledge to be learned
even if it is a negative finding) losing sight of scientific knowledge and significance
-Contribution to science
-What knowledge we gain from it
-Can look over some relationships and effects
-Arbitrary
-intended for first look, it wasn't meant for it to be determining factor
Solutions✔✔-Look at the scientific significance
-Look at how it contributes to the field
-Meaning of the results
-Look at trends
-What do the trends tell you
-Don't focus on p-values
-Multiple experimental methods
-Do you get the same answer?
-Look at the quality of the experimental design and data collection
questions and answers
2 x 2 Table✔✔2 variables:
-1 variable: treatment/ Intervention/ Factor
-2nd variable: outcome/ Event
There are frequencies analyzed in a 2 x 2 contingency table
Risk✔✔-Probability an event will occur
-Risk in one group: # of individuals with the outcome/ total # of individuals in that
group
Relative Risk:✔✔-Aka risk ratio
-Ratio of risks for 2 groups
-RR = (risk of intervention group/risk for reference group)
-RR = 1 means risk is the same
-RR > 1 patients with paclovid are more likely to be hospitalized than placebo
-RR<1 patients with paxlovid less likely to get hospitalized than placebo
Relative Risk reduction (RRR):✔✔-proportionate reduction in risk for the intervention
group relative to refrence group
-RRR = ( risk for reference group - risk for intervention -group / risk for refrence
group ) x 100
-Interpretation example: RRR= 88.8837%, paxlovid groups risk of HOD is reduced by
88.9% relative to the placebo group
-Reference = placebo group
-Intervention = paxlovid
How was alpha =0.05 determined?✔✔-R.A. Fisher wrote a book in 1925 he put
statistical tables used to determine statistical significance
-Needed a separate table for each df which was too many
-He chose alpha= 0.05
-Is arbitrary, feels reasonable, it works, socially accepted
-intended as a first look
-If alpha = 0.02 that means there is a 2% chance of making a type one error.
Decreasing type one error we increase the chance of a type two error ( missing
significant findings )
Using Alpha and Null Hypothesis significance Testing( NHST)✔✔-Caused by the p-
value there is dichotomy( 2 outcomes ) which are significant p< alpha and non
significant p> alpha
-Significant also called positive findings ( most published)
-non significant also called negative findings (put in the file drawer)
Importance of negative findings✔✔-In scientific knowledge they are important
-Scientists may repeat studies because they may not know about the unpublished
studies (negative findings usually dont get published)
, -They are scientifically significant
Publication Bias:✔✔positive findings are more likely to be published (95%)
Is Alpha = 0.05 meaningful?✔✔-Depends on how you look at it
-It seems like a rational number and it works, but we can adjust it depending on what
we are looking for.
-Medically, we may be wanting a p value of 0.01 in order to have most accurate
results
-For ecology, we may change the p value to 0.1
-For many statisticians it is valuable, and they use it as a determining factor of
wether or not they want to publish their data.
-The problem with this is that a lot of the time scientist repeat experiments due to
unpublished negative findings
-Another outlook is that p value was meant to be a first look of findings
Since many scientists see p value as meaningful, they stop their studies when they
get negative results even if there is a correlation. R.A Fishcher meant for it to be
more of a first look at your findings
-Many statisticians in the community have began no longer using a p value for the
reasons listed above, and therefore the alpha can be meaningful, it varies from
statistician to statistician, along with the field of work.
P-hacking✔✔-When a statistician is trying to get positive results they add or take
away data to get the positive results they want.
-This makes the data no longer accurate/ should no longer be published ( not
meaningful)
-Therefore focusing on p values makes us ignore trends that are still important
-The trend could also possibly be due to chance
-Managing data can be p-hacking depending on the study
biased
Why is focusing on p-values problematic?✔✔-Scientific significance/significance
meaning ( + results wanting to be published, there is still knowledge to be learned
even if it is a negative finding) losing sight of scientific knowledge and significance
-Contribution to science
-What knowledge we gain from it
-Can look over some relationships and effects
-Arbitrary
-intended for first look, it wasn't meant for it to be determining factor
Solutions✔✔-Look at the scientific significance
-Look at how it contributes to the field
-Meaning of the results
-Look at trends
-What do the trends tell you
-Don't focus on p-values
-Multiple experimental methods
-Do you get the same answer?
-Look at the quality of the experimental design and data collection
