Unit 3 Flashcards
Human studies - ethics
Informed consent - potential subjects are given details before they participate.
They have the right to withdraw.
Information should be anonymous and confidential.
Participants should come to no harm.
Refinement
Techniques should be adjusted to minimise harm to the animals.
Reduction
animal studies
Reduce the number of animals used in a study without compromising validity or reliability.
Replacement
animal studies
Avoid using animals when an alternative cellular or model system exists.
Animal studies
Used due to the complex nature of cells and systems which can’t be modelled outside the body.
References
Included in any scientific writing.
Must contain enough information to allow the reader to find the original publication.
Essential for avoiding plagiarism.
Harvard system - bracketed citations with the name of the author and the year of publication. eg. (Lloyd and Morgan, 2019)
Vancouver system - shows citations using numbers either in brackets or as superscripts. These direct readers to the reference list, which has the references in numerical order.
Citation
An acknowledgment that appears in the text of a paper where the work of another scientist is being described or quoted.
Scientific ethics
Distinguishes between right and wrong. Scientists should behave with integrity and honesty and results should be unbiased.
Null hypothesis
States that there will be no statistically significant effect as a result of the experimental treatment.
Hypothesis
A potential explanation for an observed event or the predicted outcome of an experiment.
Scientific knowledge
The current best explanation which may be updated after evaluation of further evidence.
It is constantly evolving and changing as new data is generated, leading to a refinement of ideas.
Scientific cycle
A series of steps that biologists use to investigate phenomena in a testable, measurable and reproducible way in order to explain and predict future observations by refining a hypothesis.
Primary paper
Communicates new research findings.
Includes an explanatory title, abstract (summary of aims and findings), introduction (purpose and context), method (detailed enough to allow it to be repeated), results (raw and processed data and data analysis), conclusions (supported by the data and refer to aim and hypothesis) and references for all cited work.
Review articles
Used to summarise the current knowledge in a particular field and to discuss recent or novel findings.
Usually written by experts, and provide a summary of current knowledge in a particular field of study.
Peer review
Scientific papers are read and analysed by experts before publication.
Reviewers assess the scientific quality of the manuscript and recommend whether it is suitable for publication.
Reviewers may request alteration to experimental design, additional experiments or replicates, more suitable data analysis or inclusion of missing background information.
Pilot study
Used to develop and improve scientific protocols, ensuring that the experimental design is suitable for investigating the aim and hypothesis.
Avoids wasting time and money on a full scale investigation using flawed methods.
Used to practise difficult or unfamiliar techniques.
Used to determine a suitable range for the independent variable (eg. by dilution) and to determine the number of replicates required to ensure that the data is representative.
Validity
Variables must be controlled so that any measured effect is likely to be due to the independent variable.
Experiments must have sufficient control groups and appropriate randomisation of experimental subjects
Reliability
Consistent values are obtained in replicates (multiple repeats within the same experiment) and independent replicates (repeating the whole experiment and obtaining an independent data set).
Accuracy
Data (or means of data sets) are close to the true value.
Assessed by calibration against a known standard. eg. you could assess the accuracy of a balance using a reference weight of known mass.
Precision
Measured values are close to each other.
Assessed by taking multiple measurements and calculating a mean and standard deviation.
Independent variable
The variable that is deliberately changed by the investigator to determine if it has an effect on the outcome of the experiment.
Dependent variable
The variable that is measured to determine if changing the independent variable has an effect.
Selection bias
Individuals selected for a study are not representative of the population as a whole.
Sample size
The number of tests or individuals in treatment and control groups. Should be large enough to allow results to be tested statistically and should aim to be representative.
Observational studies
Use groups that already exist (eg. smokers/non-smokers), due to the ethics or logistics involved in setting up 2 randomised experimental groups. Unable to control the independent variable.
Good for detecting correlations (negative and positive), although correlation does not mean causation. (Just because there is a correlation, it does not mean that a change in one variable causes a change in the other.)
Confounding variable
A variable other than the independent variable, which may affect the dependent variable.
Randomised block design
Minimises the effect of uncontrollable confounding variables.
Subjects are grouped randomly, so that the different groups contain a broadly similar range of confounding variables.
The influence of confounding variables is likely to be the same across the different groups.
Negative control
Provides data for what happens in the absence of a particular treatment.
Used to check that false positives are not occurring.
Shows that the dependent variable remains constant in the absence of the independent variable.
Unexpected results indicate that a confounding variable is not being adequately controlled.
Positive control
Provides data to show a positive result when it occurs.
Used to check for false negatives. A positive control should show an effect on the dependent variable - if not, the design of the experiment is not valid.
Placebo effect
A placebo is a treatment which lacks the independent variable being investigated.
The placebo effect is a measurable change in the dependent variable as a result of patient expectations, and is not due to the presence of a drug.
In vivo studies
Studies involving whole living organisms. Provides data for effects in whole organisms and allows the study of complex interactions. (Multifactorial)
Expensive and time consuming, may have ethical and legislative concerns, difficult to control confounding variables, results may be hard to interpret, difficult to prove causation.
In vitro studies
Simpler and less expensive, easier to control confounding variables, interpretation of results is simpler, can demonstrate causation.
Difficult to extend results to whole organisms or different species, hard to model complex interactions.
Representative sample
Must take into account the natural variation within a population and how the population is distributed.
The greater the variation, the larger the sample required.
A representative sample should have the same mean and degree of variation about the mean as the population as a whole.
Random sampling
Allows each member of the population to have an equal chance of being selected.
Non-subjective and unbiased
Random number tables/generators should be used to select samples.
Used for sampling large areas with uniform populations.
Systematic sampling
Selects members of a population at regular intervals, eg. along a transect.
Stratified sampling
Divides the population into categories, which are then sampled in proportion.
Used when the population being sampled is not uniform. eg. patches of plants in a grassland which would be missed by random sampling.
Variation
The differences between individual members of a species.
Can be discrete or continuous
Qualitative data
Descriptive and difficult to measure directly.
Recorded using direct counts or observation. eg. colour
Difficult to analyse. Presented as a bar graph or pie chart.
Quantitative data
Measured directly and recorded as numbers. Easier to analyse.
Presented as a line or scatter graph, or as a histogram. eg. temperature
Ranked data
Data is put into order of magnitude (smallest to largest) and is presented as a bar graph. eg. plant abundance.
Normal distribution
The most common distribution in nature.
Seen as a symmetrical bell-shaped curve.
The central value is the mean (average), mode (most frequent) and median (middle value).
Skewed distribution
Can be negative (long tail to the left) or positive (long tail to the right).
The mean is not the most common value, and the mode is not central.
Error bars
Lines through a point on a graph which show the variation within the data.
Used to show standard deviation (difference from the mean) or probabilities.
The smaller the error bars, the less variable the data.
Missing error bars indicate that only one sample has been taken.
Confidence interval
A statistical estimate of the range of values within which a certain percentage of the population would be found.
A 95% confidence interval shows that the range of values would include 95% of the population.
