Experimental Design
What is the independent variable in an experiment?
(IV) is the variable which is manipulated/changed on purpose by the experimenter during an experiment. Used to directly affect DV.
What are systematic errors?
- Are predictable variations caused by inbuilt inaccuracy in an experiment. - result in readings/measurements that differ from true value by a consistent amount in one direction.
How should a hypothesis be written?
- It is expected that... - refer to IV and expected affect on DV. - mention how DV will be measured
What information should be included in a conclusion?
- It was found that as the IV..., the DV...as measured.. - This does/doesn't support the hypothesis due to.... being/not being significantly higher than... - refer to aim/this suggests that... - use specific evidence - If required refer to other evidence/data eg. outliers, errors
What to include when designing/writing an experimental method?
- Size & Similarity (establish a large sample size of 100ish & identify similarities eg. same age) - IV (describe how the IV differs in each group) - DV (explain exactly when & how to record the DV eg. after 3 weeks measure the height of the plant in cm) - controlled variables x2 (all variables, other than the IV and DV are controlled such as...) - control group - repeat - results conclusion (hypothesis & significantly).
Define negative & positive control groups and their functions.
- a negative control group is exposed to NO value of the IV, thus providing a known level of the DV without the IV... - a positive control group is exposed to a known/high level of the IV, thus providing a known level of DV with the IV... ...that experimental groups are compared to determine IVs affect on DV.
How is the validity of an experiment improved?
- by decreasing systematic errors - if reliability of experimental method is high
How can the accuracy of results be improved?
- by reducing impacts of systematic errors
What are some causes of systematic errors?
- faulty calibration of instruments - subjective reading by same person
How is the reliability of an experiment improved?
- minimise errors - large sample sizes & repetitions
How can systematic errors or their affects be reduced/eliminated?
- repeating and calculating average DOES NOT improve accuracy of results - reduced their impacts by using calibrated equipment and reducing observer bias
What are some causes of random errors?
- sampling error (when a sample doesn't represent a population) - rounding error - inability to read instrument because reading fluctuates during
What is a dependent variable?
A dependent variable (DV) changes in response to the the IV. An experimenter counts or measures changes in the DV due to the IV.
Why may it be impossible to assess accuracy of results?
Accuracy is based on the systematic errors of the experiment, and since it is impossible to discern all systematic errors it is impossible to assess accuracy.
Define accuracy in terms of experimental results.
Accuracy refers to how closely the experimental results agree with the accepted or true value.
How can human errors or their effects be reduced/eliminated?
By repeating an experiment with more care.
What is an continuous variable?
Continuous variables involve variables that are measurable and thus they can involve any value between a certain set of real numbers including fractions and decimals.
What are discrete variables?
Discrete variables involve values based on a count from a set of distinct whole values. They do not involve fractions between values.
Define validity in terms of experimental results.
Experimental validity refers to how well the experimental design matches the requirements of the investigation to produce results that address the aim. An experiment is valid if: - there were no uncontrolled variables that could affect DV - the results were measured objectively (not biased) - systematic errors were minimal
What are human errors?
Human errors result from actions that were not intended or desired. Unintended 'mistake' that tends to cause outliers.
For which kinds of experimental results is it appropriate to use a line graph?
If the IV and DV are continuous then it is appropriate to use a line graph, because it allows for predictions of values that fall in-between the data.
For which kinds of experimental results is it appropriate to use a bar graph?
If the IV is qualitative/discrete it is appropriate to use a bar graph because it operates data into categories that can be compared.
Give an example of a set of precise data.
If the true value was 9.6 a precise set of data might be; 10.1g, 10.0g, 10.2g, 10.1g. (however this is not accurate)
When can results be considered valid?
If they match the requirements of the investigation and address the aim.
Give an example of accurate results.
If true value was 9.6 accurate results might look like; 9.5, 9.7, 9.8, 9.6, 9.4. (These could also be considered precise.)
How should categorical data/qualitative variable be graphed?
In a bar graph.
Define reliability in terms of experimental results.
Is the ability of an experimental method to obtain the same results when repeated. An experiment is reliable if the method minimises errors and the results are consistent.
What is an error?
Is the difference between a measured value and the accepted true value. Sometimes an error can be recorded as e.g.±0.1ml to account for uncertainty of a measurement.
What are outliers and how do they affect experimental results?
Outliers are values that vary significantly from other results. Outliers must be further analysed rather than being automatically dismissed. Repeating readings can be useful to further examine outliers.
Define precision in terms of experimental results.
Precision refers to how closely two or more experimental results agree with the average.
What are quantitive variables and what are the two types?
Quantitive variables count or measure a quantity as a number. They can be discrete or continuous.
Reliability vs Validity
Reliability is the ability to obtain the same results if an experiment is repeated. Validity refers to whether your results are real results and whether they apply to all situations (only the IV influences the DV).
What can reliability also be referred to as?
Reliability/repeatability/reproducibility.
How can the precision of results be assessed?
Results are relatively precise, as each trial gives the same results as its second. There may be slight imprecision due to random errors mentioned above.
Precision vs accuracy
Systematic error = inaccuracy Random error = imprecision Accuracy is the ability to obtain the correct measurement. Precision is the ability to consistently obtain the same measurement.
How can the level of accuracy for a particular value be determined?
The level of accuracy for a particular value can be quantified as 'measurement error' by calculating the difference between the true value and and measure value.
How can the level of precision for a set of data be determined?
The level of precision of a set of data can be quantified as 'uncertainty' via calculating standard deviation from the mean. e.g.. 2.7g±0.1g is less precise than 2.7g±0.001g
What are some causes of human errors?
They include mistakes and miscalculations like; - misreading a scale - forgetting a value before calculating averages
What is the purpose of controlled variables?
This is to ensure that only the IV affects the DV and therefore that the results of an experiment are valid.
What is uncertainty?
a quantitative measure of precision - shown as standard deviation from mean eg. -+0.1g
What is a placebo?
a treatment (often sugar pill) designed to have no medical effect on a disease as it lacks the factor being tested. It acts as a control to ensure that any change in a patient's symptoms are only due to the medical drug being tested and not caused by the psychological effect of having a placebo.
How do systematic errors impact experimental results?
affect accuracy of results
How do random errors impact experimental results?
affect precision
What is qualitative variable?
are categorical and describe a quality or characteristic usually in words.
What are random errors?
are unpredictable variations caused by chance during an experiment. They result in readings/measurements that differ from the true value by an inconsistent amount in both directions.
How can the precision of results be improved?
by reducing the number of random errors
How do human errors impact experimental results?
can affect accuracy and precision
What can a qualitative variable also be described as?
categorical
What are controlled variables?
controlled variables are deliberately kept constant (everything other than IV & DV) during an experiment.
Give an example of referring to results in a method.
eg. If (plants grown with fresh H₂O have a) significantly (higher photosynthetic rate than those with salty), the results would support the hypothesis
Provide an example for a qualitative variable.
eg. gender, species type, small/medium/large
What is an example of a continuous variable?
eg. length 7.85m eg. pH 6.5
What is an example of a discrete variable?
eg. number of DNA fragments
What can a quantitative variable also be described as?
numerical
What is the relationship between accuracy and precision and reliability and validity?
precision= reliability accuracy = validity
How can random errors or their effects be reduced?
repeating measurements and calculating averages can improve accuracy of results and reduce impact of random errors.
On which axis of a graph is the IV plotted?
the horizontal (x) axis.
On which axis of a graph is the DV plotted?
the vertical (y) axis.