Scientists observe something they don't understand Come up with a hypothesis (possible explanation) Test Hypothesis: Make a prediction Test by gathering evidence If evidence, backs up prediction, step closer to finding out if prediction is true Share findings in peer reviewed journals or at conferences Peer Review - other scientists test results and explanations to make sure they're reliable Helps detect false claims, but not always correct - mistake could just not be obviously wrong If an experiment doesn't fit with the hypothesis (and others can replicate results), then hypothesis may have to be modified or scrapped Accepted hypothesis are known as theories Currently accepted theories have survived 'trial by evidence' - tested many times and survived Theories are not indisputable fact New evidence that can't be explained: Hypothesising and testing starts all over again A representational model is a simplified description or picture of what's going on Can be used to explain observations and make predictions I.e : Lock and key model, which shows how enzymes work, and can be used to explain how enzymes catalyse reactions Computational models use computers to make simulations of complex processes Used when there's lots of different variables Can easily change design to take new data into account All models have limitations on what they can explain or predict i.e. Climate change models can't take into all of the biological and chemical processes that influence climate and it's difficult to include regional variations in climate
Important to communicate discoveries to the public Some show that people should change habits Some provide ideas that could be developed into new technology "Gene technologies are used in genetic engineering to produce genetically modified crops. Information about these crops needs to be communicated to farmers who might benefit from growing them and to the public who can make informed decisions on what the buy and eat" CGP Biology Revision Guide Scientific evidence can be presented in a biased way Reports about scientific discoveries in the media aren't peer reviewed Data may be presented in a way that's over simplified or inaccurate, opening it up to misinterpretation People making a point can present data in a biased way i.e. A scientist could overemphasise a relationship in the data An article could only give evidence that supports an idea, and none against it Developments could raise issues Economic issues - Society can't afford to do thing recommended by scientists without cutting back elsewhere Social issues - Decisions based on scientific evidence that affect people's lifestyles Personal issues - Some affect individuals - i.e NIMBYs Environmental issues - Human activity could effect the natural environment Science can't answer every question - including ethical issues Some can't be answered because there's no data to be collected or support a theory 'Should we do this' questions can't be answered
Nothing is completely risk free Hazard - Could potentially cause harm Risk - Chance the hazard could cause harm New technology can bring new risks, and they need to be considered alongside the benefits Estimate the size of the risk by number of times it occurs in a large sample To make decisions about activities with hazards, you have to take the risk and the seriousness of the consequences into account. High Risk - Involves a likely hazard with serious consequences Decisions about risk People are more likely to accept a higher probability of a minor and short lived consequence More willing to accept risk if they choose to do something (i.e skydiving) Less willing to having risk imposed onto them (nuclear power station built next door) Perception of risk isn't always accurate- people tend to see familiar activities as low risk and unfamiliar as high risk. People overestimate risk of long term or invisible effects Hazardous Investigations Common Hazards: Microorganisms - bacteria can make you ill Chemicals - Acids can burn skin and alcohols are highly flammable Fire - Bunsen Burners and fire Electricity - Faulty equipment Part of planning to make sure it's safe Reducing risk: Safety goggles Gloves Heat proof mats for Bunsen Burners Etc.
Investigations provide evidence to support or disprove a hypothesis Observe things and come up with hypothesis To do an investigation, you need to predict what you think will happen based on the hypothesis Investigations can be used to find patterns or relationships between the variables Data needs to be Repeatable, Reproducible, and Valid Repeatable - Does experiment using same method and equipment and gets same results Reproducible - Someone else can do the experiment, or a different method and get similar results Confidence in repeatable and reproducible evidence as they are reliable Valid results - repeatable, reproducible, and answers original question Fair Tests In an experiment, usually change one variable and measure how it affects other variables Everything else has to stay the same (control) Independent variable - You change Dependent variable - You Measure Control variable - You stay the same Can't always control all the variables, so have to do a control experiment Control experiment - kept under the same conditions, but doesn't have anything done to it See what happens when nothing is changed at all Sample size Data based on small samples isn't as accurate as data based on larger samples A sample should represent an entire population Small sample sizes make it difficult to find anomalies Larger the sample size, the better, but scientists have to be realistic when choosing it More realistic to study a 1000ish people of mixed gender, age and, race, than an entire population
Should Repeatable, Reproducible, Accurate and Precise Check repeatability and that the results are similar Repeat at least 3 times Check being reproducible by taking a second set of readings from other instruments or observers Accurate results are close to the true answer, and accuracy depends on the method and equipment used Precise data is data that's all close to the mean of the repeated results Equipment has to be right Measuring equipment has to me sensitive enough to measure change Resolution - Smallest change a measuring instrument could detect Equipment needs to be calibrated, by measuring a known value and seeing if there's a difference. Then use this to correct the inaccuracy Errors and Anomalies Results will always vary due to random error Random Error - Unpredictable differences caused by things like human errors in measuring Reduce effect of random errors by calculating the mean Systematic error - When a measurement is wrong by the same amount each time Repeating experiments won't correct a systematic error Zero error - a systematic error that is due to equipment not being zeroed properly You can compensate systematic errors if you know about them Anomalous result - A result that doesn't fit in. You should investigate and try to work out what happened, and you can ignore it if you can work out what happened
Can only conclude what the data shows, and nothing else State the pattern between the dependant and independent variables Only conclude what the data shows and not go any further Justify your conclusion using the results Refer back to the hypothesis in the final conclusion (supports or goes against) Correlation doesn't equal cause If there's a relationship between the variables, but one might not me causing a change in the other Reasons for correlation: Chance - Correlation could purely be due to chance Linked by a 3rd variable- often a different variable Cause - A change in one variable causes a change in another
SI units are used universally Same units in order to compare data SI units - Standard units Scaling prefixed can be used for large and small quantities SI Base unit + Prefix: To go from a bigger to smaller unit, you multiply by the conversion factor To go from a smaller to bigger unit, you divide by the conversion factor Useful conversion factors: Kg and g = 1000 Mm and um = 1000 Min and s = 60 m^3 and dm^3 = 1000 Dm^3 and cm^3 = 1000 Check values in equations have the right units Equations and formulas show the relationship between variables To rearrange an equation, you must do the same to each side For an equation, you have to know all but one of the variables Always check that you have the right units, and convert them if not
Uncertainty = the amount of potential error Often get a slightly different figure with every repeat due to random error Error also due to limitations and resolution of measuring equipment Mean set of results will also have some uncertainty Larger the range, the less accurate precise the results and more uncertainty Uncertainty is shown with the ± symbol Uncertainty = Range/2 Example: 20.1 19.8 20.0 Mean = 20.0 Range = 0.300 Uncertainty = Range/2 0.300/2=0.150 Mean = 20.0 ±0.150 Measuring a greater amount (i.e. Over a longer period of time in comparison to a shorter amount of time) will reduce the percentage uncertainty of the results Evaluations A critical analysis of the entire investigation Comment on the method - was it valid? Were all variables controlled? Comment on quality of results - Enough evidence for a valid conclusion? Results are repeatable? Reproducible? Accurate? Precise? Any anomalous results? Say if none, try to explain the ones there - errors in measurement?, Other variables that could have impacted the results? Comment on the level of uncertainty Analysis shows how confident you are with the conclusion Suggest any possible changes to method that could improve the quality of results, so there's more confidence in the results - changing the way a variable is controlled, more measurements at narrower intervals Make more predictions based on the conclusion, then further experiments that could be used to test them.