Mater Academy Lakes High School



SCIENCE NOTES“In Science, what is transiently known as truth is dependent on a collectively reviewed burden of proof that is driven by the power of imagination and the tools our creativity designs to further the understanding of the reality in which they are based upon."Introduction to ScienceThe Basics of ScienceBody of knowledge (accumulated of over thousands of years)Process: Study method or organized process of using evidence to learn about the natural worldField of study thatExplores the natural world (empirical)Collects and organize information (methodic)Looks for patterns/connections (logic / deductive reasoning)Proposes explanation that can be evaluated by evidence (imagination / inductive reasoning)Goal: Investigation/Observation to understand/explain the natural world in order to make useful predictionAlways changing (dynamic) through inquiry (raising questions) and peer-review (collaborative evaluation)The Importance of Creativity: Designing experiments & explaining observationsMultidisciplinary science: Integrating multiple fields of studyScience and EthicsScience is often used to guide decision making in society and thus society guides science as wellIf a scientist is not careful he/she can go too far, too fast and do harm in an effort for learning more. To set limits one should use certain guiding lights or criteria:Consider consequences to society / nature (Social / Natural cost)Economic considerations (Financial Cost)Laws (Legal cost)Moral principles (Right vs. wrong cost)What should not be used as models:Authority (What someone in power says)Majority (What most people say its true)Religion (What faith says is true)Precedent (It’s been done before)Costs should always be weighed against benefitsPracticality mattersAll guiding lights should be used in conjunction in case one fails (for example: authority, precedent, and majority sometime determine laws, but even if laws say it is okay it may still not be RIGHT to do so or ECONOMICAL or good for NATURE.Science vs. PseudoscienceHow to tell if something is truly scientific or just sounds like it: Science can be tested and is backed up by evidenceScience can be proven wrong or “not right”Science changes with time, it never stays the same (re-examined)Scientists accept criticism and contest evidence instead of attacking the critic itselfScientists present their information in a fair way; no holding back, no secrets, everything is clearly laid out and explained. Science & PhilosophyExamples of pseudoscienceInternet scamsAstrologyPhrenologyMost often than not science focus on how/when/what questionsMost why questions are for religion/philosophyBe a critic, a skeptic. Do not accept what others say at face value.ASK QUESTIONS!!! ASK FOR PROOF!Steps of the Scientific Method:Asking Questions6ObservationHypothesisTestingAnalysisConclusionCommunicationSteps of the Scientific Method in DetailText in red is advanced and may will be included in examintions e at this level (but it helps further understand the practice of science)Text in green explains how to each step translates to sections of scientific study reports. It may help with writing good lab reports Test in blue provides an example of the step being appliedAsking QuestionsIdentify a problemDiscovering a solution always comes after finding a problem (Without a problem there is no need)Challenging old / creating new (Established ideas & new discoveries)Start from general question and then get more specific (Why is the sky blue? What chemicals are in the sky? What chemical is responsible for making the sky blue? Does the high concentration of nitrogen in the sky have something to do with it being blue?)Can be phrased as a problem statement, instead of a question. “The effect of ____________ (Independent Variable) on ____________ (Dependent Variable)”When writing problems statements (which are sometimes included together with the background and hypothesis in an introduction section of scientific study reports) one should follow the format above. This statement should summarize the objective of the study and specifically mention the problem being addressed. Describes the reason for performing experiment and what are the expected outcomes. This should provide the overall direction for the laboratory investigation and must be re-addressed in the conclusion.(The effect of nitrogen on the color of the sky)ObservationCollecting information (data)Qualitative: Description with words (adjectives)Quantitative: Description with numbers (quantity; measurements; statistics)Understanding the problem before attempting to solve itIn a way, this is the “beginning” since it can be argued that a problem cannot be identified without observation. But sometimes, the problem is whether or not observation was made before the problem was generated, one thing is for sure: Once there is a problem, it is easier to solve it once it is understood After completing observations researchers should be able to include the following in the background/introduction part of their reports: Define the independent variable (what is manipulated) and justify how it will be manipulated Define the dependent variable (what responds to the manipulation) and justify how it can be measuredExplain, with sufficient research evidence, the rationale for the expected relationship between the variables (reasoning or basis for the hypothesis = Data-driven explanation for why the hypothesis is the answer for the research question). Make sure to choose valid sources. Explain how the independent variable should be changed and the reason for the expected change in the dependent variable from successive change in the independent variable (how many levels of the independent variable, or different experimental groups, should be tested and what each will do and why) Note other things that could affect the dependent variable (other than independent variable) and why each should be kept constant during the controlled experiment (how can other things affect the results if not controlled) (Research independent variable, dependent variable, possible links between them, other variables that could also affect the dependent variable, etc.)HypothesisDirect answer to Step 1 (Question) Based on inferences (educated guesses about the phenomenon; based on observations)Note: Remember that inferences are not the same thing as hypothesis. Hypothesis are much more than guesses, they are explanations Proposed scientific explanations for a phenomenonUnconfirmed predictions that must be tested (results must either reject it of fail to reject it)Even if supported in Science, nothing is “right”. Just… “not wrong” until better evidence says otherwise. Important even when they are not supported by data. A good hypothesis leads to further investigation, whether right or wrongSteps to creating a good hypothesis:Use the Magic Sentence:Rephrase question as a problem statement: The effect of ______________ (IV) on _______________ (DV)Identify Variables (using blanks in the statement)IV – Independent (Manipulated) Variable:The ONLY thing changed on purpose to study the effect on the dependent variableDV – Dependent (Responding) Variable : What is measured to see if it is affected by the IV (or what is its response to changes in the IV)Note: The dependent variable is NOT the one that does not change. It may change between groups and that is the difference what must be measured. But it is not the one that is DIRECTLY changed.. It is the one that changes because of the change made to another variable (IV)Formulate hypothesisShort and to the pointDo not explain variables just explain the relationship between themUse if / then statement or one of similar format that establishes a direct one-to-one relationship: If _______________ (IV) then ____________ (DV).(If this is done to the independent variable, then this will happen to the dependent variable)When writing hypothesis (which are sometimes included together with problem statement and background in an introduction section of2 - scientific study reports) one should follow the following guidelines: Note: It is possible to write an okay hypothesis without following a few of these guidelines, but a great hypothesis is the result of following them all.It should be stated directly. Without opening remarks such as “I think”, “I believe”. If necessary, introduce it as: “Researchers hypothesize that….” Or “It is hypothesized that…” It should be a complete, but concise statement (If is too long it is probably not very good)It should establish a one-to-one relationship between variables (Easy with IF.. THEN… statements)It should not introduce another factor that may affect the dependent variable. It should only include the relationship between 2 variables: one manipulated and one being tested (For example, Bad Hypothesis: If there is more wind and rain, then there will be more erosion. In this example, there are 2 variables wind and rain, affecting erosion)Note: There is an exception to this. In some advanced experiments, comparing the effects of multiple variables is actually the point. In such experiments, researchers are actually doing multiple experiments in one. For example, one can increase rain and see what happens to erosion. One can also increase wind and see what happens to erosion. One can also do both at the same time and see what happens to erosion. Such experiments called factorial designs, do all of the above and then analyze data from all three parts together to see the combined effect of 2 or more independent variables on the dependent variable, as well as the relationship between the multiple independent variables. Such experiments either have multiple hypothesis or complex hypotheses with multiple variables. For example: “If there is more wind, then rain will cause more erosion.” & “If there is more wind, then thre is more erosion” & “If there is more rain, then there is more erosion”. Even then, there is always a way to phrase a hypothesis as a relationship between two variables: “Rain erosion effects are directly related to amount of wind”. When possible, this is preferred and is a mark of a good scientist. Explanations for intermediate variables should be discussed in the background section or implied in the hypothesis (For example, Bad hypothesis: Days are hotter than nights because the sun is hitting the unobstructed ground directly during the day. A good hypothesis skips the sun as an intermediate variable and saves the explanation/rationale for the background section. “Days are hotter than nights”. Alternatively, as mentioned in number 4, it is possible to work around this and find a way to phrase the explanation for the phenomenon without introducing an additional variable: “Periods with direct unobstructed ground sunlight exposure are hotter than periods without it.” But remember that hypothesis do need to be considered as possible). Variables are not defined in the hypothesis. All definitions and explanations for variables belong in the background (This is related to item 5. A variable should not be explained within the hypothesis. For example: Bad Hypothesis: Days, or period when the ground is directly exposed to the sun when not obstructed, are hotter than night, when the sun does not directly shine on the unobstructed ground. Either state the summarize definitions as seen in 5, or just use the terms and remember to be as consider as possible). Hypothesis implies clear order of success for all levels of the independent variable (The order of success is clear even if more than 2 groups are used in the design. For example, a study compares how the speed of an athlete depends on hours of training and includes 3 groups: no training, 10 hour, and 20 hours of prior training. A bad hypothesis: “Athletes speed will be best after 20 hours of training”. While this may be true, it does not address the differences between the 10 minute and control group. A better hypothesis would be: “If training time increases, then athlete speed increases”. This serves as a predictor of any possible group. Note: Sometimes it is necessary to make multiple hypothesis for one experiment because one is testing several different groups at once. For example, if one is testing how 3 different schools will perform in a test based on the programs they institute, several hypothesis or a very complex one will be needed to describe the possible results: “Schools using program A will earn higher scores than schools using program B.” & “Schools using program B will earn higher scores than schools using program B” OR “Educational programs used will determine school scores in the following increasing order of success: A, B, and C.” Of course, at times all one cares about is which one is the best so it does depend on the situaton. But in general, it is often preferable to hypothesize about a general pattern than about what happens in a single situation. It should be a testable and falsifiable statement (First, the hypothesis is not overtly obvious or impossible to falsify. There is a chance that it can fail. For example, Bad Hypothesis: “Eating rice ultimately causes death”. Impossible to falsify since every single person who ever ate rice has or will eventually die, unless humanity finds a way to stop death period. Second, a testable statement means that the hypothesis implies the way the variables are manipulated and tested. For example, Bad Hypothesis: “Bad kids are less intelligent”. This does not imply how one manipulates the character of the child, nor how intelligence will be measured. A better hypothesis would be: “Kids who study less tend to perform worse in tests”. The second hypothesis suggests a comparrison between children that study with different intensities and that success will be measured through test scores. This is referred to operational definition of variables. (If children play aggressive video games, then they will be involved in more violence-related incidents in school)Testing Correlational Studies (Data Collection & Observation Only)Sometimes one cannot perform experiments (against the law, cost too much, not practical, would influence the phenomenon being studied; for example: smoking & cancer; animals in the wild; the sun; etc)Then, the best one can do is methodically collect information to establish a connection between eventsRecording of relationships; there is NO true control over variablesResults in a measurement of how closely related two variables are (correlation)If A increases as B increases, then there is a positive correlation, or direct relationship If A increases as B decreases, then there is a negative correlation, or indirect relationshipIf what happens to A has no bearing on what happens to B, there is no correlation. Stronger correlations mean that the relationship is very likely to be true in any particular eventSignificant correlations mean that the relationship (strong or weak) is likely not due to chance aloneResults in Correlation Data Conclusion statements say: A is related to B (or connected to B)Problem with Directionality: One cannot imply A leads to B. B could also lead to A (For example, if one makes a survey comparing trouble-maker kids and well-behaved ones and their video game predilections and find that trouble-makers play more violent games, what does that mean? Do troublemakers prefer violent games? Or is because they played violent games that they are trouble makers?) Problems with Causation: Cannot imply causation (more than one thing is changing at once therefore the outcome cannot be attributed to a single cause. For example: is smoking causing cancer or is the food, exercise, amount of smoking, etc that actually does so? One can only observe what happens to people who smoke versus those who do not. One cannot legally or morally control what, how much, and how people smoke, or what else they do. Even if it was permissible to expose someone to that kind of risk, one would have to force habits and conditions upon subjects in a controlled environment. Therefore no one can say for sure if smoking causes cancer.)Examples: Observation; Archival research; SurveysImproving correlational data:Partial correlations: If there are two variables A & B that could affect C, but one wants to examine the correlation between A and C only, one cannot simply calculate the correlation between A & C and ingnore B. Instead, correlations between all three variables should be calculated. Based on how closely A & B are correlated, how closely A & C are correlated, and how closely B & C are correlated, one can determine how much of the effect betweeen A & C is accounted by the effect of B on C. (For example, if one wants to know if domestic violence is related to school aggression, but is worried that how much violent media counts too, then one needs to see how they are all related to each other and do math/statistics to determine how much of the relationship seen between domestic violence and school aggression has nothing to do with watching TV) Multivariate correlations: One way to introduce control is to measure relationships between IV and DV, as well as several other variables and then use math/statistics to determine the extent to which the effect of the IV and DV depends on other variables or is solely based on the IV. This is like partial correlations, but with many variables all at once. Avoiding representation errors: Making sure the group or model studied represents the whole. For example in a survey one does not want to choose a sample of the population based on first-available, only those interested, or any other specific characteristic, as there could be something special about the group that happened to be available or chose to do the survey (reason why those internet surveys are often useless).Cross-legged design: Measuring correlations between variables at several time intervals may help solve directionality problem(Survey people who play two different types of games and correlate with their satisfaction or fun after it)Controlled Experiments (Manipulating Variables)CONTROL over variables with only ONE thing changing at time (IV); Everything else stays the same (Constants)The same thing is measured in the same way across all groups (DV)Includes a control group where the IV is not changed or manipulatedBased on the comparison between groups (Experimental vs. Control) Results in EMPERICAL DATA ? conclusion statement can say that: A causes BEstablishing causation between IV and DV (Because one makes sure only one thing changes, one can say that change is what caused the outcome observed. For example: One set of people plays a game in an HD 42in flat screen TV, with surround sound, on the best gaming system, with brand new control, in a nice couch, in the comfort of their home, with no distractions. A completely different set of people plays a different game in a standard definition 5in screen TV, with no sound, on the oldest possible system, standing up on rough flooring, inside a bare lab room, with people yelling they are bad at it. Then one cannot fairly compare the groups and say WHICH factor caused them to have more or less fun than the other. So much changed between them that there would be no way to say WHICH factor was the one to blame for different levels of fun. Controlled experiments do the opposite. If one makes similar people play different video games and make sure they play on the same TV, with the same quality sound and video, on the same room, for the same amount of time, using the same system, same interface, and everything the same except the game, then and only then one can compare how fun each game is.)5 Basic Parts:IV – What is changed on purpose to study the effect on the dependent variableDV – What is measured to see if it is affected by the independent variable Constants – What is held at the same level across groups to make sure it will not affect the results Note: It is important that every variable is defined in a way that implies how it can be measured (this is called operationalization of variables) Control Group – No special treatment (IV not purposely changed)Note: The control group IS NOT the group where the DV does not change. There may still be s change, but it will not be because of a change in the IV because no change in IV was introduced in this group. This is important because one can expect the same random change to have occurred in the experimental group (if all other conditions between them are the same, as it should be). This usually indicates that some variable that affects the DV was not properly controlled. Experimental Groups – Different level of the change of the same variable (IV)Constructing a good experimentMake sure to go through steps 1-3 of scientific method firstOnce a questions is raised and understand it (Problem and background) and hypothesis is formulated using the method above (which means variables were identified). One can set up groups:Control Group – Receives no special Treatment (Tx): Independent variable is not changeExperimental Groups – Each receives different levels of the treatment (at least one group)Establish how the IV will be changedEstablish how the DV will be measured to compare the groupsCreate a procedure to do so while making sure to keep everything CONSTANT except the IVSummarize it all in a step-by step procedure (A well written procedure minimizes errors)Note: Good experiments address safety and ethical concerns raised by the protocol. Controls & Methods to Increase Validity:Constants: Keeping other variables that could affect the DV the same ensures that the change in the DV is because of the change in the IV (see establishing causation above)Control Group: Having a group where the IV is not changed allows for analysis of what the change made in the experimental group truly caused (One cannot say that what a change in the dependent variable was caused through manipulation of the independent variable if not comparing a group treated to a group that was not. For example, if one wants to prove earthquakes move furniture, it is necessary to have furniture that is NOT exposed to an earthquake to compare the earthquake group to). For another reason why this is important see the note under control group above. Placebos: One special type of control, especially in human research is to use a placebo. It compares groups receiving treatment with a group that believes they are receiving treatment. Real treatment is not given to this group. For example, this is done in drug trials by giving sugar pills to the placebo group. This ensures that it is not the BELIEF of receiving treatment that makes the Blind: Sometimes placebo studies are done in with a BLIND. In this example, the main researcher does not know which group is which. This is done to make sure that the researcher does not inadvertently or purposefully treat the control/experimental group in a special way to influence (or bias) the results.Double blind: This is when a placebo and a blind are used at the same time (researcher does not know who are the control, placebo, or experimental groups)Sampling: This an important experimental method. When testing a population too big to be measured individual by individual, scientists use samples that must represent the population (See more on sampling under Ways of measuring later in this document)Representative Assortment to groups: When creating experimental, placebo, or control groups, researchers make sure that groups are “equivalent”. This ensures that different in results are not because of a pre-existing difference between the groups. Random assortment: Like random sampling, in random assortment, every member of the sameple has an equal chance of being assigned to each of the groups in the experimentStratified assortment: Researchers make sure all groups have the same number of subjects in a certain category (for example all groups have the same numbers of males and females) Stratified random sampling: There are categories to be equally filled in each group, but members that fit a category have equal chances of ending up on either group. Pre-testing: Another way to make sure experimental groups are equivalnet is to take a baseline measurement and making sure they are all the sameSee more on types of assortment by reading about “Sampling” under Ways of Measuring later in this document)Repetition: Another important experimental method in which researchers take multiple measurements or repeat the entire experiment several times. (not the same as replication)Using averages instead of single measurements ensures researchers that it was nothing peculiar about that ONE time when the measure was made that influenced the results. Ensures researches that the data is trustworthy (reliable)Developing better tools and methods:Using better tools allows for more precise dataUsing better methods reduces experimental errorsUsing Models (Types: Computer / Physical / Mental / Replacement) & Limitations of ModelsThe Importance of Creativity: Designing experiments & explaining observationsMultidisciplinary science: Studying the same thing from different angles(While maintaining other factors constant, randomly sort children into groups and make them play two different violent games to see which one is graded as more fun)General CommentsCommon Sources of Errors:Bias = Intentional or unintentional experimental preconception or pattern of deviation in judgment leading to invalid resultsDesign error: Limitation inherent in the experimental methods Tool error: Limitation inherent from lack of precision, reliability, or validity of the tool used to measure resultsHuman error: Incorrect application of methodsAttrition: Loss of subjects mid-way through the experimentRepresentation error: Sample/model does not properly represent the population or object of study Chance/Random events: In statistics, it is said that any event, no matter how unlikely will happen if there are enough attempts. So long as the probability of an event is not zero, there is always a chance that something might happen. So it may just be that the one time it was done, random chance caused made it all wrongWays of Measuring ControlCorrelational = Relationship between variables (Observe aggression of children in natural situations)Empirical = Controlled comparison of variables (Change situation and study children’s aggression)Number of Measurements:Single Measure Study: Single measurement taken (Aggression on 5 year old children) Repeated Measures: Measurements taken many times over time (Aggression over the years) Timing of StudyLongitudinal: Measurements taken over a period of time (Aggression on children across the years) Cross Sectional: Measurements taken at the same time for groups at different points of the process studied (Aggression on children of many ages compared all at once)Repeated Cross Sectional: Measurements taken at the same time from groups at different points of the process, but done many times over the years (Select children of different ages to begin with and compare them and then continue to do so as they age) Direction of Study: Prospective = Study that looks at data as it happens and on (Aggressive children are tracked forward)Retrospective = Study that looks at data that already happened (Aggression of a child is tracked based on past data) Type of DescriptionsQualitative = Description with words (adjectives) Contextual information (How did the child behave) Quantitative = Description with numbers (quantity; measurements; statistics) Quantifiable information (How many times did the child show aggression)Types of Data CollectionIdeographic = Look at singular causes for one particular event or subjectNomothetic = Look at causes of many events or subjectsNumber of Subjects:Case Study = Single subject (The aggression of one child)Case Series = Compares subjects with pre-existing condition with others that did not have it (at one time or over time)Normative Study = Mean of many subjects (Average aggression of several children)SamplingProbabity Sampling = Attempts to remove bias (gives all individual in the population equal chances of being selected)Random sampling: every member of the population has an equal chance of being selected (randomization tables; pick-out-of-a-hat method, etc)Systematic sampling: Advanced sampling method that delivers efficient and quantifiable probabilities of participants to be in each group by selecting every “nth” number of the populationStratified random sampling: This is done by random sampling off groups formed by dividing the population by certain criteria in order to assure presence of Proportionate sampling: Variation of stratified random sampling when you make sure that sampled group sizes reflect population representationCluster sampling: Unlike stratified samples that make sure there are enough subjects with each characteristic in each group (selected by researcher). In cluster sampling, groups are divided based on pre-existing groups in the population (Introduces the difference between the groups as variable. Usually avoided in true experiments)Multi-stage sampling: Subdividing populations in smaller groups by randomizing in smaller and smaller categoriesNon-probability = Does not attempt to limit biasHaphazard sample: Sampling first available or easiest to find (introduces bias and is avoided in science)Convenience sampling: Studying most readily available subjects (Introduces a variable since there is a chance that there is something peculiar about the group that is most easily available)Snowball sampling: Researcher asks subjects to nominate othersSequential: Sample some and then some more if neededPurposive sampling: Selecting or choosing the group without randomization, but instead targeting a specific groupPhilosophy of BiasSubjective = Perspective, feelings, beliefs, and desires of experimenter are part of the processObjective = Elusive (hard to create/find) manner of research where results are independent of experimenter (Heisenberg uncertainty principle: The closer one tries to measure something, the less control one has over controlling something else, and vice versa. Therefore, studying something actually affects the precision of what is measured)Number of StudiesSingle = Study performed onceMeta-Analysis = Study that combines or correlates results from multiple studiesNumber of manipulationsSingle = Standard controlled experimentFactorial Design = The effects of multiple variables and the interaction between them are accessedOriginalityOriginal = First time measurement is made vs Replication = Study replicates an original study to verify resultsGood designs obey the following criteria (applies to materials and procedure section):Every single thing used is fully described to the extent that anyone replicating the experiment knows exactly what to useList/drawing of all laboratory apparatus used in the investigation. Detailed enough to illustrate the configuration and with measurements and variables to me measured included. If a population is being studied, sample characteristics should be described (For example, in a study of college students the following information should be included: number of people in each sample, how sample was acquired, and demographics such as gender, ethnicity, country of origin, language, age, socio-economic status)Procedures should be written in complete sentences in paragraph format, or given in an organized recipe-like, outline format. Procedures should include every single step necessary for anyone replicating the experiment to do exactly what was done the first time around. This includes:How the independent variable was manipulated (experimental process)How the dependent variable was measured (data collection process)The presence of experimental and control groups (control process)Efforts to maintain control or maintain constant, variables other than the independent could affect the dependent variable (control process)Safety and ethical concernsSteps taken to maximize validity (reliability, precision, and accuracy), including: Sampling procedures (if applicable)Assortment procedures (if applicable)Repetition (if possible)Blinds or placebos (if applicable)Efforts taken to use better tools and methodsEfforts taken to compare results with that of others for validity and generalization purposesData analysis procedure (Appears in the results/analysis section in most advanced scientific reports. But it is also important to specify data analysis procedure including which statistical/quantitative analysis tests or qualitative methods were performed)AnalysisChecking data (checking validity: can this data be trusted and does it to say what it is supposed to say?)Is the data reliable? (Consistency): Measurements yield the same number for the same thing (Example: Scale always give the same weight if the weight itself and the method of weighting does not change. No matter how many times it is used, the result for a certain weight is always the same)Is the data accurate? (Correctness): Measurements yield values that are close to the actual value (Example: The scale give the weight that actually matches the objects weight)Is the data precise enough? (Exactness of measurement): Measurements use tools that can be closer to the smallest measurement possible (Example: Using a stopwatch, one can be more exact than using a head count or a wall clock to measure the passage of time)Determining Confidence Interval / Error Margin: Steps taken to measure the degree can the data to be trusted or the chances the chances that it is wrong. In statistics, there is a degree to which the data is acceptable as representative or reality (95% confidence interval)Measurements are sometimes given with an error margin that depends on the tools used to make the measurement (For example, 150km +/- 3m)Understanding the data (What is the data saying about the Hypothesis?)Visualize data: Charts, graphs, tables, maps (See note below)Bar graphs: Compare different groups Line graph: Change over time (2+ lines for comparing more than one group changing over time)Pie chart: % distribution or how pieces of a whole compare to othersScatter-Plot: Relationships (Common in correlational studies) Stem-and-Leaf / Box-Plot / Histogram / O-give graph: Data distribution or frequency distribution (used to see patterns in the data) Compute data: Do math & statistics to see if the results are significant. Sometimes one will find averages are different between the groups, but considering how many subjects where in each sample, a small difference may have been a simple “chance event”. To say there is a significant difference, the difference must be large enough to be considered something that could not have happened solely because of chance in the confidence interval used in the experiment. (For example, if one compares 2 groups of 100 people to check if listening to music helps test scores and find that on average the music group got 1 extra question right out of 50 when compared to the non-music one, was the difference of 1 point actually significant enough for one to say that music makes a difference in test scores? That’s what statistics is for!)Explain data: Use words to explain resultsWhen writing results/analysis sections of scientific study reports one should follow the following guidelines:Due to the importance of proper note-taking during the experiment, it is advisable to keep and submit notes all notes taken during the lab. This should include quantitative and qualitative information, such as measurements and other observations (errors, visuals, sounds, etc.)Results from notes sheet should be summarized clearly in complete verbal sentencesIf applicable, results section should include a descriptions of the procedures used to analyze data qualitative or quantitative ((This often includes a description of the test performed followed by a verbal description of the result in words and the scores and confidence interval used in numbers. For example, results of a t test indicated a significant difference between the groups, t = 5.64, a = 0.05)Results should also be displayed through the assistance of tables. Whenever possible, graphs should also be used. All tables and graphs must be labeled (columns, rows, titles, axis, lines, etc) directly or through legendsUnits must be included for all measurements and results should include correct number of significant figures. Though results should be listed and statistically analyzed (if applicable), they should not be interpret in this session. That is part of the conclusion(Data Check: Check that variables were defined correctly, measured with the best tools possible, and that averages from multiple measurements were used) (Correlational: Computing the correlation between groups with multivariate analysis for other variables that could also affect the relationship. Design tabling/graphing to demonstrate the relationship.)(Experimental: Computing whether there was a significant difference between groups. Design Tabling/Graphing to demonstrate the relationship and perform statistics.)Note: Look for more information on the website, on how to create good graphs and tables. Mr. DRY MIX: the Dependent variable is the Responding variable that in a graph is recorded on the Y-axis; the Manipulated variable is the Independent variable and is graphed on the X-axisAll graphs and tables should be labeled (columns, rows, titles, axis, lines, etc) and legends should be used when necessary. ConclusionRestating what is being researched, what the results said about it, and how that can be explainedIn other words, this step clarifies the meaning of the results. Good conclusions follow these criteria: Should begin with conclusion analysis statement phrased like: “Results rejected / failed to reject the hypothesis that _____________________ (restate hypothesis)Note: Read about scientific laws later to understand why it MUST be phrased this wayThis statement should be followed by an interpretation of the results that explains why the hypothesis was or was not rejected. If possible and applicable, comparisons with other similar investigations should be madeThen one should infer as to why the results turned out the way they did (especially if they rejected the hypothesis) Limitations or errors during the experiment should be acknowledged (see above) and inferences should be made as to how they affected the resultsThese errors and limitations should be addressed with suggestions on how to improve the study in future replications.Real-life and/or science applications for results should be suggested, with notes as to how generalizable the experiment is (The degree to which the control conditions of the experimetn match the real life conditions being studied is an important factor in determining how and if the experiment should be applied in the field or in real life)Conclusion should suggest future experiments or further research that could further explore the area or phenomenon(Results fail to reject the hypothesis that if children play aggressive video games, then they will be involved in more violence-related incidents in school… )Communication (Sharing & Peer Review)Sharing results with peersPresentations at meetings/conferencesJournals, news, and other media publications (audio/video)Internet, phone, texting, emailing, social networkingEthical Research Conduct Extends to Communication Process: Everything should be reported (even if the data rejected the hypothesis it is important to divulge results). All data helps enhance scientific knowledge. Note: This is a problem in practical science, as some researchers tend to hide or ignore data that contradicts hypotheses of interest. Researchers that cannot produce results come under pressure from those paying for their research or expecting to profit through pratical applications from it. There is also the fear of ridicule in the scientific field or wishes for fame that make scientist ignore so-called “bad results”. Also, scientific publications often reject the publishing of experiments with no viable results. Together, this leads to general distrust of science and to a “drawer-effect” where scientists waste time repeating experiments that already failed in many other lab and to the slowing of scientific progress because of the lack of information on what is NOT acceptable or right. Data should never be tempered with (tell others what actually happened)Note: Under pressure to perform, scientics falsify data (For example, it haappened with global warming research on the late 2000’s). This leads to distrust of science and dangerous ethical repercussions (imagine what would happen if you hide bad results about whether a drug works of its dangers)Allows scientists to conduct peer-review:Verification of data, procedures, and conclusionCriticizing/analyzing the works of othersReplication and exploration of results (different from repetition within an experimental design)Expanding research of others (confirming, refining, or correcting hypotheses)Essential to the advancement of scienceAllows scientists to build upon the works of othersImproves scientific field and humanity as a wholeEnsures that science is always evolving as a method and body of knowledgeSee how to write a report by reading the green text through this section of the notes)(Scientists discuss conclusions with colleagues, present it at conferences, publish articles in scientific journals, and communicate it to the news)Theory vs. Hypothesis vs. Law vs. Principles vs. PostulatesHypothesisBefore testingProposed explanation; predictionBased on inferencesNeeds proof (based on inferences/observations)Describe the relationship between 2 variablesInitial steps of research and very specificDesign to be refutedTheoryAfter testingWorking explanation; based on conclusionsBased on a lot of empirical or at least correlation dataHas proof (Cumulative body of evidence)Describe the phenomena as a wholeOften the result of years or research combining hypotheses and the study of several phenomenaMost powerful explanation available (simplest / best)Hard to discard, so often adapted (well-substantiated)Not the same as the “theory” as it is used in societyLawsNot the same as society lawsStatement of fact that universally and perpetually explains a specific natural phenomenonConcise and specific description of a natural relationship that is constant throughout the universeAnalytic statement with an empirically determined constant (description based on evidence)In other words, it is completely accepted as universally trueUniversally true and invariable concept = Applies in many scenarios (though often not in every single one) & Should not change (has been observed many times with little or no contradiction)Theories often contain sets of laws or a theory can be implied from lawsTheories do not become laws or vice-versaExamples:Newton’s Laws of motionKepler’s Laws of planetary motionGravity: “Objects that have mass attract others” (Gravity)Thermodynamics (1st Law): “Energy/matter can be changed from one form to another, but it cannot be created or destroyed. The total amount of energy and matter in the Universe remains constant, merely changing from one form to another.” (Thermodynamics 1st Law)Thermodynamics (2nd Law) “In any closed system, the entropy (disorder) of the system will either remain constant or increase.” OR “A cyclic transformation whose only final result is to transform heat extracted from a source which is at the same temperature throughout into work is impossible.” OR “A cyclic transformation whose only final result is to transfer heat from a body at a given temperature to a body at a higher temperature is impossible.”Thermodynamics (3rd Law):It is impossible to reduce any system to absolute zero in a finite series of operations. The entropy of a perfect crystal of an element in its most stable form tends to zero as the temperature approaches absolute zero. As temperature approaches absolute zero, the entropy of a system approaches a constantThe Idea of scientific law is problematic because it assumes that it is something is set in stone and ALWAYS true. The very nature of science is to question things. Even as it tries to understand and predict it all, science accepts the impossibility of ever reaching a final conclusive solution. Very little is truly universal and never changing. Even concepts like entropy, conservation of matter/energy, and gravity, and laws of motion do not apply in “every” scientific scenario. A few examples of ideas previously seen as universal scientific laws being contradicted: In certain places of the universe, matter and energy disappear and/or appear into/from nothingness (black holes for example). The effect of gravity is inexplicably negated at a cosmic scale by dark energy which causes the universe to expand at an ever increasing rate. Even though it was postulated as impossible, absolute zero is the actual theoretical end for the Universe. Newton’s laws of motion have been expanded, revised, and completely challenged by ideas such as relativity. Objects are no longer described as falling on straight lines, but as rolling down gravity hills of bent space-time continuum around massive objects! Conclusion: There should be no “scientific law”, just very good theories. However, people use it to describe precepts that seem to be almost always universally true and very unlikely to change because they are based on theories that have survived testing a multitude of times in many different situations/applications. They are useful descriptions of the nature of things that is so widely accepted that they go beyond theory (working explanations). .Note: In the past, it was much easier to get to this point as skepticism in science was not as common as it is today. If a famous scientist or a scientist that pioneer a brand new idea came along with a theory, he/she often call the precepts of his theory “Laws” (as Newton did with his Laws of Motion or Gravity)PrinciplesDictionary: Rule that must be followed, that is usually followed, or that can be chosen to be followed. Also the inevitable consequence of somethingIn science: a rule of conduct; a fact of nature; a descriptive or fundamental law Examples: “The Scientific Method” is a scientific principle, or rule of conduct; “Birds are alive” is another example of a principle that is a fact of nature, or descriptive; “The speed of light is the ultimate speed limit of unaltered space-time continuum in universe”, is an example of a fundamental law that if you were to change would change everything we know about physics. PostulatesA logical assumption not necessarily proven or demonstrated by evidence but considered as either self-evident or necessarily trueExample: “It is possible to draw a straight line from any point to any other point”; “light comes from a sourceLevels of Truth: Change & Durability of ScienceThere is a reason why stats professors tell us to say results failed to reject hypothesis (as opposed to supported). It's as if in science, nothing is ever RIGHT. Just "not wrong", for now...until better/more data fails to reject new hypotheses and thus help create a new theory. One can never say a hypothesis was proven, confirmed, or supported. One can only say that the results cannot disprove it. A hypothesis will only stand for as long as it is not rejected by data. Science does not prove things, it explains them as best it can for the time being. It focuses on bringing more clarity by questioning things. Therefore, scientists never accept things. Scientists say that based on current data the present explanations should not be rejected. Nothing in science is sacred or forever protected from change. EVERYTHING in science can change. Science welcomes criticism: Skepticism is at the core of scienceScientific knowledge is constantly under scrutiny: RIGHT vs. NOT WRONGDebate, replications, logic/critical thinking, argumentation, consideration of alternative data/explanations are constantInformation sources, evidence, and methods are constantly challenged for accuracy/reliability/validityLearning from right/wrong or success/failureResult is durable knowledge that withstands the test of time or changes according to new data or interpretationsThus scientific knowledge is both robust/durable AND open to changeExample of the Method in ActionAsking QuestionWhat causes erosion?ObservationStudy what is erosion and possible causes (wind, deforestation, water, grade or inclination of ground, and other factors that could explain increasing levels of it). Learn about the problem before trying to solve it. For example, one finds that the trees prevent erosion through their roots as they keep the soil in place and drain water.Hypothesis: Step 1 (Rephrasing Question as Problem Statement using Magic Sentence):“The effect of cutting down trees (IV) on erosion (DV)”Step 2 (Identify variables)IV – Cutting down trees (Deforestation) & DV – Amount of erosionStep 3 (Formulate hypothesis using concise if/then statement that does not explain what the variables are, but does explain the relationship between them)“If more trees are cut down (change on the IV), then there will be more erosion (predicted change in the DP)”TestingOne can simply observe erosion and collect data on where it happens more based on the number of trees, but then there could be no causal relationship. One could only say deforestation is related to cutting down trees. Or one can do an experiment: IV - # trees cut downDV – Amount of erosionCG – No trees cut downEG – Trees cut downEG2 – More trees cut downConstants: Inclination or grade of floor; type of soil; type of plants; amount of rain/wind; etcDescription (This is in recipe-like, outline format. But to save space, it is presented as a text)1) Create lab with fields of the same type of soil in the same gradient (The same amount of initial soil should be used in all). 3) Plant the same number of trees in all soils (All trees should be of the same type). 4) Count the numbers of trees in each field once enough time has been given for the trees to reach maturity and cut down extra trees to make sure that each group has the same number of trees. 6) Manipulate the number of trees per field in the following manner: In the control group, no additional trees will be cut. In the experimental group 1, 50% of trees will be cut. In the experimental group 2, all trees will be cut. 7)Introduce rain/wind through lab environmental control in the same fashion for all groups. 6) Measure the amount in tons of soil runoff in through collection bins at the corners of lab fields connected to gravity-sensitive scales. 7) Repeat experiment at least 3 times. 8) Perform multigroup anova statistical test to test for significant difference between average soil runoff of groups.AnalysisConsider validity, reliability, precision, and accuracy of measurementsRepetition and calibration of tools leads to reliability and accuracy. Bins will be weights by balances that calculate weight down to cents of a gram (for increased precision)Graphs, plot, chart, or table results; Explain results; Do statistics to compare groups with mathResults indicate that there are 3 times more runoff in setting with no trees than in the control group and 1.2 times more runoff if 50% of trees are cut down.Conclusion (incomplete, this is just a sample of what should be said)Summary statement: Results failed to reject the hypothesis that cutting down trees increases erosionThis means that more cutting down trees makes it more likely for rain/wind to carry off top soilFarmers and people should refrain from cutting down trees if soil is to be preservedIt is possible that some of the runoff did was not properly collected into the bins. Future research should use a system that catches all the runoff through grates on the ground bellow the lab assuring that all runoff is accounted forFuture research should also explore how different types of soil, plants, and varying amounts of wind/rain affect erosion.ShareResults discussed with colleagues through networkingResults are published in journal of agriculturePresented at geology conferenceAppeared in the news on TV and radioOther scientist replicated and expanded the research…Theory vs. HypothesisUnlike the hypothesis above, a theory would come after many scientists spend time studying erosion and find a lot of evidence to support many hypotheses about it. A theory of erosion may include many more factors that cause it. For example, it discusses amount of wind and air, types of soil and plants, gradient of the floor, quantity of plants, etc. This is a working explanation for erosion based on collected evidence, not a prediction or possible explanation (which would be a hypothesis)Law: As part of the working theory of erosion, statements that describe the known phenomena which have been consistently observed in many circumstances are called the Laws of Erosion. But since this may yet change if new data adds to the understanding of erosion, it is pointless to call it a Law, as if it is always true and it will never change – no exceptions. Still, if a lot of research is done and no exceptions, or better explanations are ever found for a very long time… then people cannot resist calling it a law. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download