Biological Analysis

A researcher measures a distance at 1.315 m The actual value is 1.200 m. What is the percent error?

% error = [(1.315 m - 1.200 m) / 1.200 m] X 100 = 9.583%

Give an example of a 2 x 4 factorial design in an experiment. Explain why it is factorial and why it is 2 x 4.

Four different doses of insecticide are exposed to flies and the percent of dead flies are measured after 1 and 2 hours. A factorial treatment has two independent variables. The two independent variables in this example was the dose of insecticide and time after exposure. In 2 X 4 factorial design, the 2 is the two time periods and the 4 is the four doses of insecticide.

What can graphs tell you about a data set that simple descriptive statistics cannot?

Graphs can show relationships and reveal qualitative patterns better than descriptive statistics. Graphs can reveal or hide the meaning of the results based on how they are constructed.

Give an example situation where a researcher's experimental design does not show good internal validity.

A researcher hypothesizes that people getting at least 30 minutes of daily exercise will have a lower BMI than people getting less than 30 minutes of daily exercise. However, if the researcher does not account for the diet of the subjects, the diet can be a confounding variable. The presence of a confounding variable shows poor internal validity.

Give an example situation where a researcher's experimental design does not show good construct validity.

A researcher wants to test the effects of classical music on the growth of plants. However, the researcher plays rock music to the plants. This shows poor construct validity as the methodology does not address the actual scientific question.

What is a scientific argument?

A scientific argument uses evidence to make a case for whether a scientific idea is accurate or inaccurate. Scientific ideas, expectations, and observations make up a scientific argument. A scientific argument is an argument about a scientific explanation for something and both use facts and empirical evidence and data to back up their claims, not just arguing based on opinion and feelings. A scientific argument uses evidence to make a case for whether a scientific idea is accurate or inaccurate. There are three components of a scientific argument - the claim (or the explanation), the evidence (or the observations), and the rationale (or the reasoning). Scientific argumentation requires scientists to support their claims (either for or against a particular idea or explanation) with evidence that has been gathered through observation or experimentation and then to use logic and reason to justify why that evidence supports their claims

Why can't a scientist work in isolation?

A scientist cannot work in isolation because an individual may be biased. Involving the scientific community can counteract the bias of the individual. Also, scientific conclusions must be replicated by others to be accepted as a sound explanation. Furthermore, working with other scientists can provide different points of view.

Give an example of a nest treatment design. Explain what makes it nested.

An example of a nest treatment design is if a researcher is testing the effect of a cholesterol lowering drug among males and females of Caucasian and non-Caucasian descent. This is a nested design because the analysis of the effect of the drug on males and females is under the broader categories of Caucasian and non-Caucasian. The analysis is only within the subject groups and interactions cannot be determined.

Give an example of a single blind study. Why was the single blind study used?

An example of a single blind study is where the researcher either performs actual knee surgery or a placebo sham surgery on patients with knee pain, but the patients do not know which one they received. The researcher knows which treatment the subjects received but the subjects do not know. The single blind study can prevent the subjects from reacting and behaving in some particular way based on what treatment they received. This helps reduce bias.

What are pseudoreplicates? Give an example.

Pseudoreplication is the process of artificially inflating the number of replicates. Pseudoreplicates are subsamples presented as true replicates. They misidentify the experimental unit. For example, researchers who collected multiple leaves from one oak tree instead of collecting leaves from different oak trees and presenting them as true replicates. This would be pseudoreplication because the leaves are from only one subject.

Give an example of the use of randomization in an experiment. Be specific.

Randomization is a method in which study participants are assigned to a treatment group based on chance alone. In an experiment to test the effects of a weight loss treatment in women, the researcher will randomly assign a number to each of the female test subjects. The researcher can use a random number generator to divide the female test subjects into two groups for the treatment and control group.

What is randomization in an experiment and why is it important?

Randomization is dividing the test subjects to treatment and control groups randomly based on chance. Randomization is important because it helps reduce bias. It helps us obtain a more representative measure of a population.

Distinguish between rationalism and empiricism. How does each "work?" What are the pros and cons of each?

Rationalism is thinking in which knowledge is developed through reasoning. Information is carefully stated in a major premise, minor premies, and finally a conclusion is drawn. The pros of rationalism is that it does not require any type of measurement because of well-established premises. The cons of rationalism is that false conclusions can be drawn if any of the premises are wrong or bad reasoning is used. Empiricism is gaining knowledge through observation. It is knowing by experiencing with our senses. The pros of empiricism is that it can explain new phenomenon and support conclusions based on evidence and observations. The cons of empiricism is that senses can deceive us and yield incorrect data and conclusions.

Give an example of when it might be useful to use a reference line or reference area in a graph?

Reference lines are used to compare, reference, or measure against the data values displayed in the graph. For example, in a graph showing the grades of all the students in Biological Analysis class, the reference line could show the average grade of the class. Also, in a graph showing the monthly sales of a company, the reference line could be the sales goal.

What are advantages and disadvantages of the repeated measurement design?

Repeated measurement design is measuring the same subject more than once. Advantage of repeated measurement design is that it can show the effect over time because the same subjects are measured over time. It may reduce confounding variables caused by individual differences. Also, fewer subjects may be needed. Disadvantage of repeated measurement is that time can become a confounding variable because the subject may have done something unrelated to the treatment that may affect the results. It also can be hard to replicate if the time between measurements has to be long.

What is the difference between sample size and the number of replicates in an experiment?

Replication assigns more than one experimental unit to a treatment group and assesses random variation. Sample size is the total number of experimental units. Sample size is the total number of units tested in the experiment. The number of replicates is the number of units that are in each treatment group.

Explain why we replicate experiments.

Replication provides a way to assess random variation. Replication helps to isolate the effects of treatment by controlling for variation caused by chance effects. Replication can also improve the precision of the estimates of the variables in the study. Replication that increases the sample size can increase the power or the likeliness of detecting a significant difference if one exists.

What about the attached graph is misleading?

The dates of the different graphs are confusing and misleading. The top graph and the bottom graph have different dates. The top graph shows a range ('67-'68), but the bottom graph only has one number ('67). The reader does not know if the '67 refers to '66-'67 or '67-'68. This can lead to misrepresentation and misinterpretation. Also, the 3D format of the top graph makes the data and the differences among the data seem bigger. The 3D format can also make the data hard to read.

The results of an experiment in which teaching method is the experimental treatment, used with a class of low achievers, do not generalize to heterogeneous ability students. What type of threat to validity would this be? Explain.

The external validity is threatened. External validity asks whether the results can be generalized to other situations. Since the results of this experiment cannot be applied to a group of heterogeneous ability students, the external validity is threatened.

What is Tufte's Rule?

Tufte's rule gives guidelines to making effective figures. Tufte's rule says to maximize the ratio of ink to data, to maximize dimensions of information, and minimize dimensions of presentation. Minimize the ratio of ink to data. This basically means use the maximum amount of data with the least amount of ink.

What is required to come up with a valid conclusion?

Valid conclusions need a question and hypothesis. The hypothesis or idea has to be supported through observations and data. Also, valid conclusions can be formed by deducing from premises or deductive reasoning.

What does the "validity" of an experiment refer to?

Validity refers to how well supported the conclusion drawn from an experiment is. One of the fundamental tasks of research is to ensure the validity of the experimental procedures and conclusions. Four types of validity include internal validity, external validity, construct validity, and statistical validity.

What is it important to show variation in your graph?

Variation can be shown through standard deviation, standard error, or confidence intervals. Error bars can indicate how spread apart the data is compared to the mean. Small variance will give the reader more confidence about the results. Large variance shows that the results of the experiment are not meaningful. Variation can help show the difference between the two variables that the graph is presenting.

Can the sample size be too large? What happens if the sample size is too large?

Yes, a sample size can be too large. If the sample size is too large, the power of the experiment goes up. If the power goes up, you may be more likely to detect an insignificant difference. The results may be too general to be significant. Also, a large sample size might take up too much time and resources.

I was determining how lethal a new pesticide is to fruit flies. I cultured 1000 fruit flies in 20 vials (50 in each vial). Five vials received the pesticide and five vials did not Do I really need five control vials? Explain. Analyze this experiment for replication. What is the sample size (n = ?)?

1. Five controls vials might not be necessary because the experiment is only testing the effects of one pesticide on fruit flies. The flies in the control vial did not receive any pesticide, and the control vial is only used as a comparison to the vial that received the treatment. There is no need to separate the flies into different vials for the control because they are not receiving any treatment. Decreasing the number of vials and increasing the sample size of fruit flies in the vial may be a more effective way to conduct the experiment. However, it is good that you have some type of replication, but 5 may not be necessary. 2. The experiment has internal replication as there are 5 vials with 50 in each vial receiving the treatment and 5 getting the control which gives lots of replication to both the treatment and the control groups. You also have 50 flies in each vial which gives replication within each vial as well. Replicating the experiments can help assess random variation and increase the validity of the results. The sample size is the total number of experimental units in the experiment. The experimental unit in this experiment is the number of vials. The data that we are collecting is the percentage of flies that die in each vial exposed to pesticide. We are treating the vial as an experimental unit and not the individual flies themselves. So the sample size of the experimental vials would be 5 as five vials are exposed to pesticide. The sample size of the control in this experiment would be five. Total sample size = 10

What are the graphical cues ranked from most to least accurate graphical perception?

1. Position along an axis 2. Length 3. Angle or slope 4. Area 5. Volume 6. Color and shade

Describe four things that can be done to keep the behavior of the researcher from threatening the validity of the study.

1. Using double-blind or triple-blind experiments so that even the researchers are blinded can help reduce bias of the researchers 2. Deciding what statistical test to use to analyze the data before the experiment is conducted so the researcher cannot choose a statistical test to match the results that he wants 3. Using a standardized protocol and data collecting procedure that is set beforehand so the researcher cannot change the procedure to suit his data 4. Obtaining objective measures that are clearly specified and having multiple observers can help reduce the bias of the researcher

What is a bivariate graph?

A bivariate graph shows the relationship between two variables that have been measured on a single sample of subjects. Bivariate graphs have two variables. The independent variable is on the x-axis, and the dependent variable is on the y-axis. There are tick marks that are intermittently labeled.

Give an example of a type of investigation what would have a low level of constraint.

A naturalistic study would have the lowest level of constraint. Naturalistic studies aim to study the subject and its behavior in the natural setting. There is no manipulation of objects or the settings. Controlling the confounding variable can disrupt the natural habitat of the subject and mess up the experiment. An example would be observing the way that baby tigers learn how to hunt.

How is a hypothesis different than a theory? How are they similar?

A theory is usually a broader explanation of a phenomena than a hypothesis. A hypothesis is formed through observations and logic. A hypothesis has to be tested through the scientific method to be either supported or disproved. A theory is an explanation of a phenomena that has already been supported by data and experiments. A hypothesis is someone's explanation for why something happened- it will be tested and may or many not be shown to be true. a theory has been tested and for the moment has consensus on why it occurs and what it is. they are similar because both still are not complete answers and 100% verifiable. theories do change, although they change much less than hypothesis do.

How can accuracy and precision of a measurement be increased?

Accuracy can be increased finding the percent error and subtracting that value from our measurements to remove the systemic error from our measurements. We could also calibrate the balance or the equipment. Precision of a measurement can be increased by using the smaller increments on a measurement device. We can also measure in the same way to get similar results. You can increase accuracy by doing more trials/experiments as well as calibrating the balance or instrument. You can increase the precision by taking smaller measurements several times, being very careful, and using a more exact measuring device.

Distinguish between accuracy and precision

Accuracy characterizes the system's ability to provide a mean close to the true value when a sample is measured many times (related to systematic error). Accuracy is basically how close you are to the actual value. Precision characterizes the system's probability of providing the same result every time a sample is measured (related to random order). Precision is basically how close your measurements are to each other.

Why are assumptions necessary in science?

Assumptions are a starting point in science. A hypothesis can be considered an assumption the scientists makes before the experiment. Also during the experiment, the scientist assumes that all variables other than the variable being tested will remain constant. Assumptions are necessary for experiments.

Distinguish between basic and applied science. Is one "better" than the other? Explain.

Basic science is based on understanding fundamental aspects of science that are fairly constant and can be broadly applied. Applied science is the application of basic information to solve practical problems. One is not necessarily better than the other because basic science and applied science are codependent. Basic science provides the basic components of a topic which is then used by applied science to solve practical problems.

Write a good hypothesis and justify why it is good. Tell why it is: a) Testable b) Specific c) Show direction d) Establish variables

College students who sleep for more than 8 hours a night will receive a better grade in Biological Analysis than students who sleep less than 8 hours a night. This hypothesis is testable because you can randomly divide the class into two groups, and have one group sleep for at least 8 hours a night and the other group less than 8 hours a night. The hypothesis is specific in that the subjects of the test are limited to college students taking Biological Analysis. The hypothesis shows direction because it predicts that getting at least 8 hours of sleep will improve the student's grade in the class. The hypothesis has independent (amount of sleep) and dependent (grade in Biological Analysis) variables.

What makes pie charts easy to misread?

Comparing data on pie charts can be difficult especially if the data is not displayed next to each other. Also, the angle or area of the slices of pie charts can be difficult to guess just by eye. This makes it difficult to estimate the percentage of each slice. Differences of the different slices may go unnoticed by the reader. Pie charts have humans distinguish between both areas and angles. Humans have a hard time perceiving areas and angles compared to other things like position and length. Sometimes to fix this, pie charts give the numbers and percentages as well which basically wastes space because they could have just given the numbers, if you have to use both, it is a waste. Sometimes giving percentages can lead to faulty conclusions as well. Pie charts are mostly used for visualization, and many times can be hard to understand due to the things mentioned earlier which can lead to false conclusions. In the end, it really does not easily compare values.

What is a contour plot?

Contour plots are a way to show a three-dimensional surface on a two-dimensional plane. The plot shows the relationship among three variables (X, Y, and Z) in three dimensions while plotted in a two-dimensional format. It graphs two predictor variables X and Y on the axes and a response variable Z as contours. Contour plots are a way to show a three-dimensional surface on a two-dimensional plane. They work by plotting constant z slices, called contours, on a two-dimensional format. For a given a certain value for z, lines are drawn for connecting the x,y coordinates where that z value occurs.

Why isn't high correlation enough to establish causation?

Correlation focuses on quantifying the relationship between two or more pre-selected variables. Correlation states the relationship but does not give an explanation of the relationship. Causation states that any change in the value of one variable will cause a change in the value of another variable. Just because two variables generally rise and fall within the same time frame, it does not mean that one variable influenced a change in the other variable. It could be due to coincidence and be completely unrelated.

What is the difference between a correlational study and an experimental study?

Correlational studies focus on quantifying the relationship between two or more pre-selected variables. Correlational studies have a lower level of constraint than experimental studies. Experimental studies aim to make comparisons under different and controlled conditions. Although causality can sometimes be inferred, results may not be applicable outside of the experimental setting.

What is count/enumeration data? Give an example.

Count data/enumeration data combines observations into frequencies. It is nominal discrete. An example is a number of vacation days versus the number of work days in a year. Another example is the number of guys versus the number of girls attending Southern.

What is the role of creativity in science?

Creativity plays a big role in science. The beginning of the scientific method is observing the natural world and asking questions. The scientist then forms a hypothesis that may provide possible explanations. Creativity is central in forming a hypothesis to a question that does not have an answer yet. Also, advancements in any field, including science, can only be made by a person thinking outside the box and trying something new in a creative way. Creativity also helps scientists move away from tenacity/tradition or authority as their method of acquiring knowledge to rationalism and empiricism.

What is data analysis? Why is it necessary in science?

Data analysis is applying data obtained through experimentation to statistical or logical techniques to interpret the data. Analyzing data allows it to become evidence. Analyzed data can either support or disprove a hypothesis. Data analysis is necessary in science because raw data does not mean anything. Raw and uninterpreted data cannot be used to support a hypothesis. Raw data also leaves room for other people to misinterpret it and draw incorrect conclusions.

How could we control for edge-effect? Give a specific example.

Edge-effect is a subtle confounding variable that occurs due to the positioning of the samples. For example, plants in the same greenhouse can receive different treatment due to their position in the greenhouse. Different places in the greenhouse might have more access to sunlight or water. Rotating the plants and spacing out the plants can help reduce the edge-effect.

What methods are used to reduce error variance? Be specific.

Eliminating confounding variables can help reduce error variance. A large sample size and randomization of subjects can also reduce error variance. Taking careful measurements with accurate equipment will also help. Objective measure should be used instead of subjective measure. Replication can also help reduce error variance.

What is an empirical observation? What makes it empirical?

Empirical observations are obtained using the senses. They begin as descriptive observation and then become quantitative measurements. They are empirical because the observations are capable of being verified or disproved by further observation or experiment.

Define/distinguish between empirical and deductive science. Explain the strengths and limitations of each.

Empirical science is a systematic approach to understanding that uses observable, testable, repeatable, and falsifiable experimentation to understand how nature commonly behaves. The strength of empirical science is that a scientist can manipulate a variable to perform experiments to obtain useful data. The weakness of empirical science is that the conclusions formed are specific which may be too subject to be generalized or even invalid. Also, experimentation is not possible in all situations. Deductive science builds models of complex systems to explore relationships. The models are used to make predictions and then compare the predictions with actual measurements. The strength of empirical science is that it can be used to form the correct model for a complex system and can be used to make accurate prediction to give us a clearer understanding. The weakness of deductive science is that it may be hard to figure out and fix a model if the model is wrong. It may also not be as detailed when it comes to testing.

How do rationalism and empiricism work together?

Empiricism gathers knowledge through observation and data to form a conclusion. Through repeated experiments, the conclusion can be validated. The validated conclusion can be used in rationalism as a premise. The premise can be used to finally arrive at a broader conclusion. Empiricism can deal with the measurable, and rationalism can deal with the unmeasurable or ideas that can only be arrived at through reasoning.

What do you learn by each of the following: exact replication, systemic replication, and conceptual replication?

Exact replication: Repeating the experiment as nearly as possible in the original way helps determine the accuracy of the original results. It can show any possible errors in the original experiment. Systemic replication: Doing the study with some variation as compared to original study can help show how different variables can change the results. It can also increase the generalizability of the study. Conceptual replication: Testing the same question by a different manner can help show the best method for testing the question. It can reveal any biases in the experiment.

Define/distinguish between experimental and historical science. Explain the strengths and limitations of each.

Experimental science is based on conducting experiments to test hypothesis and discover properties of nature. The strength of experimental science is that it can be used to repeat experiments to further support the existing conclusions or come up with new conclusions. However, the weakness of experimental science is that not everything can be tested. Some natural phenomena such as black holes or space are just out of reach for experimentation. Historical science seeks to reconstruct the past and seeks to explain the present by reference to the past. The strength of historical science is that it can be used to make predictions because law of natural are fairly constant. However, the weakness of historical science is that information about the past may be limited and incomplete. It can also fail to make new discoveries.

Give an example of a type of investigation that would have a high level of constraint.

Experimental studies would have the highest level of constraint. For example, a study that tests the effects of a new drug would need a high level of constraint because the drug could have multiple effects. The researcher may have to be specific and test for a specific variable such as the effects of the drug on blood pressure. The researcher would need to control the confounding variables.

Distinguish between factorial, nested, and gradient designs.

Factorial treatments can determine interactions between multiple independent variables. Nested treatments cannot determine interactions. They can only analyze within subjects. Gradient treatments are often used when direct manipulation is less feasible. The variable is treated as continuous.

What does it mean when we say that an investigation has a high level of constraint?

High level of constraint means that there is a rigorous control of the confounding variables. High constraint helps the researcher have high confidence that the results are not confounded and helps establish cause and effect. However, the results may be less generalizable to natural conditions.

What is the difference between a histogram and a bar chart?

Histograms are usually used to present continuous data. They show distribution of a given value in intervals. The bars in histograms are adjacent to each other no spaces in between. Bar graphs are usually used to display categorical data. They show comparison between data. The bars in bar graphs are separated from each other so there are spaces between them.

Do you think that the term scientific creationism is an oxymoron? Why or why not?

I do not think the term 'scientific creationism' is an oxymoron. Methodological naturalism (demarcation argument) says that scientific explanations must only appeal to natural processes. I personally believe that science can be more broadly applied than just to natural laws. I think natural laws can be interpreted through the Bible. I also believe that observations and data can be used to support creation. Science and creation can exist together and support each other.

"The behavior of the researcher may influence the independent variable, thus threatening the validity of the study." Give an example identifying a possible behavior that is a threat to validity.

If a researcher was hired by a pharmaceutical company to research the effects of a blood pressure medicine created by the company compared to other brands, the researcher's behavior may be biased. The researcher may be biased because the researcher is representing the company and is getting funding from the company. The researcher's goal is to show through research how this specific brand of medicine is the best. So, the independent variable (the brand of blood pressure medicine) can be influenced. The researcher could intentionally compare the company's medicine to other brands that he knows is not as good and leave out certain brands. He might also run different statistical tests to support the result that they want. This threatens the validity including statistical validity of the study.

What happens when the sample size is too small?

If a sample size is too small, the power of the study will decrease. This makes it less likely to detect a significant difference if there is one. Also, you might not get accurate results because outliers and mistakes can greatly affect the results. The results also cannot be generalized to larger populations.

Define love. Give an operational definition of love - one that can be used to make scientific observations

Love is an intense feeling of deep affection. Love is choosing someone over yourself. An operational definition of love can be measured through different acts such as telling the person that you love them and prioritizing the other person.

Why is publishing critical to progress in science?

If scientific research is not published, the study cannot be peer reviewed. Peer review is necessary to check for the validity of the study. Publication also allows the population to access the study and learn new things. Also, other scientists can learn about the study and the results. The new explanation can answer questions and provide further knowledge.

What does it mean when it says that the results were confounded?

If the results were confounded, something other than the independent variable could have affected the results. Factors other than the treatment influenced the results.

Give an example of a double blind study. Why was the double blind study used as compared to a single blind study?

In a study that researches the effects of a cholesterol lowering medicine, a double blind study would mean both the subjects and the researcher are blinded. Both the subjects and the researcher do not know which group received the actual cholesterol lowering medicine and which group got the placebo. A double blind study was used instead of the single blind study to reduce bias of the researcher. The researcher cannot give better observations toward the ones who received the treatment more than the ones who received placebo.

List and briefly describe the different types validity.

Internal validity aims to identify and account for potential confounding variables. It looks for any potential error in the experiment. External validity aims to see if the results can be generalized to other situations. Construct validity aims to see if the methodology of the experiment sufficiently addressed the scientific question. Statistical validity aims to ensure that the correct statistical analysis was used.

Does science only work by means of experimentation? Why or why not?

No, science does not only work by means of experimentation. There are phenomenon which cannot be conducted in an experiment such as outer space. However, observations can be made about the phenomenon, and conclusions can be drawn. Also, theories in historical science cannot be tested or experimented directly, but is still considered science.

Nominal data must be mutually exclusive and exhaustive. Explain and give an example to illustrate.

Nominal data is mutually exclusive in that each observation (person, case, score) cannot fall into more than one category. For example, if the two categories are male and female, a person cannot be both male and female. Nominal data is exhaustive in that the system of categories system should have enough categories or all the observations. For example, when we roll a die, we expect to get either 1, 2, 3, 4, 5, or 6. Every time we roll a die, we will get one of these results. Sticking with this example, the data is mutually exclusive because each time we roll a die, the results can only be one number.

Distinguish between objective and subjective measures. Give an example of each.

Objective measure is a impartial measurement that is less prone to bias. No emotions or opinions influence the measurement. An example would be measuring the number of students taking a class. Subjective measure is based on personal opinion and is more prone to bias. An example would be measuring the happiness of pets.

What is peer review? Why is in useful? What are potential problems?

Peer review is when experts in a particular field evaluate the work and research done by other scientists. Peer review is useful to check the quality and validity of the study. It maintains the quality of scientific standards. Peer review is useful because it can show the author potential mistakes and different ways to improve the study. It is also useful to build trustworthiness of a study. Peer reviewed studies are accepted to be reliable. Some potential problems of peer review may be authority. Some people might accept something as true just because it was from a peer reviewed article and not by critical thinking. Also, studies that attack or contradict commonly accepted ideas may have a harder time being approved by fellow peers due to bias.

What is a placebo? Why is it needed? Explain.

Placebo is a pill, medicine, or procedure that has no therapeutic effect. It is used as a control in testing new drugs. It is used to see if the results in the study were due to the treatment or due to unrelated causes that might be related to psychological processes. The placebo helps account for the placebo effect.

What is the difference between statistical power and effect size?

Power is how likely you are to detect a significant difference if one exists. As sample size goes up power goes up but you may be more likely to detect an insignificant difference. Effect size is how meaningful the difference is.

Why is science difficult to define?

Science can be difficult to define. The term "science" can apply to a broad set of studies on different phenomenas. The science checklist states qualities (such as focusing on the natural world, aiming to explain the natural world, using testable ideas, relying on evidence, involving the scientific community, leading to ongoing research, and benefiting from scientific behavior) that make something scientific. However, the science checklist does not cover all parts of the scientific process as stated in the Demarcation arguments. For example, the Demarcation arguments sate that science must be directly observable. However, many things in science are not directly observable such as sub-atomic particles or black holes. Also, historical science looks at past events which are not observable. Past events are explanatory, not natural law.

What implications does the following statement have for science? "Our worldview impacts our interpretation of the evidence."

Science is more subjective than widely believed because a person's worldview can cause different interpretation of the same set of evidence. Our preconceived beliefs can affect the way we see the data and lead to different conclusions. Worldview can be a source of bias.

Why isn't science predictable and predetermined?

Science is not predictable and predetermined because science is a process. Scientists do not have all the answers. Scientists are continually learning through observations and experiments. Since scientists do not know all the answers, the conclusions formed from the data can be unpredictable and not predetermined. As scientists learn new information and gain new data, old theories and conclusions may be shown to be untrue. Science is not predetermined, but continually growing and changing.

What makes scientific idea more trustworthy and more likely to be true?

Scientific ideas are more trustworthy and more likely to be true when the idea is supported by observations or data from experimentation. The results from the experiment are more reliable when similar or identical results can be obtained from repeating the experiment. Also, when different scientists repeat the experiment and obtain similar or identical results, the scientific idea is more reliable. When the results of the experiment and the scientific idea is peer reviewed by other scientists in the field, the idea is more trustworthy. Scientific ideas that can be more broadly applied and are more consistent with neighboring fields are more likely to be true.

Why is scientific replication important?

Scientific replication can help validate the results of the study. If the same results are obtained through performing the study again using the same methods, the results are more reliable. If the same results are not obtained, the researcher can see their mistakes and modify the experiment. Scientific replication can reduce bias of the researcher. Also, the original results may have happened through chance. But repeating the experiments helps the scientist be confident that their results were not due to chance alone.

How do scientists use evidence to choose between multiple competing ideas?

Scientists can analyze the collected data to rule out ideas that are not supported by the data. If the data supports one idea, then the scientist can choose the idea that the data supports. Based on the observations, data, and evidence, the scientist may choose one of the ideas or revise one of the ideas to match the evidence.

Give an example of a situation where a stratified design would be desirable.

Stratified design is useful when the researcher wants a representative sampling from various stratified groups. An example of a situation where a stratified design would be desirable would be testing a drug that affects men and women differently such as antidepressants. Stratifying the group so that there is an equal distribution of men and women in the treatment and control group will eliminate gender as a confounding variable.

When would a subjective measure be preferable to an objective measure?

Subjective measure could be preferable to an objective measure when the measurement cannot be measured through an instrument. Some measurement cannot be measured objectively because it might vary with each subject. An example would be measuring happiness in people.

What do you learn from subsampling?

Subsampling is drawing smaller samples from the original sample. Subsampling is a type of replication. Subsampling can show consistency of results and check for precision of the instruments used in the experiment.

How is knowledge obtained outside of science?

Tenacity/tradition - People accept something as fact simply because it has been around for a long time. There is no demand to check on the accuracy of the ideas. Intuition - People believe something because it feels right. A hunch or gut feeling about something. Can be based on unexamined experiences or even divine revelation. Authority - People accept ideas because some respected authority asserts that the ideas are true

Critique the attached graph. Describe at least two things that could be improved.

The 3D format of the graph make it very hard to read. The colors of the different variables are hard to distinguish. The graph needs axes titles that contain units. Graphs do not need titles.

Identify the variables in the experiment below. Pesticide exposure has repeatedly been associated with cancers, although the molecular mechanisms behind this association are largely undetermined. Abnormal DNA methylation plays a key role in the process of some disease. However, little was known about the effect of pesticides on DNA methylation in the common carp. In this study, we investigated the mRNA levels of DNA methyltransferases (DNMTs) and methyl-CpG-binding protein DNA-binding domain protein 2 (MBD2) as well as the DNA methylation levels in the liver, kidney and gill of the common carp (Cyprinus carpio L.) after 40-d exposure to atrazine (ATR) and chlorpyrifos (CPF) alone or in combination, and a 40-d recovery period. Juvenile common carp were exposed to various concentrations of ATR (at concentrations of 4.28, 42.8 and 428μg/L), CPF (1.16, 11.6 and 116μg/L), and an ATR/CPF mixture (at concentrations of 1.13, 11.3 and 113μg/L). The results revealed that the levels of genomic DNA methylation decreased in all tissues after 40d of exposure to ATR and CPF either individually or in combination. Moreover, the mRNA expression of DNMTs was down-regulated in all treatment groups. In contrast, the mRNA expression of MBD2 was up-regulated. These results demonstrated that long-term exposure to ATR, CPF and ATR/CPF mixtures could disrupt genomic DNA. It might imply that DNA methylation is involved in the toxicity caused by ATR and CPF in the common carp.

The independent variable is the 40 day exposure to atrazine (ATR) and chlorpyrifos (CPF) alone or in combination. The dependent variable is the mRNA levels of DNA methyltransferases (DNMTs) and methyl-CpG-binding protein DNA-binding domain protein 2 (MBD2) as well as the DNA methylation levels in the liver, kidney and gill of the common carp. Some confounding variables include the specific species of the carp, the size of the carp, the age of the carp, and the temperature during the 40 day exposure.

Identify the variables in the experiment below; all the variables. "This study determines the effect of atrazine and fenitrothion no-observed-effect-levels (NOEL) on the binding of corticosterone (B) to corticosterone-binding-globulin (CBG) in an amphibian and a mammal. Plasma from five cane toads and five Wistar rats was exposed to atrazine and fenitrothion at the NOEL approved for Australian fresh water residues and by the world health organization (WHO). The concentration required to displace 50% (IC50) of B binding to CBG was determined by a competitive microdialysis protein assay. Competition studies showed that both atrazine and fenitrothion at NOEL are able to compete with B for CBG binding sites in toad and rat plasma. The IC50 levels for atrazine in toads and rats were 0.004nmol/L and 0.09nmol/L respectively. In the case of fenitrothion the IC50 level found in toads was 0.007nmol/L, and 0.025nmol/L in rats. Plasma dilution curves showed parallelism with the curve of B, demonstrating that these agro-chemicals are competitively inhibiting binding to CBG. The displacement of B by atrazine and fenitrothion would affect the total:free ratio of B and consequently disrupt the normal stress response. This is the first time that the potential disruptive effect of atrazine and fenitrothion on B-CBG interaction at the NOELs has been demonstrated in amphibian and mammalian models."

The independent variable is the concentration of atrazine and fenitrothion no-observed-effect-levels (NOEL). The dependent variable is the binding of corticosterone (B) to corticosterone-binding-globulin (CBG) in an amphibian and a mammal. The type of subject used in the study (species of amphibian and mammal) can be a confounding variable. Other confounding variables can include the number of amphibians and the number of mammals tested, the consistency of size of the test subjects, the environment the test subjects were kept, the type of plasma used, the way the competitive microdialysis protein assay was performed, and using the same concentration of atrazine and fenitrothion no-observed-effect-levels (NOEL).

Distinguish between the independent and the dependent variable in an experiment. Give an example.

The independent variables are factors that may cause a change in the dependent variable. The independent variable is the part of the experiment that is often manipulated by the researcher. For example, in an experiment to test the difference in mating behavior in male guppies after different amount of prenatal exposure to low doses of atrazine, the independent variable is the amount of prenatal exposure to low doses of atrazine. The dependent variable is the variable of interest in the experiment. This is the variable that will change as a result of the experiment. The dependent variable is dependent upon the independent variable. For the example used above, the dependent variable would be the mating behavior in male guppies.

What are some past technologies that have facilitated large scientific advances?

The invention of the microscope allowed scientists to see small objects that cannot be seen with the unaided eye. The invention of the polymerase chain reaction by Kary Mullis allowed scientists to make millions of copies of DNA for further research. The invention of gel electrophoresis allowed scientists to separate charged molecules such as DNA, RNA, and proteins. X-ray crystallography used by Rosalind Franklin allowed for the formation of the DNA model.

Critique the attached graph. Describe at least two things that could be improved.

The pie chart does not need to be in 3D. The 3D aspect of the graph makes it harder to read and understand. The figure caption should be below the graph instead of beside the graph. The title is not needed in a graph. The percentages are too far from the pie chart slices. The percentages should either be closer to the pie chart or just be included in the figure caption.

I wanted to study the mating behavior of a species of rock iguana in their natural habitat. What are variables I would need to control for?

The researcher should control for the temperature, the length of the experiment, the specific species of iguanas, the mating season, the location of the iguanas, the size of the iguanas, the age of the iguanas, and the number of males and females.

Discuss what determines the number of times an experiment should be replicated?

The sample size can help determine the number of times an experiment should be replicated. Sample sizes can be constrained by practicality, and studies often have smaller sample sizes than theoretically desirable. Replicating a study using a larger sample size will increase the power. Power is how likely you are to detect a significant difference if one exists. However, if the sample increases, the effect size will decrease. The effect size is how meaningful the difference is. A balance of power and effect size can affect the number of times an experiment should be replicated.

What are strengths and weaknesses to having a high level of constraint?

The strength of high constraint is that the researcher can have high confidence that the results are not confounded. High constraint can help establish cause and effect. Also, the researcher can obtain very specific findings. The weakness of high constraint is that the results may be less generalizable to natural conditions.

Summarize the general principles to consider when constructing a bivariate graph?

There should be two variables. The independent variable should go on the x-axis, and the dependent variable should go on the y-axis. Don't label every tick mark. The lettering should be large enough so that when the graph is reduced, it is still readable. Make sure the data points are large enough. If more than one symbol is used on a graph, make sure to put the figure legend within the graph. Keep the symbols simple and easy to distinguish between. Adjust the X and Y axes minimal values so they are appropriate for the data. The origin does not have to be zero. Lines can be added to emphasize trends, but lines are not data.

I set up a blind to observe the behavior of lions. What type validity might this threaten? Explain.

This might threaten the internal validity because the methodology of the experiment cannot account for all the potential confounding variables. If the lions are being observed in the wild, the experiment has low constraint.

In a physical performance experiment, the pre-test clues the subjects to respond in a certain way to the experimental treatment that would not be the case if there were no pre-test. What type of threat to validity would this be? Explain

This would threaten the construct validity. The pre-test clues influenced the subjects and may have biased the results. Construct validity looks at whether the methodology actually addresses the scientific question. However, the pre-test clues altered the methodology and potentially the results, so the results may no longer address the original scientific question.

A researcher conducts a survey over time of a certain species of mice at different elevations. She captures ten mice at an elevation of 250', ten at 1000', and ten at 10,000' with all mice were collected on the same day. This procedure is followed once a month for three different consecutive months. Each mouse was massed three times. She determined the average mass, with variance, of mice captured at each elevation for a single day. She then compared the average mass of mice at each elevation on each of the three months. Then she pooled the data by elevation and compared differences in average/mean mass at each elevation. Why capture ten mice at each site instead of just one? Why mass each mouse three times? What do I learn by doing this? Why calculate the variance of the mean/average? Are there any possible confounding variables? If so, how can I control for them? Are there any threats to validity? How many times was the experiment replicated? Explain and justify your answer. What is the sample size? Was there any subsampling in this investigation? Explain and justify your answer.

a. Capturing one mouse at each site means the sample size of the site is one. A sample size of one is too small. Any outliers or abnormalities will mess up the results. A sample size of ten will reflect the population of mice at that site more accurately. b. Massing each mouse three times helps verify the precision of the instrument used to measure the mass of the mouse. Taking the mass three times also helps reduce experimental error. We can learn about the precision of the equipment and experimental method. c. Variance measures how far a set of numbers are spread out from their average value. A low variance means that the data are close to the mean. A high variance means that the data are far away from the mean. Calculating the variance helps us see how precise our measurements are and if there are any outliers present. d. Gender of the mice can be a confounding variable as the male mice may weigh more than the female mice. The researcher can either just study one gender of the mice or collect an equal number of males and females at each elevation. Also, the habitat of the mice can be a confounding variable because the habitat may affect their diet. If the mice of one elevation live in a wet habitat, the researcher should collect mice that also live in wet habitats at the other elevations. e. There are threats to internal validity because there may be many potential variables. Experiments outside of the laboratory and in a natural setting usually have low constraint because the potential confounding variables may not be accounted for. There may also be threats to the external validity. The mice in this particular setting may have unique characteristics that may make it hard for the results to be generalizable to mice all around the world. f. Internal replication is collecting independent samples as contemporaneously as possible. When the researcher collected ten mice at each elevation, she may have been conducting internal replication. However, she pooled all the data together so the results do not show any replication. Massing the mice three times was subsampling. (External replication is collecting samples over time. The researcher collected data once every month for three months. She externally replicated three times, but once again, the results do not show replication as the data has been pooled together.) g. The sample size is 90. The sample of size per elevation was 10. There were 3 elevations. She repeated the experiment every month for three months. A total of 90 mice were used in the experiment. h. Yes, there was subsampling in this experiment. The researcher weighing the mice three times was subsampling. The researcher replicated by subsampling to ensure the precision of the measurement. (Subsampling is taking variables from the same location and testing from different areas. There is subsampling because for each elevation level, 10 mice were taken from several spots in the same relative area, not all from the same exact spot, but representing the same sample. Because of that, we do get 10 samples from each elevation,)

Classify each type of data. Justify your answer. a. 23.4 m b. 45.6 girls c. Ford d. 24 freshmen, 35 sophomores, 45 juniors, 1 senior e. Seven f. 2 agree, 3 disagree g. 35 km/h h. Class two tornado i. 2 juveniles (2-12 cm), 4 adults (12-20 cm) j. The number of wins and losses during the basketball season

a. Numerical and continuous because it reflects a measurement of length. The measured value has a decimal and have a meaningful unit associated with it. b. Discrete because means of discrete data can have decimals. Also discrete reflects a number obtained by counting, and you can count the number of girls. Also, it is nominal because of the category of girl. c. Nominal because Ford is the name of a category. d. Discrete and ordinal because the numbers are obtained by counting and there are no decimals. Also the the categories of class ranking are ordered and ranked. e. Discrete because it is a whole number. But it could be continuous if it was a measurement like 7 cm. f. Discrete because there are no decimals and only whole numbers. Categorical because there are only two possible answers that are mutually exclusive. g. Derived because it was acquired by mathematically manipulating actual measurements. Continuous because it is a measurement with units. h. Categorical and ranked because it is a category that is ordered. i. Discrete because there are whole counted numbers. Categorical nominal because there are only two possible answers. j. Categorical nominal because there are only two possible answers. Discrete because there can only be whole numbers