Lesson: 2.1 Data and Decision Making

The data collection method is related to the nature of the problem to be solved...

...and the ethical and practical constraints of collecting data in some particular environment.

In a statistical analysis, it is usually...

...not possible to recover from poorly measured concepts or badly collected measurements

The data described are often collected to satisfy state and government...

...regulation as well as for business purposes, such as to examine trends in a particular stock price or interest rates. These data values are not the result of a designed experiment.

Explanatory Variable

A statistics instructor wants to determine if the amount of time students spend studying for the final exam has any influence on their final exam grade. The amount of time students spend studying for the final exam would be considered the ______.

Double Blind

An experiment in which both nurses and participants do not know whether the participant will receive a placebo is ____

Observational

Determine whether an observational or experimental study is appropriate to address the following statement. A deli cart entrepreneur wants to find out if her customer base is shrinking.

Determine whether an observational or experimental study is appropriate to address the following statement. A car wash operator wants to find out if dressing an employee in a gorilla costume will bring in more customers.

Experimental

Treatment Group

Select the term which best describes the following scenario. A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug.

completely randomized design

The term that best describes an experimental design in which experimental units are randomly assigned to two or more different treatments is ___.

Randomization

What method is used to control for bias?

Control Group

is the group of subjects that does not receive the treatment

Experimental Group

is the group of subjects that receives the treatment

Response Variable

is the variable of interest in an experiment.

Without measurement...

... standards would not be possible. Without complex measurements and standards, just about all of the conveniences that we take for granted (telephones, automobiles, refrigerators, televisions, and computers) would not exist.

The main objective of any business is...

... to earn a profit and consequently companies depend heavily on measurement, data, and statistical thinking.

The response variable is measured in the control group at the beginning of the study, and then a treatment is applied...

... to the control group. After the treatment is applied, the control group becomes the experimental group. The response variable is again measured after the treatment has been applied. If the treatment affects the response variable then there should be a difference between the value of the response variable for the control and experimental groups, presumably caused by the treatment.

Suppose a finance instructor wants to determine if there is any beneficial effect of studying statistics, particularly regression, before taking finance. The instructor obtains records of her finance students....

....and compares the group of students that had statistics or are taking it concurrently with those who have not. The average finance grade for those having had statistics is a great deal higher than those that had not. The conclusion reached from these data is: The study of statistics improves one's understanding of finance.

Collecting data is...

...a natural part of our lives. For example, consider the everyday question, What will I have for dinner? Although virtually no one formally applies the scientific method to such a problem, most people perform experiments and collect data (by eating). This leads to generalizations such as I hate asparagus or I like ice cream. After sufficient experimentation, these generalizations become personal preference laws.

Unless the data are gathered with a...

...controlled experiment, it may not be possible to untangle the effects of the causal factors.

For example, scientists cannot...

...exclude data in an analysis for the purpose of validating their theory or viewpoint.

A large part of using statistics to make good business decisions...

...is developing an ability to appraise the quality of measurements. For many problems, what you measure and how you measure it is more important than how you analyze the data. Thus, it is not surprising that the science of statistics is just as concerned with producing good data as it is with interpreting it.

When confronted with statistical evidence of any kind, regardless of whether or not the statistical analysis is done in good faith, it...

...is ultimately up to you to ask reasonable questions about the data.

Fuzzy concept definitions produce fuzzy measurements. One could devote an enormous amount of time developing measuring instruments for fuzzy concepts such as...

...love, rivalry, and prejudice, and still have a poor measurement. These concepts are fuzzy because they are perceptions. No person can be sure that their perception of love, rivalry, or prejudice is the same as someone else's. Science that relies on fuzzy measurements usually makes the assumption that everyone's interpretation of the concept is more or less alike.

There is no experimentation to see how...

...manipulating one variable will affect another variable or variables. Virtually all of the data we routinely encounter are observational.

The accounting system is designed to...

...measure profitability and to inform management of potential problems. The inventory system is a measurement tool designed to measure the status of inventory, indicate when orders should be placed, and spot potential inventory theft or shortfall. The cash flow system is a forecasting system that measures the company's need for cash.

Because virtually all surveys produce...

...observational data, survey research belongs to steps one and two of the scientific method.

The second step in the decision-making method...

...suggests defining objectives and developing criteria in order to evaluate various alternative decisions. Not all statistics are simple means, proportions, or standard deviations. Managers and researchers often develop their own statistical measures for summarizing some aspect of a phenomenon.

The Consumer Price Index (CPI) is a summary statistic...

...that describes the overall price level in the United States. This statistic is an economic measure of inflation and is used in labor contracts to escalate wages as well as to calculate cost of living increases in social security payments.

Measurement is the first step in understanding....

...the "reality" of any circumstance you wish to change in a predictable way. Thus, measurement is a fundamentally important link in controlling our environment. That is why measurement is so pervasive in our culture. We seem to want to measure just about everything in the physical world: temperature, weight, distance, pressure, hardness (Mohs scale of hardness), wind speed (Beaufort wind scale), earthquakes (Richter scale), and so on. We even try to measure feelings, like love.

The conclusions suggested by statistics can be no stronger than the quality of the measurements which produced...

...the statistical evidence. Fuzzy or confounded measurements must produce fragile conclusions.

What should be measured depends on the problem...

...to be solved. Sometimes what should be measured is obvious and relatively easy. If the problem is to maintain or improve a system, key variables are monitored and decisions are made based on the level of these variables.

Statistics and data are fundamental...

...to the scientific method. Data from carefully designed experiments are the ultimate evidence that support or discredit new theories.

Graphical and numerical summaries are frequently useful in discovering the existence of a problem as....

...well as in shedding light on what some of the potential causes may be. In many instances, problems are caused by systems that do not operate as they are designed. Collecting data (finding out system "reality") and using simple statistical tools to monitor a system are the most common ways of ensuring a system performs properly.

Your career will essentially be a choice of the kinds of problems...

...you desire to solve. The more difficult the problems you decide to solve, the more you will depend upon data and measurement to solve them.

A technology company decided to change the amount spent on marketing on the two social media sites, Instagram and Twitter, to determine the effect on their social shares by users. Their original strategy was to split the budget 50/50 and market to both media sites equally. They switched their strategy to only spend 25% of their marketing budget on Instagram and 75% on Twitter. After switching to the new strategy and collecting data for two months, they found that the social shares increased with the new marketing strategy. Step 1 of 3 : Is this an observational study or a controlled experiment? Step 2 of 3 : Identify the response and explanatory variables. Step 3 of 3 : Identify the control group and the experimental group.

1.) Controlled Experiment 2.) Response Variables: Social Shares Explanatory Variables: Budget Strategy 3.) Control Group: Original Budget Strategy (50/50) Treatment Group: New Budget Strategy (25/75)

Problems are not always the result of a diagnosis of some complaint or system malfunction.

A "problem" may well be the result of an inquiry into unexpectedly good system performance. That is, oftentimes, improving a process does not imply that something was wrong with the process. In this context, a problem presents itself as an opportunity for improvement. Suppose, for example, a finance instructor develops a new method for teaching introductory finance. Measurements are kept on his students as well as students taking subsequent finance courses. Using statistical methods to compare students using the new method to those using the old method can be valuable in pointing out potential educational improvements that may be used at other institutions. As another example, consider the development of the Intel chip. Improving the speed of its processors/chipsets is a constant goal for research and development, regardless of the current speed of the chip. It is not a matter of the chips being too slow, but as technology and hardware advance, the chipsets need to "keep up" so that computers using the chips maintain (or improve) their processing speeds.

explanatory variable

A college administrator wants to know if the SAT scores of incoming freshmen have an effect on their end of year freshman GPA. The SAT scores would be considered the ___.

Confounding Variables (CV)

A company did a study on the physical activity level and weight gain of its employees. Age was not considered so it might be an example of a(n) ____

response variable

A farmer experiments with different amounts of fertilizer in a cornfield to determine which amount results in higher yields. Corn yield would be considered the ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ A farmer experiments with different amounts of fertilizer in a cornfield to determine which amount results in higher yields. Corn yield would be considered the ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

explanatory variable

A farmer experiments with different types of soil in a tomato field to determine which type results in higher yields. The type of soil would be considered the ______.

explanatory variable

A farmer wants to know if there is a relationship between the frequency of irrigation and the yield of his new variety of cucumbers. The frequency of irrigation would be considered the ______.

Bias

A horticulturist consistently overestimates the growth of the plants he uses in his experiments. This is an example of _____.

Placebo

A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug, another group is given a sugar pill. What can we call the sugar pill?

Bias

A quality control technician, using a set of calipers, tends to overestimate the length of the bolts produced from the machines. This is an example of _____.

response variable

A researcher wants to know if there is a relationship between the amount of an anti-nausea drug given and the time for the patient to feel better. The time for the patient to feel better would be considered the ____.

Bias

A sensor consistently overestimates the temperature in the room. This is a good example of ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Response Variables

A statistics instructor wants to determine if the amount of time students spend studying for the final exam has any influence on their final exam grade. The final exam grade would be considered the _______.

Double Blind Study

A study in which the subjects are not told whether they are members of the experimental group or the control group and the evaluators are also not told whether their subjects are members of the experimental or control group

Procedure: The Decision-Making Method

Clearly define the problem and any influential variables. Decide upon objectives and decision criteria for choosing a solution. Create alternative solutions. Compare alternatives using the criteria established in the second step. Implement the chosen alternative. Check the results to make sure the desired results are achieved. Notice the first step in the decision-making method is to define the problem. This is important because almost any solution to the right problem is better than the best solution to the wrong problem.

observational data

Data that we routinely encounter and do not involve experimentation are called ____. A distinguishable fact about observational data is that the data are collected without any influence or interaction with the experimenter.

There is a vast difference in measuring the height of a person and measuring his or her intelligence.

Defining height is relatively simple. It is nothing more than the vertical length of an object. There are standard scales, such as inches and feet, that everyone agrees upon, that can be used to measure height. The National Institute of Standards and Technology maintains rods which define a government standard for distance measures (feet and inches). Because these standards are widely accepted, if ten different people measure a person's height, there should not be large differences in their measurements. This is not true when measuring intelligence. The National Institute of Standards and Technology does not have a measuring rod for intelligence. It is unlikely that ten randomly selected people could agree on a definition of intelligence, much less on how it should be measured. Intelligence is a fuzzy concept because there is no universally accepted definition and hence there can be no universally accepted standard of measure. If a concept cannot be precisely defined, it cannot be precisely measured. How do you measure fuzzy concepts?

Observational

Determine whether an observational or experimental study is appropriate to address the following statement. A gas station owner wants to find out whether customers are satisfied with the service they receive.

Determine whether an observational or experimental study is appropriate to address the following statement. A coffee shop manager wants to find out if providing a discount for local residents will generate additional revenue.

Experimental

Determine whether an observational or experimental study is appropriate to address the following statement. A coffee shop manager wants to find out if using a new cleaning product will reduce staffing costs.

Experimental

confounding variable

Fill in the blank with the most appropriate term. A company did a study on the physical activity level and weight gain of its employees. Age was not considered so it might be an example of a(n) _____.

confounding variable

Fill in the blank with the most appropriate term. Factors that influence the results of a study that the researcher did not, or could not, account for are known as ___?

Procedure: The Scientific Method

Gather information about the phenomenon being studied. On the basis of the data, formulate a preliminary generalization or hypothesis. Collect further data to test the hypothesis. If the data and other subsequent experiments support the hypothesis, it becomes a law. If the experiment does not support the hypothesis, construct a new hypothesis and start the method over again. Sometimes, even if the hypothesis is supported, you may want to test it again but in a different way.

Steps to Consider When Conducting a Survey

Have specific goals. Consider alternatives for collecting data. Select samples to represent the population. Match question wording to the concepts being measured. Pretest questionnaires. Construct quality checks. Use statistical analysis and reporting techniques. Disclose all methods used to conduct the survey.

Specify which of the following variables are not well-defined. Select all that apply.

Hot and Attractive

Double-Blind

In a _____ experiment, neither the subject nor those conducting the experiment know whether the subject is in the control group or the treatment group.

Procedure: Do we have good measurements?

Is the concept under study adequately reflected by the proposed measurements? Are the data measured accurately? Is there a sufficient quantity of the data to draw a reasonable conclusion?

Scientific Method

Measurement and data are an integral part of science. Methods for exploring research problems have been developed over a long period of time and have become standards in the scientific community.

Determine whether an observational or experimental study is appropriate to address the following statement. A car wash operator wants to find out the average age of his clientele.

Observational

Determine whether an observational or experimental study is appropriate to address the following statement. A deli cart entrepreneur wants to identify the number of cars passing a billboard in a day.

Observational

Determine whether an observational or experimental study is appropriate to address the following statement. A gift shop owner wants to study his average annual profits.

Observational

For this example of a completely randomized design:

Response Variable: Students' grades in the finance course. Explanatory Variable: Whether students take statistics before finance. Randomization is often used as a method of controlling bias and is an important principle in the design of experiments.

Single-Blind Experiment

Select the term which best describes the following scenario. A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug, another group is given a placebo. The individuals do not know if they are taking the drug or the placebo.

single-blind experiment

Select the term which best describes the following scenario. A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug, another group is given a placebo. The individuals do not know if they are taking the drug or the placebo.

Confounding Variable

Select the term which best describes the following scenario. A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug. It is not identified prior to the study what is causing their pain, and the source of their pain impacts their response to the drug.

Confounding Variables

Select the term which best describes the following scenario. A study is done an a new drug designed to lower blood pressure. Individuals with high blood pressure are selected for the study and are prescribed the drug. In addition to the drug, some of these individuals start exercising regularly, which was not part of their instructions for the study. This has some impact on their blood pressure. What term can we assign to the effect of exercising regularly in this study?

Double-Blind Experiment

Select the term which best describes the following scenario. A study is done an a new drug designed to lower blood pressure. Individuals with high blood pressure are selected for the study and some are given pills containing the drugs while others are given a placebo. The participants and the researchers tracking the participants blood pressure are not told whether they are taking the drug or not.

Placebo Effect

Select the term which best describes the following scenario. A study is done an a new drug designed to lower blood pressure. Individuals with high blood pressure are selected for the study and some are given pills containing the drugs while others are given pills that do not contain any of the drug. Even the group which did not take the drug sees an improvement in their blood pressure. What could be causing this?

Placebo

Select the term which best describes the following scenario. A study is done an a new drug designed to lower blood pressure. Individuals with high blood pressure are selected for the study and some are given pills containing the drugs while others are given pills that do not contain any of the drug. What can we call the pills that do not contain the drug?

Placebo

Select the term which best describes the following scenario. Researchers are investigating a drug to help reduce gastrointestinal issues in individuals with a gluten sensitivity. There is a group that is given pills to take they are told will help, but the pills do not contain the drug. It is studied whether their symptoms are relieved. What can we call the pills this group was given?

Control Group

Select the term which best describes the following scenario. Researchers are investigating a gluten-free diet. There is a group of people who they will monitor that they have told to continue eating normally.

confounding variable

Select the term which best describes the following scenario. Researchers are investigating a gluten-free diet. There is a group of people who they will monitor that they have told to eliminate any gluten from their diet. In addition to the lack of gluten, there may be some impact of an altered caloric intake which was indirectly caused by eliminating gluten. What can we call this impact?

Placebo

Select the term which best describes the following scenario.A new painkiller is studied. A group of people with chronic pain participating in the study are administered the drug, another group is given a sugar pill. What can we call the sugar pill?

Inferential Statistics

The branch of statistics that develops theories to test the hypothesis using data collected from an experiment and makes formal conclusions about a population or parameter.

Descriptive Statistics

The branch of statistics that focuses on exploratory methods for examining data that help to formulate a hypothesis.

The data collection process in steps one and three of the scientific method can be quite different.

The first step of the scientific method is exploratory, finding out what "reality" is about the subject under consideration. Since the data in this phase need not produce convincing evidence, whatever data are available are used to generate ideas. However, the third step begins the validation of a hypothesis. Scientists are trained to be critical thinkers. If a new idea is to be accepted by the scientific community, convincing evidence must be developed at the third stage. The manner in which the data are collected is an important part of that evidence. If the evidence is to be persuasive, a data gathering strategy (an experimental design) that will produce data without the unwelcome influences of confounding variables is required. For example, a manager wants to know if a new marketing plan has increased sales. In the analysis of the data, the manager must ensure that any increase in sales is due to the marketing plan and not other factors such as an increase in the sales force or a reduction in the cost of the product, for example. These "other factors" would be considered confounding variables.

Placebo

The term that best describes a fake treatment that contains none of the drug being tested is ____.

Double Blind Study

The term that best describes a study in which neither the subjects, nor the evaluators are told whether the subjects are members of the experimental group or the control group is ....

Explanatory Variable

The term that best describes a variable that affects the variable of interest in an experiment is ....

explanatory variable

The term that best describes a variable that affects the variable of interest in an experiment is ___.

Explanatory Variable

The term that best describes a variable that affects the variable of interest in an experiment is....

Confounding Variable

The term that best describes a variable that was not controlled or accounted for by the researcher and thus damaged the integrity of the experiment is...

Controlled Experiment

The term that best describes an experiment in which a researcher attempts to control the environment so that the effect of one variable on another can be isolated and measured is ___.

Controlled Experiment

The term that best describes an experiment in which a researcher attempts to control the environment so that the effect of one variable on another can be isolated and measured is...

completely randomized design

The term that best describes an experimental design in which experimental units are randomly assigned to two or more different treatments is .

Controlled Experiment

a researcher attempts to control the environment of the experiment so that the effect of one variable on another can be isolated and measured. The purpose of controlled experiments is to reveal the response of one variable (sales, the response variable) to changes in another variable (amount of price increase, the explanatory variable).

Confounding Variables

are additional variables that were not measured but affected the outcomes in a study. Factors that influence the results of a study that the researcher did not, or could not, account for are known as ___?

Causal Factors

are factors or variables that influence the response variable.

Placebo

is a fake treatment that contains none of the drug being tested.

Simpson's Paradox

is a phenomenon in statistics in which an effect or trend appears in several different groups of data when considered separately but disappears or reverses when the groups are combined. This paradox is often seen with reports based on frequency data when making inferences about relationships between two or more variables.

Explanatory Variable

is a variable that affects the variable of interest (response variable) in an experiment.

Confounding Variable

is a variable that was not controlled or accounted for by the researcher and thus damaged the integrity of the experiment.

completely randomized design

is an experimental design in which experimental units are randomly assigned to two or more different treatments.

Placebo Effects

is the belief that the subject improves (or has a reaction to the placebo) when they haven't received the treatment.

Treatment

is the factor that changes the level of the explanatory variable.

Essentially there are two ways to obtain data:

observation and controlled experiments.