First Exam Study Guide
First Exam Study Guide MGMT 276
Popular in Statistical Inference in Management
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
Popular in Business
This 15 page Study Guide was uploaded by Alex Batali on Thursday February 19, 2015. The Study Guide belongs to MGMT 276 at University of Arizona taught by Delaney in Spring2014. Since its upload, it has received 304 views. For similar materials see Statistical Inference in Management in Business at University of Arizona.
Reviews for First Exam Study Guide
I'm a really bad notetaker and the opportunity to connect with a student who can provide this help is amazing. Thank you so much StudySoup, I will be back!!!
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 02/19/15
Study Guide Management 2 Contrasting the measurement of observable actions with the theoretical constructs associated with those measurements 0 A measurement is what ls being observed and recorded while a construct is what is being measured 0 An operational de nition tells you the de nition of a construct and how it is going to be measured in the context it is in o Validity is how well something measures what it is supposed to measure while reliability is how accurate you are every time High reliability means that you hit the same spot every time while high validity means that you are hitting the quotcorrectquot spot 0 A metaanalysis is a statistical procedure that summarizes a large body of evidence Other little things this chapter 0 Dependent variable is what is being measured while an independent variable is something that is changed to see a change in the dependent variable 1 independent variable univariate data set 2 independent variables bivariate data set 3 independent variables multivariate data set The control group is the group that does not have anything done to them while the experimental group is the group that is getting tested A within participant design is where the same subjects are used for every level of testing in the experiment For example the subjects would have to travel from city to city Between participant design is where there are different groups of subjects that compare against each other 0 Random assignment is randomly assigning participants to certain groups 0 Random sampling is where you pick randomly participants from a population of people 0 A population is the entire group of people while a sample is just a smaller group within that large group A parameter is a characteristic of a population while a statistic is something that describes the population Sample data ls data that comes from the sample Sample mean is the xbar Population mean is the mu sign lnferential statistics is where you make a decision of statement about the population based on studies and or experiments that you have ran Descriptive statistics is where you organize and look at information in a certain way Levels of measurement 0 Nominal level Qualitative No order Color gender jersey number 0000 O o Ordinal Qualitative Order to the variables Class standing quotgoodquot or quotbetterquot type rankings 0 Interval Quantitative Fahrenheit temperature scale Size of dress 0 Ratio Quantitative 0 point Weight height distance number of patients seen number of calls made Surveys 0 O O A goal of a survey is to get a representative idea of what people in a population want A census describes an entire population and comes to a conclusion about everyone while a sample is just a portion of the population and therefore only describes a small part A sample can be representative if there has been random sampling or random assignment It would be biased without these things Sampling techniques 0 Probability sampling is where everyone in the population has an equal chance of being chosen same probability Probability sampling is used for precise measurements Simple random sampling Probability sampling technique This uses random numbers often generated from a computer database Systematic random sampling Probability sampling technique Works perfectly Strati ed random sampling Probability sampling technique Where people from the population are sorted by something as in their major and then the computer database chooses them from there Cluster sampling Probability sampling technique Organizing the population into groups and then randomly selecting them from there Convenience sampling Nonprobability sampling Sampling from people that are closest to you or because it is easy for you to sample them Snowball sampling Nonprobability sampling Where you pick someone to lead and decide on the rest of the participants Someone leads the sampling that person was decided upon by you 0 Judgment sampling Nonprobability Sampling the people based on your own judgments and stuff 0 5 principles of questionnaire construction 0 Make sure items match their research objectives and identify what constructs you are trying to understand be explicit in identifying your constructs o Responders have the answers to our questions Tapping into their attitudes and beliefs You have to think like them and understand them Be clear and concise in your questions 0 Assessment should feel easy and clear unthreatening Should start with the most threatening questions rst Avoid double negatives agree or disagree to a negative question 0 Avoid ambiguity and bias in your items Avoid doublebarreled Avoid leading or loaded questions 0 Consider lots of different formats for the questions Consider open vs closeended questions Bring in other data collection in your questionnaire Pilotfeedback xpilotanalyze xpilot The big 4 for ethics 0 Informed consent 0 Con dentiality o Debrie ng o Deception A contingency table is one that I use to classify observations according to two identi able characteristics 0 A tree diagram is a diagram that is helpful in organizing calculations that involve several stages 0 A Pareto chart is one that has the categories in descending order according to frequency 0 A linear relationship shows correlation because the two variables ow together in a consistent way that creates a straight line while the curvilinear relation has two variables that create a curve when working together 0 Guidelines for constructing frequency diagrams 0 Classes should be mutually exclusive Set of classes should be exhaustive All classes should have equal intervals Selection of number of classes and interval width is multifaced Class width should be round Avoid openended classes OOOOO Cumulative frequency is when you just add up the frequencies from the bottom to the top and your top cumulative frequency number will be the total number of data values Relative frequency is each frequency number divided by the total number of data values while cumulative relative frequency is the cumulative frequencies divided by the total number of data values for each one A frequency histogram is when you have all of the frequencies stacked on top of each other in a histogram diagram A frequency polygon is when you make a line across the frequency histogram that has all of the midpoint of each bar connected 000000 0 O O O O O O O The characteristics of the normal curve Measured on continuous scale Possess clear central tendency Have only 1 peak unimodal Tapering tails Symmetric around the mean Bellshaped 3 measures of central tendency Mode most common score Median middle score 50th percentile means 50 of people are better than you while 50 are worse A quartile is where all of the data values are split up into 4 equal parts Box plots use median quartiles minimum and max scores Mean Average score also the balance point of a distribution Trimmed mean is when maximum and minimum scores are not included in the mean The weighted mean is the one that includes all of the scores Normal distribution meanmedianmode Positively skewed distribution Meangtmediangtmode Negatively skewed distribution Modegtmediangtmean What they are good for O O 0 Mode is good for nominal or ordinal data Median is good for ordinal data Mean is good for interval or ration data Probabilities o 68 that a score will fall between 1 standard deviation of the mean z1 o 95 that a score will fall between 2 standard deviations of the mean z2 o 997 that a score will fall between 3 standard deviations of the mean z3 Variability o The range is the largest score minus the smallest score 0 O Variance is the standard deviation squared Standard deviation is the amount that observations deviate on either side of the mean O O O The standard deviation is each score minus the mean and then squared All of those numbers are added up and divided by N or Nl depending on if it is the entire population or only a sample If we took the average of the deviation scores we would get the standard error The mean deviation is the absolute values of the deviation scores The standard deviation is calculated relative to the mean is saying that every score that is calculated uses the mean 0 The mean is a measure of quotpositionquot means that it lives on one location on the curve 0 The standard deviation is a distance score of the spread of the distribution PLOUS A pseudoopinion is when someone is familiar with the context or something more than what the question is asking Attitudebehavior inconsistency is when someone39s attitude towards a situation or something differs than their behavior toward the situation 0 Response scales are great for people to evaluate their own behavior off of their answers 0 Extreme anchors can affect responses because they can lead some responses astray by being heavy on one side or the other Study Guide Management First test Describe how familiarity with statistical methods can be 0 Associated with power in the workplace Knowing statistical methods can help you to conduct simple research that will allow you to make more informed decisions in the workplace 0 Lead to advancement in the workplace Gives you the ability to work up in the workplace and obtain better jobs than what you 0 Provide a type of literacy that crosses many domains Your quotliteracyquot now includes anything with numbers and graphs It moves past just speaking English every day 0 Help to control for cognitive biases and illusions By understanding the statistical methods and what they are all about a person can know not only the stats but also how to conduct tests to test for certain things 0 Be useful in improving communication computer skill information management technical literacy career advancement and quality assessment By having the ability to describe just about anything with statistical methods a person can really get ahead of the rest of the population in understanding people and things like that 0 Be applied to several domains within business You can use statistical methods to do just about anything inside of an organization Statistics is the quotscience of collecting organizing presenting analyzing and interpreting data to assist in making effective decisions 0 The ratman illustration shows that there are more than 1 way to see the world Almost everything can be seen from a different angle or perception o The blob shows us that we construct a lot of what we see of the world from our preconceptions and what we already knowcan attribute to what we are seeing 0 Attention bias shows us that we can be vulnerable to bias because sometimes we just miss something that other people can sometimes see 0 Awareness of these issues allows us to be able to distinguish and pay closer attention in situations like those Knowing that we miss a lot of things in our everyday lives can help us to pay closer attention to the things around us Contrast the measurement of observable actions with the theoretical constructs associated with those measurements 0 A measurement is what is actually being measured while the construct is the overall overarching concept that is being measured 0 An operational de nition is the de nition of a construct and how it is going to be measured in the context that it is in o Validity is how accurate the test actually is while reliability is how consistent the results are A test with high reliability and low validity would be if someone who is shooting darts at a dartboard consistently shot in one spot but that spot wasn t the center where it needed to be Metaanalysis is a statistical procedure that summarizes a large body of evidence 0 Contrast a dependent variable versus independent variable 0 O O A study with two variables in a correlation would be associated with a bivariate data set A study with two or more variable would be associated with a multivariate data set Control group is the group that is not getting any tests done on them they are the test that the experimental group is based off of The experimental group is the group that is getting tested and will produce the results that distinguish the research Within participant design is where the same subjects are used throughout the entire experiment on every level of the independent variable In other words the subjects would have to travel to another city when comparing two different cities Or the subjects will use both types of beds to see which one is more comfortable Between participant design is where each group only does one job So for the bed example one group would test one bed while the other group would test the other bed Random assignment is randomly assigning people to a certain group in an experiment Random Sampling is picking your participants randomly out of a population Population is the entire group of people while the sample is just the smaller group A parameter is a characteristic of a population while a statistic is something that describes a population The sample mean is represented by XBar The population mean is represented by mu A placebo is the part of an experiment that is not actually real For example a placebo in an experiment that tests how well a pill makes people sleep a placebo would be a pill that had no sleeping aid in it A placebo is trying to see if people experience the same results even from something that isn39t real The placebo effect is what happens when someone takes a placebo and thinks that it is the real thing This causes the person to experience the same effects that the real thing would cause without actually taking the pill A double blind procedure is where the researcher and the experimental group don39t know what is going on This eliminates all bias from the tests A single blind procedure is just where the experimental group doesn39t know what is going on This could lead to researcher bias The experimenter will sometimes be bias if he or she expects something from the experimental group Random assignment is very important in differentiating between quasi and true experiments Genderpolitical af liation is associated with quasiexperimental de gn With no random assignment we are very vulnerable to selection bias and confounds Cause and effect means that one thing causes another to happen In true experiments there can be cause and effect because there is random assignment with both an independent variable and a dependent variable These can be manipulated to show that there really is a cause and effect relationship In a quasiexperiment there is no random assignment and therefore everything that happens can t be deemed as cause and effect Descriptive statistics has to do with organizing and looking at information in a certain way lnferential statistics is taking the information and making a conclusion based on the sample that you used Continuous variables are ones that don39t have a domain and could potentially go on forever Discrete variable can assume any number within the given domain Categorical data is also called qualitative data and assumes any nonnumeric values Numerical data is also called quantitative data and assumes any numeric values 0 Someone might code mae1 and female 2 but it does not make quotgenderquot a quantitative variable 0 A binary variable is a variable that can only take on 2 different states Levels of measurement 0 Nominal No order to the labels Qualitative Color gender jersey number 0 Ordinal Some order to it Qualitative quotgoodquot or quotbetterquot type ratings class standing 0 lnterval Quantitative Fahrenheit temperature scale Size of dress 0 Ratio Quantitative 0 point Weight height number of patients seen number of sales calls made distance A naturalistic observation is where the researcher goes into the experimental groups natural habitat and does not disrupt them while observing them quotnaturallyquot A eld observation comes when the researcher observes what is happening in a more controlled environment something the researcher set UP Time series comparison looks at the same subjects over a period of time to identify trends A crosssectional comparison looks at different groups at a period in time to notice differences between the groups A Likert Scale is something that is used in questionnaires It allows for the participants to choose from a range 15 of how much they agree or disagree with a statement A survey is used to gather information about how people feel about a certain topic or upcoming situation A census is a conclusion based off the entire population It includes everyone A sample is just a portion of the population A population is everyone almost like a census except conclusions aren39t drawn from them A sample is a portion of that population 0 Sample is preferred in experimental situations where it would be extremely unnecessary to look at the entire population A census might be used when everyone NEEDS to be looked at o A sample can be representative if there was random assignment or sampling taken Without this the sample could be biased based on who the researcher chooses The sampling frame are all the people that are chosen that could technically participate in the experiment or whatever they are doing 0 A sampling frame would be different than a target population because the sampling frame is an entire population that is t for the sample The target population are the people that should be included in that sampling frame Larger samples would be preferable to smaller samples because they give you a more representative group of the entire population A higher response rate is better because it gives you more data to base your conclusions off of A response rate is important for the sample to remain unbiased because it if only certain people respond then there is a possibility for the sample to be biased Probability sampling is where every person in the population has an equal chance of being chosen This is used for precise measurements In nonprobability sampling the probability of each person is not known This is quite common and useful in many situations 0 Simple random sampling is using a random numbers table to generate the participants in the study 0 Systematic Random Sampling is regular random sampling that works perfectly o Strati ed random sampling is where the participants are sorted by something as in their major and then a computer program chooses the participants from there 0 Cluster sampling is picking people from something that identi es them such as their major Then you can pick the cluster that you want and go from there O 0 Convenience sampling is sampling people because they are either close to you or because that is the easiest way to choose people Snowball sampling is where you pick a someone at random to lead you and decide on the other people to be chosen Judgment sampling is a nonprobability sampling technique 0 An interview can be described as just about any time that you are questioning someone or researching into someone39s mind 0 A structured interview is one that is set of formally and has a more questionresponse type set up An unstructured one could be when you are just asking someone about their lives Gathering quantitative vs qualitative is relatively easy for someone to do within an interview You can ask questions that give quantitative results by asking about certain number gures However on the opposite side you can ask people about qualitative measures as well Informal interview would probably not be directly face to face If it was then this type of conversation would be looser talking and not so much just questions and answers A highly structured interview would be in a room with someone just face to face with another person These types of interviews are not fun A focus group much like an interview can ask questions that pertain to either qualitative or quantitative data You can use the sampling in the same way as surveys Probability sampling could be used by just randomly sampling people from the population 0 A questionnaire is something that contains questions that are also speci ed for a target audience and to measure one single construct O Questionnaires have very speci c sampling and wording of their items that is all constructed to match with their aims 5 principles of questionnaire construction 0 Make sure items match research objectives and identify what constructs you are trying to understand be explicit in identifying your constructs Responders have the answers to our questions Tapping in to their personal beliefs Understand them have a target audience Have simple short questions Assessment should feel easy and clear Start with questions that are the most quotthreateningquot foot in the door phenomenon Avoid double negatives Agree or disagree quotTeachers shouldn39t have les contact with their parentsquot Avoid ambiguity and bias in your items Avoid quotdoublebarreledquot Avoid leading or loaded questions Consider problem of acquiescence Consider lots of different formats for responses Consider open or closed ended questions Consider complementing your questionnaire for other forms of data collection 0 Pilot Feedback x pilot analyze x 0 A rating scale is any way that the researcher nds a way to analyze the data that is given to him or her an rate the surveys 0 An anchored point is one that has writing or numbers at the end of the points which would mean that a fully anchored rating scale is one that has the ends capped o Ranking scale is a type of question where the responder is asked to rank their opinion on something For example this could be from like quotrank something on a scale of 15quot 0 A semantic differential technique uses a ranking system that also includes having words at the end to describe how the person should grade it o A summated response works by having a scale of 15 that a person can chose from and the researcher adds up all of those points at the end to see how much some quotprefers classical music over rock musicquot 0 Pilot testing a questionnaire can be helpful because you can see how it really does when tested 0 Peer review of a questionnaire can be helpful because you get another person39s opinion on the questionnaire 0 Pros of administering surveys are the responders are required to understand the questions just by reading and surveys are cheaper than interviews Cons would be that some people may not just answer the survey or miss certain vital questions 0 Ways to administer Face to face Mail Internet Telephone Focus groups 0 The big four for ethics Informed consent Con dentiality Deb e ng Decep on OOOO Describing Data Visually Data in a dot plot is stacked on top of each other In other words each person has a single dot that is placed on top of other peoples dots and they stack up 0 A frequency distribution for grouped data has some data that is stacked on top of the other data 0 One for ungrouped is much more spread out and covers a larger distance 0 Guidelines for constructing frequency distributions 0 Classes need to be mutually exclusive Set of classes should be exhaustive All classes should have equal intervals Selection of number of classes and interval width is multifaced Class width should be round 0 Avoid openended questions Cumulative frequency adds up the frequency data from the smallest to the largest numbers Relative frequencies are the percent of how frequent the frequency is and relative cumulative frequency is the percentage of people who did worse than you A frequency histogram is a graphical way to describe the relative frequency Frequency polygon is the midpoint of each bar on a frequency histogram with a line drawn through each one A contingency table is a table that used to classify observations according to two identi able characteristics A tree diagram is a graph that is helpful in organizing calculations that involve several stages 0000 Frequency histograms o Skewed left means that the graph trails off in the left direction while skewed right means that the graph trails off in the right direction 0 Bimodal skewedness is the exact same except bimodal has two peaks 0 The skewedness stays the same no matter what you are looking at o Symmetric graph would look like a normal one A bar chart is just a regular chart that is commonly used A Pareto chart is where the categories are displayed in descending order according to frequency A stacked bar chart is where you would stack multiple people or other colors on top of each other in a bar chart A simple line chart is just a chart that has the line that represents that data These are fairly easy to read and show trends A pie chart is great for showing ratios and other percentage data like that Correlation A correlation is where two variables react with each other but do not cause each other A scatterplot is what shows correlations and how you plot dots as data 0 These can be both strong positive relationships and strong negative relationships 0 There are also such things as a zero patter and also a curvilinear patter The correlational coef cient is the number that describes the correlation This is anything between negative 1 and positive 1 The closer the number is to 1 the higher the correlation whether it is positive or negative Identifying the correlation coef cient is relatively easy 0 Examples of correlations quotheight of mothers and height of daughtersquot A correlation does not imply causation at all There is no cause and effect relationship that comes from correlational studies o A linear relationship is one that contains a line and a curvilinear relationship is one that contains a large curve that shows the correlation between the two variables 0 Guidelines for constructing frequency distributions 0 Classes should be mutually exclusive Set of classes should be exhaustive All classes should have equal intervals Selection of number of classes and interval width is multifaced Class width should be round Avoid openended classes 0 Review how to nd class midpoint and class interval 0 Look at frequencies and relative vs cumulative frequencies OOOOO Distributions 0 Three characteristics of distributions 0 Central tendency measure of location Includes mean median and mode Where are all the data values concentrated 50th percentile means that 50 of people are above you and 50 are below you 0 Dispersion measure of variability Standard deviation variance range mean absolute deviation 0 Shape How we analyze the shape of the graph or distribution Are the data values distributed symmetrically 0 Normal curve 0 Measured on a continuous scale when range is large discrete variables become continuous Holds clear central tendency Has only one peak unimodal Has tapering tails Symmetric around the mean 0 Bellshaped 0 Measures of central tendency and location Mode the most common occurring score 0 Median the middle score in the distribution 50th percentile o A quartile separates a set of observations into 4 equal parts 0 A box plot plots the minimum score maximum score and the rst three quartiles Mean average score also called the balance point A trimmed mean is one that has removed a small proportion of the smallest and largest values in a distribution The weighted mean is one that contains every value and therefore bears the weight of the entire set of data values 0 In a normal distribution meanmedianmode Positively skewed distribution meangtmediangtmode Negatively skewed distribution modegtmediangtmean O 0000 O 0 Mode is good for nominal or ordinal data 0 Median is good for ordinal data 0 Mean is used for interval or ratio data 0 Normal curve 0 The empirical rule for the normal curve is that 68 of scores fall between 1 standard deviation of the mean 95 of scores fall between 2 standard deviations of the mean and 997 of scores fall between 3 standard deviations of the mean 0 Normal distribution 0 A zscore is the number of standard deviations that a set of scores is away from the mean The probability that a score will fall above a z of 0 is 50 The probability that a score will fall between 1 and 1 standard deviation of the mean is 68 21 The probability that a score will fall between 2 and 2 standard deviation of the mean is 95 22 The probability that a score will fall between 3 and 3 standard deviation of the mean is 997 23 0 Measures of variability 0 Range smallest score subtracted from the largest score 0 Variance standard deviation squared 0 Standard deviation typical amount that observations deviate on either side of their mean the data values always add up to O o The formula for standard deviation is the change squared of values Then from there all you do is square root all the numbers and you get the standard deviation 0 A deviation score is the amount of each standard deviation 0 The formula for sample vs population are different in the sense the samples for the population just use the number of observations in the population N while for the sample the bottom of the equation is Nl 0 Standard deviation formula is the one with the square root symbol o If we took the average of the deviation scores we would get the standard error of the mean 0 The mean deviation uses the absolute values of the deviation scores A measure of the average distance between an observation and the mean of the observations o If the standard deviation is calculated relative to the means then that means that the standard deviation is based off of the mean 0 Standard deviation can also be estimated by range6 o The mean is a measure of position means that it only lives on one location of the curve 0 The standard deviation is a distance score of the spread of the distribution Plous Text Chapter 1 selective perception When you expect something to show up you are more inclined to see it as showing up When you think that you are drinking alcohol you tend to feel some of the effects before they really come Sometimes your perception can change the way you think about a game or something Every person has a different perception of a football game for example If we expect a democratic person to win we would think that the media covering a republican could be biased and therefore get mad at what we are seeing Chapter 5 plasticity The order of the way that things are presented de nitely causes a change in responses 0 The example of asking about the US and communism and the order of the questions affects the outcome of the answers Generally pseudoopinions are offered by 25 to 30 percent of responders o This is when someone is familiar with the context or part of the question When they are trying to lter out these opinions researchers are attempting to see if people really think that Consistency is the way that someone answers questions is in a consistent way while plasticity is already having prior knowledge and the attempt to be biased in your responses to questions Attitudes about abstract propositions are often unrelated to attitudes about speci c application of the same propositions Attitudebehavior inconsistency arises when someone has an attitude about something but then their actions are completely different Chapter 6 Wording and Framing Responses to openended questions tend to be lengthier and cause problems because there is a wide variety of answers that could be given People take the opportunity to write to write many weird things that they wouldn39t have with a closeended question Respondents tend to look at rating scales and be able to determine their own level of behavior or response based on the information given The words quotforbidquot and quotallowquot have de nite altercations to the way that people respond to questions A decision frame is a controlled part that causes the respondents to respond in a certain way Researchers can frame the outcome of the answers that they would expect from people Chapter 13 anchoring and adjustment There is an initial starting value that acts as an anchor to the responses that it gets Extreme anchors can affect responses because they can lead some of the responses astray and make them lean heavily on one side or another
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'