POL 51 Midterm Exam FORM 1

Name:______

Directions: Please turn in this signedexam along with your blue book. Write all answers in your blue book including answers to multiple choice. On the front cover of your blue book, please write the number of the form of the exam you are taking (1,2,3,4, or 5). Cell phones are not permitted; use of them will result in a failing grade. Cheating is not permitted: do not try this (the person sitting next to you has a different exam than you do). Do not tear pages out of your bluebook. If you have pages torn out upon turning in the exam, this will count as a failing exam. If you have pages torn out in advance of the exam, notify me IMMEDIATELY. You have 1 hour and 20 minutes to take this exam. At 11:50 A.M., I will end the exam and immediately collect all exams. The only exception are for students who have obtained PRE-APPROVAL for extra exam time. There are 90 points possible on this exam.

Multiple Choice (1 point each):

  1. In examining aggregate data, a researcher finds a positive correlation between literacy rates in a state and the number of immigrants in the state. The researcher concludes immigrants are therefore more literate than non-foreign born residents. This conclusion best exemplifies:
  2. Simpson’s Paradox
  3. The Ecological Fallacy
  4. Benford’s Law
  5. The Hawthorne Effect
  1. A researcher wants to study the relationship between expenditures on public education and property values. She randomly samples 100 U.S. cities and then within each city, collects data on school district funding and median housing price in each school district. In this study, what constitutes the unit of analysis?
  2. school districts
  3. cities
  4. census tracts
  5. neighborhoods
  1. In the previous question (right above this one), what is the sampling frame?
  2. school districts
  3. cities
  4. census tracts
  5. neighborhoods
  1. A researcher believes that individuals who were raised (socialized) in Democratic or Republican households (i.e. parents were either Democratic or Republican) are positively influenced to affiliate with the party of their parents. In turn, the researcher hypothesizes that an individual’s party affiliation (Dem. or Rep.) influences their evaluation of political figures, like President George Bush. In this setting, party affiliation of the parents is best described as:
  2. an intervening variable
  3. a dependent variable
  4. an antecedent variable
  5. a spurious variable
  1. A researcher conducts a content analysis of newspaper articles over some time frame. He hires two students, A and B, to do the content coding. After the project is over, the researcher discovers, to his dismay, there is considerable dissimilarity between student A’s coding and student B’s coding (say, the correlation is low, .10). The researcher is dismayed because of issues pertaining to:
  1. The Hawthorne Effect
  2. Simpson’s Paradox
  3. The ethics of hiring student workers.
  4. Intercoder Reliability
  1. The following hypothesis is made: the probability that an individual will vote is positively related to the individual’s level of civic engagement: the more engaged they are with politics, the more likely it is they will vote. This hypothesis best exemplifies
  2. a one-sided (directional) hypothesis.
  3. a two-sided (non-directional) hypothesis
  4. Not enough information to tell.
  5. ecological inference
  1. In the previous question (the right above this one), which variable constitutes the dependent variable and which constitutes the independent variable?

Dependent variable is: ______

Independent variable is: ______

  1. The following statement is read to the survey respondent followed by a question: “POL 51 is the best class ever. Do you strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree with this statement?” This survey question is best described as a
  2. Guttman Scale
  3. Ratio-level Scale
  4. Likert Scale
  5. Open-ended Response
  1. The Stanford Prison experiment was problematic for which of the following reasons:
  2. The lead researcher may have influenced the results.
  3. Participants were not informed of possible mental/physical harm.
  4. The nature of the design makes it hard to replicate.
  5. The research question was unimportant.

Some combination of a,b,c,d (Write letters in your blue book if you think two or more of a,b,c,d are correct.)

  1. E(P)=P implies the following:
  2. over repeated samples, the average estimate of P will equal the population parameter P.
  3. the estimated value of P always equals the population parameter P.
  4. the variance in P will be small.
  5. All of the above.
  1. Which of the following is not true about simple random sampling?
  2. Every element in the population has the same chance of entering the sample.
  3. The larger the sample, the more likely it will be representative of the population.
  4. Any statistics estimated from the sample will equal the true population value.
  5. There is still some chance the sample will not be representative of the population.
  1. The following question is asked on a survey: “Do you agree or disagree with Candidate X’s willingness to tolerate terrorism in our country?” The kind of survey that most likely had this type of question on it is best described as:
  2. a nonprobability sample.
  3. a push poll.
  4. an open-ended poll.
  5. a “person-on-the-street” poll.
  1. A “response set” is best described as:
  2. the actual set of survey respondents sampled from a population.
  3. the sampling frame from which respondents are sampled.
  4. the pattern of responding to a series of questions in a similar fashion without carefully reading the questions.
  5. the proportion of respondents selected for participation in a survey who actually participate.
  1. The “Hawthorne Effect” found by Granberg and Holmberg in their study was:
  2. People were more likely to say they voted when they did not vote.
  3. People gave socially desirable answers to questions about racial attitudes.
  4. People more likely to respond if they thought they were administering a shock to a subject in an experiment.
  5. People who were asked about voting were in turn, more likely to actually vote in an election.
  1. A sampling distribution is best described as:
  2. The degree to which the distribution of sample elements are representative of the population.
  3. The distribution of statistics from repeated samples of size n from the same population.
  4. The number of cases that end up in the final sample.
  5. The confidence interval around a statistic.

Short Answer:

1. Suppose a researcher were to take 1000 samples of size 500, 1000, and 5000 and estimate a statistic from each sample? Assume the samples are random and are drawn from a population of 1 million. Call the statistic “res” and let it correspond to the sample proportion. Below are three histograms plotting the estimates of “res.”

Sample Size=500Sample Size=1000Sample Size=5000

Describe the relationship between sample size and precision in estimating the sample proportion. In your answer, please describe the important features of the graphs given above. Your answer should be no more than 1-2 paragraphs. (15 points)

2. A researcher is interested in estimating the proportion of all UC-Davis students who would support a $100 fee in order to raise money for a new student union. The researcher devises the following question: “There is a proposal to increase fees by $100 in order to raise funds for a new student union on the UC-Davis campus. Do you support or oppose a proposal to increase fees by $100 in order to raise money for a new student union?” Respondents can answer “yes, I support” or “no, I do not support.”

In order to collect her data, the researcher assigns a random number to every available space in UC-Davis residence halls. There are 4,500 residence hall spaces. She then randomly selects N=100 students from the 4,500 spaces and asks them the survey question from before (and collects other data as well which we won’t worry about).

Based on her sampling technique, she finds the estimated proportion of students who support the fee to be P=.50. She then reports that “50 percent of all UC-Davis students would support a fee increase.” Based on this information, please answer the following questions.

  1. Is the population of interest equivalent to the population from which the researcher is sampling? Yes or no, and why or why not? (12 points)
  2. What are the units of analysis in this design? (2 points)
  3. What, if any, problems do you see with the researcher’s conclusion that “50 percent of all UC-Davis students would support a fee increase.” Please be specific. (13points)
  4. What is the level of measurement of the survey question? (2 points)
  5. In the way she asks the survey question, do you see any potential problems? (5 points)
  6. Knowing the standard error of a sample proportion is, what is the standard error around her sample proportion of .5? (3 points)
  7. Knowing the “68 percent” confidence interval is given by P plus or minus the standard error, what is the confidence interval. (2 points)
  8. What is the interpretation of this confidence interval? (9 points)

Definitional Questions. Please give a short definition (usually 1-3 sentences should suffice) of the following terms/concepts (4 points each):

  1. Simpson’s Paradox
  2. Validity
  3. Standard Error