MMRM Exam #2 Review Sheet

Surveys

Two major types of surveys (descriptive and analytical)

Advantages and disadvantages

Data collection techniques in surveys (know one advantage and disadvantage of each)

How can we (perhaps) increase response rate?

Obstacles (some) of survey research

CATI system

Placement order of questions (general to specific, sensitive issues at the end, demographic info typically near the end etc.)

Double-barreled questions, filter questions

Content Analysis

Definition of content analysis

Characteristics (objective, systematic, empirical, quantitative)

Manifest vs. latent content

Importance of inter-coder reliability

Codebook and code sheets

Composite week

Purposes of content analysis

Unit of analysis in content analysis

Can we make conclusions about media effects based on content analysis?

Experiments

Advantages and disadvantages

Typical steps that a laboratory experimenter takes

Problem of confounding variables

Importance of randomization

Experimental designs—pretest-posttest-control group design

Solomon four-group design (pretest-treatment-posttest; pretest-posttest; treatment-posttest; posttest only)

Validity and reliability in experiments

Double-blind experiments

Qualitative Research 

Four criteria used to evaluate qualitative research (article posted on blog):

naturalistic observation

contextualization

maximized comparisons

sensitized concepts

Positivist Paradigm vs. Interpretive Paradigm. Which is associated with Quantitative Techniques? Which is associated with Qualitative Techniques?

Major types of qualitative data collection techniques:

In-depth interviews
Focus Groups
Participant Observation
Case Studies

Understanding "Sense-Making"

Putting together the qualitative report (what are the steps?)

Make sure you know the following:

NOM IV + NOM DV = chi-square
NOM IV + I/R DV = t-test/ANOVA
I/R IV + I/R DV = correlation

Statistics

Definition

Central tendency vs. dispersion

Mean, mode, median

Frequencies

Type I vs. Type II error and the “null hypothesis”

Test-statistics—

Know when to use, how to solve, and how to interpret chi-square

Know when to use, how to solve, and how to interpret cross-tabulation

Know when to use, how to solve, and how to interpret t-test 

Know when to use and how to interpret correlation

Degrees of freedom

The exam will feature at least one one chi-square problem, one cross-tabulation problem, one t-test problem, and one correlation interpretation problem. 

There will also be a few questions about data interpretation. Specifically, you'll have see if a hypothesis is supported or not supported based on p < .05.

Practice problems:

Practice Statistical questions:

1. Chi-square. 

Ithaca 

School Year------------ Observed Freq. -----------Expected Freq.

Freshmen ------------------------15---------------------------- 27

Sophomores ---------------------20---------------------------- 35

Juniors ---------------------------10----------------------------- 20

Seniors ----------------------------15---------------------------- 25

Where o = observed frequency; e = expected frequency.

The table above provides the expected and observed frequencies of IC students who drop out of school during any given year. The admissions department would like to know if their retention efforts are making a difference.

Using the chi-square test, please tell me if there is a significant difference between the observed and expected frequencies (at the .05 level).

Are the retention efforts working? Why or why not?

2. Cross-tabulation.

I’m testing the following hypothesis:

Men are more likely than women to prefer TV sitcoms to TV dramas.

After collecting my data, I’m left with the following cross-tab:

	Comedy TV	Drama TV	Total:
Male	40 ( )	42 ( )	82
Female	29 ( )	57 ( )	86
Total:	69	99	168

Using chi-square, tell me whether or not the data support my hypothesis (at the .05 level). What use is this data to the ad agency representing Schick Quattro for Men (shaving products—face razors)?

T-Test

1. The following are data of TV use per two weeks by gender. Using t-test (independent samples), determine the statistical significance with probability .05 between the two groups. Are these groups statistically different or not? Why?

Where the denominator is the difference between the standard error of the mean for each group, and X is the average/mean for each group.

Gender -------------------Male------------------- Female

Mean --------------------41 hour --------------- 56 hours

Participants ---------------10---------------------------20

Standard error of mean 2.01 -----------------------0.58

What is the t-value? What conclusions can you make?

Correlation

OK, so here’s the deal. I’m a TV news investigative reporter for Newswatch 16 and I’ve got a tip that a local grocery store is knowingly selling kid yogurt that contains unsafe levels of bacteria. The thing is, I’m, uh, “allergic” to numbers and I can’t make heads or tails out of this information. The tipster, a food safety scientist from Cornell, gave me the following info from his random survey of children who consumed the tainted yogurt, but I have no idea what it means. Can you help me? Do I have a story here? What do all these numbers mean? (5 points) 

Correlation table of unsafe levels of bacteria in yogurt to intestinal illness among children 0-14:

Children ages 0 thru 2                        .38

Children ages 3 thru 5                        .17

Children ages 6 thru 8                        .66

Children ages 9 thru 11                      .22*

Children ages 12 thru 14                    .04*