Social Science Research Method

Survey designs

- similar in nature to opinion polls: a series of questions are generated, and after collecting responses to those questions, conclusions can be drawn.

How to calculate variance

1. Calculate the mean 2. Calculate deviations (from the mean) of each value. 3. Sum up all the deviations 4. Divide the sum by (number of values - 1)

List the seven key steps in quantitative research.

1. Isolate a problem 2. Select a design 3. Operationalize variables 4. Generate a sample 5. Collect data 6. Analyze data 7. Write results

Here are the ages (in years) of six people who took part in a study: 31, 21, 26, 30, 31, 24. What is the mode of these values?

31

Quantitative - Standard of rigor

Conclusions based on observed connections that are unlikely due to chance or errors on the part of the researcher

Qualitative - Standard of rigor

Conclusions that are reflective and deeply grounded in the words and behaviors of participants

Rhetorical - Standard of rigor

Conclusions that are well reasoned and grounded in a theoretically driven reading of discourses

Give an example for a study in which the data you obtain should be analyzed with the t-test.

Conduct a study to find out whether boys and girls spend time differently using social media

"The more you smoke, the fewer years you can expect to live." This is an example for a __________ bivariate correlation. (Fill in the gap with a word that describes the direction of the relationship.)

Negative

Qualitative - Role of researcher

To be a careful and reflective observer

Quantitative - Role of researcher

To be a precise analyst of data

Rhetorical - Role of researcher

To be a thoughtful and informed critic

Write the names of at least two post-hoc procedures that can be used to follow-up a significant One-Way ANOVA.

Tukey, Scheffe, Dunnet

If a study concludes that there is a relationship between two variables when in fact there is no such relationship, we say that the researchers have made a __________ error. (Fill in the gap.)

Type I

If a study concludes that there is not enough evidence for a relationship between two variables when in fact the relationship does exist, we say that the researchers have made a __________ error. (Fill in the gap.)

Type II

ratio variable

Values that lie along an evenly dispersed range of numbers where there is an absolute zero. Ex: income, Kelvin temperature, weight.

interval variable

Values that lie along an evenly dispersed range of numbers. Ex: Celsius or Fahrenheit temperature. 0 Celsius or Fahrenheit degree does not mean there is no heat.

Suppose you conduct a statistical test to see if you have evidence for a relationship between two variables. You conduct the test at the α = 0.05 significance level and find that p = 0.08. What is the conclusion that can draw from this result?

We cannot reject the null hypothesis since 0.08>0.05

When performing a t-test, the t-statistic that we compute must be compared to a critical value that we can look up in a table. When can we conclude that we have evidence for difference between the two groups?

t-value > critical value

The authors point out the fact that communication research is characterised by "a variety of epistemological perspectives". What does term epistemology mean?

the part of philosophy that is about the study of how we know things (Cambridge dictionary)

Census

the researcher can collect data from each member of the whole population of interest (rather than a subset of the population)

What does a typical ethnographer do? How does he/she do research?

- An ethnographer studies peoples and cultures (minority cultures, subcultures, etc.) with their customs, habits, and mutual differences. - The ethnographer's job is to listen to the talk, watch what happens, see what people do, to write it down, tape it, record what documents can be recorded, and so on. - The ethnographer accesses 'what is going on' in a setting through the mundane competences he or she has developed that routinely make it possible to learn about new cultures and forms of social organisation.

Write at least two advantages of modern Web-based forms over traditional pen-and-paper questionnaires.

- Cost less time and procedures. - Save more money (using electricity instead of pen and paper)

Rhetorical - Example question

- How do audience members/listeners assign meaning to messages? - How do speakers/rhetors create messages and with what effect? - How can evidence be effectively used in persuasion? - How do organizations use metaphors to describe ideal organizational cultures?

What is meant by the requirement that quantitative research must be replicable?

- Replication: other researchers can use similar methods and still get similar results. - Through replication, chance finding posibilities are reduced => more credibility can be attached to the conclusions drawn.

What is the key difference between the rhetorical and the qualitative research strategies?

- a rhetorical scholar might seek a prime example of communication to analyze (for eg: an excellent speech) - a qualitative researcher might be much more interested in studying the more typical and common day-to-day communicative behaviors of people

Similarities between the three strategies

- attempt to raise knowledge about communication practice - promote the generation of theory (trying to make some theoretical statement about communication as a result of the study) - discursive in nature (The decision to collect discourse, to engage in ethnography, or to analyze survey results is partly a question of what type of evidence you want to use in support of your claim(s)

disadvantages of a surveys

- can typically not provide compelling evidence of causation: although survey designs can help identify relationships between variables, they are not well suited to identifying cause-effect relationships among variables - For instance, a survey design might show a statistical relationship between communication ability and school achievement; however,such a design cannot show whether communication ability causes students to get better grades or vice versa

Questions of difference

- explore how patterns of behavior or perceptions might differ from one group or type of person to another - for eg: do people with disabilities experience emotional labor differently from those without disabilities? Do women perceive talkativeness (or lack thereof) differently than men? Do communication styles differ from one culture to the next?

three characteristics of "systematic observation"

- intentional - replicable - valid

What is the key disadvantage of experiments?

- most experiments must take place in ideal "laboratory" conditions => most experiments must take place in ideal "laboratory" conditions - require so much planning and control that they oftentimes represent ideal conditions that do not necessarily approximate well to the messiness of the "real world"

advantages of surveys

- obtained from large random samples of people, thus allowing for more robust conclusions - typically take place in naturalistic settings and can therefore have strong generalizability

what is meant by the term "positivism" and summarise the key elements of this approach.

- the view that the study of the social world should be conducted according to the principles of natural science. - A positivist approach to sociology holds that objective knowledge can be produced through careful observation, comparison and experimentation

The value of a correlation coefficient must be between... (Finish the sentence by specifying the lower and upper limits.)

-1 and 1

A significant correlation test indicates that we have sufficient evidence to conclude that the population correlation coefficient is not equal to _________. (Finish the sentence.)

0

You conduct an experiment in which randomly selected young women are solicited in the street by a good-looking young man who approaches each woman with the following line: "Excuse me! I think you look ravishingly pretty! I just thought you might be happy know... I'm going to work now, but do you think we could have a drink somewhere in the afternoon? If you give me your phone number, I'll ring you up." It is recorded whether the woman complies with the man's request and shares her phone number with him. The purpose of the experiment is to find out whether women's decisions are influenced by what the man is wearing. You test a random sample of 50 women under each of the following conditions: (A) the man is wearing an elegant suit, (B) the man is wearing casual clothes but well-known and expensive brands, (C) the man is wearing no-name jeans and a T-shirt. Suppose the that the main chi-square test yields a significant result and therefore you decide to conduct post-hoc analysis to determine which pairs of conditions differ significantly. Because you will conduct 3 comparisons, you wish to keep the familywise error rate at (or below) 0.05, and therefore you decide to use the Bonferroni adjustment. At what level of significance will you conduct each post-hoc comparison?

0.0167

In a study, the researchers examined the relationship between extraversion and self-confidence. It turns out that the value of Pearson's correlation coefficient is 0.57. What percentage of the variability in self-confidence can be explained by differences in the degree of extroversion?

0.57^2 = 0.325 => 32.5% of the variability in self-confidence can be explained by differences in the degree of extraversion.

Calculate the familywise error rate when conducting 2 tests, each at the 0.05 level of significance. (I.e. what is the probability that you will make at least one Type I error?)

1 - 0.95*0.95 =0.0975 => 9.75%

Calculate the familywise error rate when conducting 3 tests, each at the 0.05 level of significance. (I.e. what is the probability that you will make at least one Type I error?)

1 - 0.95^3 = 0.142625 => 14.26%

Calculate the familywise error rate when conducting 6 tests, each at the 0.05 level of significance. (I.e. what is the probability that you will make at least one Type I error?)

1 - 0.95^6 = 0.2649 => 26.5%

Here are the ages (in years) of five people who took part in a study: 18, 26, 28, 22, 21. What is the median age of this small sample?

22 (You must sort the list from lowest to highest first)

Here are the ages (in years) of five people who took part in a study: 18, 21, 22, 26, 28. What is the mean age of this small sample?

23

A researcher conducts an observational study on the Budapest Metro line. The researcher observes young people traveling on the metro line and records the time (in seconds) that passes between the moment that the person enters the carriage and the moment he/she first looks at his/her smart phone. The observation of five young people yielded the following times: 12, 13, 21, 28, 36. Calculate the range of these values.

24

I've never seen the type of design called "Four-Group Design With Pre-Post and Multiple Controls" in actual practice. The only advantage I see is that in addition to the effect of the treatment (i.e. the training), you can measure another type of effect. Do you see which and how?

???

Let us suppose you run a study to test whether a candidate's evaluation in job interviews can be improved by a recently developed one-day training course on nonverbal communication skills. In this example, O stands for mock interviews with HR experts who evaluate the participants, and X stands for the training the participants receive. The so-called "Simple Causal Design" would be an inadequate choice in this case (in fact, I think this type of design would be inadequate to test pretty much anything...) Do you see why? (Hint: How would you interpret the results of such a study?

???

Suppose that in an experiment the independent variable is the amount of alcohol consumed: 0 ml, 25 ml, 50 ml, or 100 ml. Give an example for a potential dependent variable (related to communicative behaviour). In the above example, how would you manipulate the independent variable? How would you measure the dependent variable that you came up with?

???

Suppose you want to generalise the findings of your survey to all of the (approx.) 10 000 students of Corvinus University Budapest. How would you (in principle) take a random sample of 100 students out of the 10 000? (Note: It is important that each and every student must have the same chance of being included in your sample.)

???

The "Causal Design with Control" is commonly used and gets rid of the problem with the "Simple Causal Design" that you have - hopefully - just identified. The "Causal Design With Pre-Post and Control" is supposed to be a further improvement, but I'm afraid it is flawed (at least in the form printed here). Do you see why? How could this design be improved?

???

What is the advantage of the improved version of the "Causal Design With Pre-Post and Control" over the "Causal Design with Control"?

???

When performing a t-test, we may choose to conduct a one-tailed test or a two-tailed test. What is the difference between the two cases?

A one-tailed t test is usually associated with a hypothesis that stipulates that the researcher believes one mean is greater than the other. A two-tailed test did not specify in advance which mean would be larger than the other. The researcher can rule out one of the two possibilities in one-tailed test.

A study is conducted to find out whether people are more likely to disclose personal information in a good mood vs. in a bad mood. The participants are not divided into two groups, but each participant is tested under both mood conditions. In this case, instead of the "independent samples t-test", another type of t-test must be used. What is it called?

A paired samples t test

What is meant by the requirement that measurements/observations in quantitative research must be valid?

A small mistake could lead to something entirely different in your study. => A systematic and valid observation would ensure that you are observing what you intend to observe rather than some other phenomenon.

Which Greek philosopher is mentioned by name in relation to rhetorical tradition?

Aristotle

The mean

Average

A simple linear correlation between two continuous variables is called b________ correlation. (Write the missing term.)

Bivariate

We can use the One-Way ANOVA procedure whenever the independent variable is c__________ and the dependent variable is c___________. (Fill in the gaps.)

Categorical and continuous

List the four types of variable used in quantitative research. (Just provide the terms, you need not define or illustrate each type.)

Categorical: nominal, ordinal (measured in categories) Continuous: ratio, interval (measured in number)

In some rare cases, the researcher can collect data from each member of the whole population of interest (rather than a subset of the population). In such cases, the researcher is said to have conducted a c__________. (Write the technical term.)

Census

The strength and direction of the linear correlation between two continuous variables is measured by a statistic called the correlation __________. (Fill in the gap with the proper term.)

Coefficient

Quantitative - Primary objective

Create, expand, and refine theory through systematic observation of hypothesized connections among variables

Qualitative - Primary objective

Create, expand,and refine theory by observing and interacting with people in their natural environment to discover rich explanations and unique instances

Rhetorical - Primary objective

Create, expand,and refine theory through critical analyses of public discourses guided by particular theoretical stances and perspectives

When performing a t-test, the t-statistic that we compute must be compared to a c_________ v_______ that we can look up in a table.

Critical value

The scatterplot suggests that Pearson's correlation coefficient will not be an adequate measure of the relationship between the level of anxiety and performance because the relationship is c_________ rather than linear. (Write the missing term.)

Curvilinear

When conducting a One-Way ANOVA, small or moderate violations of the assumption of homogeneity of variance do not severely affect the results of the analysis provided that the sample sizes are (approximately) e__________. What is the missing word?

Equal

(?) What is the key advantage of experiments over questionnaire surveys?

Experiments provide evidence of causation.

The test statistic we compute in One-Way ANOVA is denoted as __. (Fill in the gap.)

F-test

Give an example for the operationalisation of a variable. (Name the variable and describe how it could be operationalised.)

For example, if we are concerned with the effect of media violence on aggression, then we need to be very clear what we mean by the different terms. In this case, we must state what we mean by the terms "media violence" and "aggression" as we will study them. Therefore, you could state that "media violence" is operationally defined (in your experiment) as 'exposure to a 15 minute film showing scenes of physical assault'; "aggression" is operationally defined as 'levels of electrical shocks administered to a second 'participant' in another room'.

give an example in which the validity of an observation/measurement is doubtful

For example, you must devise an appropriate way of observing emotion: - if you analyze statements that people make during conversation, how do you know that you are learning about emotion rather than verbal ability - If you use a survey, how can you separate their emotions from what they think they should have indicated about the emotionto reflect social desirability?

Write the name of the chart type that is most commonly used to visually represent the distribution of a continuous variable?

Histogram

When calculating Pearson's correlation coefficient between two variables x and y, we assume that for each value of x, the distribution of y has the same constant variance. What is the technical term for this assumption?

Homoscedasticity (or homogeneity of variance)

Qualitative - Example question

How questions: - How do doctors provide emotional support to patients? - How do teachers communicate motivational messages to students? - How do romantic couples negotiate conflict? - What types of messages do people enact and receiveto maintainand resist the enactment of organizational culture?

What is operationalization? (Provide a definition.)

How you will define and measure a specific variable as it is used in your study.

Suppose you conduct an experiment to find out whether the type of media used for studying has an effect on learning outcomes. You randomly assign your participants (secondary school students) into two groups: members of group A receive hard copies of a history book and are instructed to study a particular section for 30 minutes, while group B can access the same material online, using their notebooks. After 30 minutes' studying, each group is given a test to measure how much of the material they can recall. What is the independent variable in this study? What is the dependent variable?

IV: type of media DV: learning outcomes

Experiments

In an experiment, the research manipulates one variable—the independent variable—to see how that manipulation affects another variable—the dependent variable.

One important assumption required for a valid One-Way ANOVA is the independence of observations/measures. How can the researcher ensure that this assumption is satisfied in a quasi experiment? (In a typical quasi-experiment, we compare the behaviours of naturally occurring groups, such as people living in different regions, people belonging to different generations etc.)

In order not to violate the assumption, we do not randomly assign participants to groups.

One important assumption required for a valid One-Way ANOVA is the independence of observations/measures. How can the researcher ensure that this assumption is satisfied in a pure experiment? (In a typical pure experiment, we observe how people behave under different conditions.)

In order not to violate the assumption, we must randomly assign participants to groups.

List the three assumptions that must be satisfied for a One-Way ANOVA to yield accurate results? (You need not explain what each assumption means.)

Independence, normal distribution and homogeneity of variance

Suppose you conduct an experiment to find out whether the type of media used for studying has an effect on learning outcomes. You randomly assign your participants (secondary school students) into two groups: members of group A receive hard copies of a history book and are instructed to study a particular section for 30 minutes, while group B can access the same material online, using their notebooks. After 30 minutes' studying, each group is given a test to measure how much of the material they can recall. Name the statistical procedure that would be appropriate to analyze the data obtained from this study.

Independent sample t-test

List at least three specific research methods that belong to the qualitative research strategy.

Interviews, focus group, ethnography, participant observation and other techniques that allow the researcher access to individuals'accounts and stories

Qualitative - Data used

Interviews, observations, and other techniques that allow the researcher access to individuals' accounts and stories

What is the advantage of increasing the sample size in a study?

Larger sample sizes provide more accurate mean values, identify outliers that could skew the data in a smaller sample and provide a smaller margin of error.

Do the authors believe that quantitative research is better than other research strategies? What is their viewpoint on this matter?

No, because each has different goals, advantages and disadvantages

Is it necessary for all researchers to be equally well-trained in every methodological tradition? Why?

No, because it's nice to have basic knowledge, but there's no need to be equally well-trained in every methodological tradition.

When calculating Pearson's correlation coefficient, we assume that both variables follow a __________ distribution. (Fill in the gap with a technical term for the distribution.)

Normal

Give an example in which a general pattern is objectively observable, yet the pattern cannot be generalised to every individual case because there are many exceptions

Number of parties attended <=> alcohol consumption

(?) What is the name of the procedure that we conduct to determine whether we have evidence for difference between the population means of any number of groups?

One-Way ANOVA or Omnibus

Quantitative - Data used

Operational variables created through surveys and/or intentional manipulation (i.e., an experiment), content analysis

What are the two most common types of graph that can visually represent the frequencies/percentages of people that belong to different categories? (In other words, what types of graph can you use to represent the distribution of a categorical variable?)

Pie chart and bar chart

"The noisier the children in the class, the more stress the teacher will experience." This is an example for a __________ bivariate correlation. (Fill in the gap with a word that describes the direction of the relationship.)

Positive

The One-Way ANOVA is often followed up by p_____-_____ comparisons. (Write the missing term.)

Post-hoc

Rhetorical - Data used

Public discourses, including speeches, documents, and other publicly available statement

Correlational Design

R O involves one observation consisting of two measures that are correlated with one another

Causal Design With Pre-Post and Control

R O X O R O multiple measurements and multiple groups

Four-Group Design With Pre-Post and Multiple Controls

R O X O R X O R O O R O multiple measurements and multiple groups

Describe an experimental study using "Causal Design with Control". You may use the symbols R (= randomisation), O (= observation/measurement) and X (= treatment/programme) to represent the design scheme.

R X O R O adds the element of a control group

Simple Causal Design

R X O involves a simple causal relationship tested with one group;

List three different measures of dispersion. (You needn't explain what they are.)

Range, Variance, Standard deviation

ordinal variable

Represent categories that can be ordered from greatest to smallest Ex: education level (e.g., freshman, sophomore) Rank in videogames (Copper, Silver, Gold, Platinum, Diamond)

nominal variable

Represents categories that cannot be ordered in any particular way Ex: biological sex (male vs female), political affiliation (Democrat, Republican, or independent), football fan affiliation, hair color,...

Which of these 3 major traditional approaches to communication has the oldest origin (going back to antiquity)?

Rhetorical

What are the three major traditional approaches to communication research?

Rhetorical, quantitative, qualitative

The purpose of most quantitative studies is to generalise from a s________ to a p________. What are the missing words?

Sample to a population

The correlation between two continuous variables can be represented on a type of graph called the _________. (Fill in the gap.)

Scatterplot

survey packet

Several surveys combined (and conducted simultaneously within one study)

If the analysis of our data indicates that the difference between the two groups in our sample would have been (highly) unlikely to occur by mere chance (if the two groups do not in fact differ), we say that the result of the analysis is statistically s__________. What is the missing word?

Significant

Here are the ages (in years) of six people who took part in a study: 31, 21, 26, 30, 31, 24. What is the median age of this small sample?

Sorting: How sdfsdfdf21,24,26,30,31,31 -> The list has even number of values => Find the average of the 2 middle numbers: (26+30)/2 = 28

The t-test is also called "_________'s t-test" because the person who developed it first published the procedure under the pseudonym "_________". What is the missing word? (The same word fits both gaps.)

Student

Relationship questions

Tend to explore how one behavior exhibited by people is related to other types of behaviors. for eg: Are students' communication skills related to classroom performance? Are certain supervisor communication skills related to the emotional experiences of employees?

Explain when the t-test is the appropriate procedure to analyze the data obtained from a study. What are the key requirements concerning the independent and the dependent variable?

The IV must be categorical with two levels, while the DV must be continuous

A study revealed that the 95% confidence interval for the correlation coefficient between extroversion and self-confidence is (0.14; 0.29). This means that we can be 95% confident that... (Finish the sentence.)

The actual correlation would fall between 0.14 and 0.29.

What are the technical terms for the two types of variable in an experimental design?

The dependent and independent variable.

When performing an independent samples t-test, what are the two key assumptions about the distribution of the dependent variable in the two populations?

The independent samples t test assumes that your dependent variable is measured on either an interval or ratio scale. The independent samples t test assumes equal variances of the dependent variable in the population. Normal distribution + Homogeneity of variance.

The median

The middle score in a sorted (from lowest to highest) distribution; half the scores are above it and half are below it

List three different measures of central tendency. (You needn't explain what they are.)

The mode, the median and the mean

Write an example for a positive correlation that you think may be true.

The more infinity stones you have, the more power you acquire.

Write an example for a negative correlation that you think may be true.

The more time students spend doing Social Science Research Methods homework, the less time they have to play video games (or doing any leisure activities).

The mode

The most frequently occurring value(s) in a distribution

What is the null hypothesis that we test in a One-Way ANOVA procedure?

The population means are all equal.

Suppose you conduct a statistical test to see if you have evidence for a relationship between two variables. You conduct the test at the α = 0.05 significance level and find that p = 0.03. What is the conclusion that can draw from this result?

The result is significant, we can reject the null hypothesis since 0.03<0.05

How to calculate the standard deviation?

The square root of the variance (do not confuse this one with the deviation).

Quantitative - Example question

What questions: - What effect do punitive behavioral control statements have on a classroom? - What communicative behaviors are associated with different stages in a romantic relationship? - What communicative behaviors are used to respond to coworkers displaying emotional stress? - What is the relationship between the amount of time someone works for an organization and his or her perceptions of that organization's culture?

What is meant by the requirement that quantitative research must be intentional?

You must have intent because without it you would not know how to focus your observations, which serves as a foundation for systematic observation

The text also mentions a group of philosophers called the Sophists? Who were they and what did they do?

a new type of intellectuals, professional educators who toured the Greek world offering instruction in a wide range of subjects, with particular emphasis on skill in public speaking and the successful conduct of life (Stanford Encyclopedia of Philosophy)

quantitative methods definition

any approach that uses systematic observations to account for and generalize about human behavior

In a study, the researchers examine the relationship between extraversion and self-confidence. They measure both variables with a questionnaire. They obtain responses from 163 people. What is the degrees of freedom for the test they will conduct to determine whether they have enough evidence for a linear correlation between the two variables?

df = Sample size - 2 = 161

In a study, the researchers examine the relationship between extraversion and self-confidence. They measure both variables with a questionnaire. They obtain responses from 30 people. They are going to conduct a test to determine whether they have enough evidence for a linear correlation between the two variables. Suppose that the researchers assume that if there is a relationship between extraversion and self-confidence, then these two variables must be positively correlated. (In other words, the rule out the possibility of a negative correlation by principle.) If they wish to conduct the test at the α = 0.05 level of significance, what is going to be the critical value for Pearson's correlation coefficient?

p = 0.306 (one-tailed test)

two types of questions in quantitative researches

questions of relationships and questions of differences